Bacteria-targeted imaging using vancomycin-based positron emission tomography tracers can distinguish infection from sterile inflammation

Abstract

Introduction: Bacterial infections pose major challenges in medicine. To guide effective infection treatment, faster and more accurate diagnostic modalities are needed. Bacteria-targeted molecular imaging can meet these needs. The present study was aimed at the in vivo evaluation of two ¹⁸F-vancomycin-based PET tracers, for detection of deep-seated Gram-positive bacterial infections. These tracers were bench-marked against the current standard of care, [¹⁸F]FDG.

Methods: The potential of [¹⁸F]BODIPY-FL-vancomycin and [¹⁸F]PQ-VE1-vancomycin ([4+2]photocycloadduct of 9,10-phenanthrenequinone-vancomycin and [¹⁸F]fluorinated vinyl ether) to distinguish bacterial infections from sterile inflammation was evaluated in a murine myositis model. Tracer specificity was assessed by infecting mice either with the Gram-positive bacterium Staphylococcus aureus (n = 12) or the Gram-negative bacterium Escherichia coli (n = 12). The contralateral leg was injected with Cytodex beads to induce sterile inflammation, or with phosphate-buffered saline for control. In parallel, mice were imaged with [¹⁸F]FDG (n = 12). Dynamic positron emission tomography (PET) measurements, biodistribution analyses, and immunohistopathology were performed to determine tracer distribution and bacterial burden.

Results: Both ¹⁸F-vancomycin-PET tracers accumulated at sites of infection, but not at sites of sterile inflammation, in contrast to [¹⁸F]FDG. The tracers exhibited distinct biodistribution profiles, with [¹⁸F]BODIPY-FL-vancomycin being cleared more rapidly. Both ¹⁸F-vancomycin-PET tracers displayed significant target to non-target ratios of 2.95 for [¹⁸F]BODIPY-FL-vancomycin and 1.48 for [¹⁸F]PQ-VE1-vancomycin.

Conclusion: Vancomycin-based PET is a potentially attractive approach to distinguish Gram-positive bacterial infections from sterile inflammation.

Keywords: 18F-vancomycin; Antibiotic-based tracer; Bacterial infection imaging; In vivo; Myositis; PET; Vancomycin.

Fig. 1

Experimental set-up and timeline of the murine myositis infection model. (A) Chemical structures of [¹⁸F]PQ-VE1-vancomycin and [¹⁸F]BODIPY-FL-vancomycin. [¹⁸F]PQ-VE1 is conjugated to the primary amine (R₁) of vancomycin, whilst [¹⁸F]BODIPY-FL is conjugated to the secondary amine (R₂) of vancomycin. (B) Simplified model of the cell envelopes of Gram-positive and Gram-negative bacteria. ¹⁸F-vancomycin-PET tracers will bind to d-Ala-d-Ala moieties in the peptidoglycan of Gram-positive bacteria, whereas binding is precluded by the outer membrane of Gram-negative bacteria. (C) In the murine left hind leg, infection was induced with S. aureus or E. coli. The right hind leg was injected with sterile Cytodex beads to induce inflammation or PBS for control. (D) Representative images of infected hind legs illustrating typical symptoms of infection by S. aureus or E. coli. Arrows indicate extensive haematoma and extravasation due to E. coli infection and indurated tissue due to S. aureus infection. (E) Timeline of the experimental procedures

Fig. 2

PET imaging of bacterial infection with [¹⁸F]BODIPY-FL-vancomycin, [¹⁸F]PQ-VE1-vancomycin or [¹⁸F]FDG. Mice with S. aureus or E. coli-induced myositis, or Cytodex bead-induced inflammation were monitored for tracer accumulation by 50-min dynamic emission scans and 10-min transmission scans in a microPET imaging system, directly post-administration of the respective tracer. PBS was injected for control. (A) Time activity curves for all PET tracers in the different ROIs; red, S. aureus infection; brown, E. coli infection; green Cytodex bead induced inflammation; blue, PBS control. Activity was measured as the SUV_mean and expressed as the mean ± SD. The list-mode data from the emission scan were reconstructed into 20 frames. (B) Tracer uptake in hind legs presented as ROIs for both ¹⁸F-vancomycin-PET tracers and [¹⁸F]FDG upon analysis of the last two 10-min frames (30 min post-injection) from the dynamic emission scan. Data is represented as the median with the interquartile range (IQR). Wilcoxon signed rank test: *, p < 0.05; **, p < 0.01, ***, p < 0.001. (C) Representative PET images (coronal view) of two animals per tracer, showing tracer signals in areas of E. coli or S. aureus infection in the left hind leg versus Cytodex bead-induced inflammation contralaterally. Arrows indicate the sites of injection of bacteria, beads or PBS. The colour legend on the right indicates the SUV_mean

Fig. 3

Biodistribution of [¹⁸F]BODIPY-FL-vancomycin and [¹⁸F]PQ-VE1-vancomycin. (A) Tracer distribution at 60 min post-injection for both ¹⁸F-vancomycin-based PET tracers. Radioactivity was measured as percentage of the injected dose per gram tissue (%ID/g), expressed as mean ± SD. Highlighted in colour are the hind leg tracer uptake, as well as control muscle tissue derived from the front limb (purple). “Sterile hind leg” is a split-bar representing the PBS control (blue) and Cytodex bead-induced inflammation (green). ‘Infected hind leg” represents S. aureus (red) or E. coli infection (brown). (B) Magnified representation of uptake of both ¹⁸F-vancomycin-PET tracers and [¹⁸F]FDG in hind legs with S. aureus- or E. coli-induced myositis, Cytodex bead-induced inflammation or control legs injected with PBS. The data is presented as the median with IQR. Wilcoxon signed rank test: *, p < 0.05; **, p < 0.01. (C) Target to non-target (T/NT) ratios expressed as mean ± SD for the three PET tracers. Non-target signal was defined as healthy muscle tissue as in (A)

Fig. 4

Histological analysis of tissue biopsies from mice with bacterial or sterile myositis. (A) H&E-stained tissue biopsies collected from hind legs of mice with myositis due to infection with S. aureus or E. coli, or due to injection of Cytodex beads. Large aggregates of bacteria and immune cells can be observed in S. aureus-infected tissue. Immune cells are marked with yellow arrowheads. Please note that injected Cytodex beads were not visualized in the Inflammation condition, which is due to the relatively low numbers of injected beads compared to bacteria and the fact that, unlike the injected bacteria, the beads will not replicate upon injection. The scale bar indicates 250 μm. (B) Immune cell infiltration in infected or inflamed tissue was visualized by DAPI-staining of the respective nuclei (purple arrowheads). S. aureus bacteria were stained with BODIPY-FL-vancomycin (green arrowheads). BODIPY-FL-vancomycin signals were completely absent from E. coli-infected tissues. Scale bars indicate 50 μm