Background: Exact localization of deep venous thrombi as well as determination of thrombus age is still a serious problem in the clinical setting and has significant implications for antithrombotic treatment regiments, since it will help to adequately adjust doses in lysis schemata. Thrombus detection by conventional ¹H MR methods, like angiography or T1/T2-weighted techniques can be difficult, because small, non-occlusive thrombi have only minor impact on blood flow and may not give rise to a clear signal in weighted images. To overcome this limitation, we have recently used ¹⁹F MRI for detection of early thrombi utilizing a targeted perfluorocarbon nanoemulsion (PFC) [1]. Due to the fact that ¹⁹F is nearly absent from biological tissue, ¹⁹F-containing agents result in an unequivocal positive contrast without any background signal. In order to further distinguish between acute, subacute, and chronic thrombi, we now used additional probes binding specifically at different stages to thrombi and coupled them to PFCs with indvidual spectral signatures. Discrimination of the targeted agents was achieved by a novel multi chemical shift selective imaging technique (mCSS-RARE [2]) which was used for simultaneous detection of three different PFCs.

Methods: For ‘multicolor’ ¹⁹F MRI, we generated nanoemulsions which contained 10% perfluoro-15-crown-5 ether (PFCE), 40% perfluorooctyl bromide (PFOB) or 40% perfluoro-1,3,5-trimethylcyclohexane (PFCH). For PFOB and PFCH a diblock (semifluorinated alkane) was added to stabilize the nanoemulsions. The diameter of the emulsion droplets was in the range of 100-200 nm, displayed a narrow size distribution and a strongly negative ζ-potential. To target the emulsions to specific stages of thrombi, we used ligands for three different components of thrombi: the peptide α2-antiplasmin (α2^AP) which is cross-linked by factor XIIIA to the fibrin network in the very early phase of thrombus formation [3], EP2104R, a peptide which binds to the fibrin network of early as well as chronic thrombi [4], and the single chain antibody anti-LIBS which binds to the activated gpIIb/IIIa receptor of platelets [5], which is predominantly invoked in the early and intermediate stage of thrombus development. Ex vivo experiments were carried out with thrombi obtained by platelet rich plasma of human blood which was incubated with targeted PFCs at different times after thrombus formation. For in vivo experiments, a murine model of deep venous thrombosis was used in that a filter paper soaked with 10% FeCl₃ was applied for 8 min to the external side of the vena cava inferior which resulted in induction of non-occlusive thrombi. Combined ¹H/¹⁹F MRI was performed at a 9.4T Bruker Avance^III WideBore NMR spectrometer using a microimaging unit (Micro 2.5) with actively shielded 40-mm gradients (1 T/m maximum gradient strength, 110 µs rise time at 100 % gradient switching). Mice were placed in a 25-mm ¹H/¹⁹F birdcage resonator and after acquisition of anatomical ¹H reference images, a ¹⁹F mCSS-RARE sequence with the following parameters was used: 2.56x2.56 cm2 FoV, 64x64 matrix, 1 mm slice thickness, TR 2500 ms, RARE factor 32, 256 averages, 20 min acquisition time.

Results: In a first step, ¹⁹F MR spectroscopy was used to identify the resonance frequencies specific for the individual PFCs, which were subsequently used for ¹⁹F mCSS-RARE to acquire artifact-free images of multiresonant PFCs (Fig. 1). Next, all ligands were successfully coupled to PFCE using sterol-based post-insertion [1] or a biotin/avidin system and were shown to specifically label ex vivo generated human thrombi. In contrast, corresponding control emulsions did not result in any detectable ¹⁹F signal within thrombi. Thereafter, EP2104R-coupled PFOB and PFCH nanoemulsions were demonstrated to be as well suitable for identification of thrombi as ¹⁹F hot spots (Fig. 2A) confirming that all PFCs (PFCE, PFOB, and PFCH) can be equipped with targeting ligands for thrombi and differentiated by ¹⁹F mCSS-RARE. Using a deep venous thrombosis mouse model, we proved that α2^AP-PFCE and -PFOB can also be detected in thrombi by ¹⁹F mCSS-RARE in vivo. In separate experiments, we were able to show that early thrombi are labelled by both α2^AP-PFCE and EP2104R-PFOB, whereas subacute thrombi are exclusively labelled with EP2104R-targeted PFOB or PFCH only (Fig. 2B).

Conclusions: The results demonstrate that the presented approach holds the potential to assess the current stage of thrombosis via ‘multicolor’ ¹⁹F MRI by simultaneous visualization of FXIIIa activity, fibrin and activated platelets. Since this technique is not restricted to thrombi, it may also be used for in situ labeling of a variety of other targets, e.g. different cell populations like immune or stem cells, and can thus substantially extend the field of application of ¹⁹F MRI for basic research and possible clinical applications.