Publications
Take a look

by Irma Mahmutovic Persson  on behalf of the TRISTAN Consortium

EMIM 2024

Abstract

To address the increasing issue of drug-induced toxicity, both clinical and experimental studies (within the EU-IHI funded project TRISTAN) are aiming to find translational Imaging Biomarkers that can identify incidence as well as reveal progression of lung injury at an early stage. Together within TRISTAN we are aiming to improve drug safety by development of novel imaging biomarkers. Various imaging techniques such as MRI, PET and CT are used for lesion assessment in lungs of patients and animal models of ILD.
The animal models developed to mimic the clinical aspects of disease have their limitations and opportunities – and the same goes for imaging biomarker development. Many previously promised applications known from oncology and brain-imaging are yet difficult to directly transfer to lung imaging. Many aspects of organ movement, influence of anaesthesia compounds and duration of scan acquisition, needs to be further optimised and automated for better and faster output of accurate data (1). Longitudinal imaging sessions and more non-invasive imaging approach is warranted in order to find the best diagnostic imaging biomarkers. In particular interest are those biomarkers that can characterize and differentiate inflammation from fibrosis within the lung.
In our experimental setting, we have employed the bleomycin-induced injury models in rats and mice and applying different MRI sequences combined with PET-CT imaging co-registration (2). By varying between different Ultra Short Echo (UTE) sequences, lesions of different characteristics can easily be visualized. With the PET-tracer [F18]FDG, typically used in inflammation-imaging, we could see increased signal uptake, significantly higher in bleomycin-challenged lungs compared to healthy controls during the early time points post-injury. However, also a late-phase [F18]FDG-uptake has been observed which is suggested to occur due to active fibrotic processes, similar to tumour metabolism known as the Warburg effect. Since the fibrogenesis might start early while the inflammatory phase is present in this type of model, we aimed to track fibrogenesis with two other, more specific, PET-tracer.
We have performed targeting of collagen type I and fibroblast activation (FAP), at various time points in the lung tissue, using the same bleomycin model in rats (2-4). Both studies were performed according to the same workflow of multi-modality approach with combined MRI/PET/CT longitudinal imaging post-instillation of bleomycin in rats. The Collagen-I tracer named CBP was conjugated to the radionuclide Cu64 while the FAP-tracer included the radionuclei Zr89. The collagen tracer uptake, targeting newly synthesised Collagen-I, was significantly increased in bleomycin-challenged lungs compared to controls, both at early and late time points. The majority of the CBP lung uptake was found at the borderline of MRI- and CT-defined lesions. This was further confirmed by autoradiography, thus on a microscopic level. The other fibrogenesis PET-tracer is targeting the fibroblast activation protein. This study in particular aimed to test the reproducibility of the bleomycin model, being confirmed by histology and cell data but also in parallel followed by [F18]FDG-imaging across the research sites Lund University and Radboud University in Nijmegen (Figure 1). Furthermore, the novel PET tracer tracking early signs of fibrosis onset was investigated, showing promising ability to detect fibroblast activation already 2 weeks after injury, while inflammation still might be present (4).
The intratracheal model has its limitations for certain ILD groups seen in the clinical settings, such as the incidence scenario observed in patients developing drug-induced ILD over time upon bleomycin treatment (5). Therefore, we are also developing animal models where systemic and chronic exposure of bleomycin is performed while also looking into other drug-induced pathologies occurring from other medication classes known to induce ILD in patient cohorts (6: see Figure 2).

Conclusion: As technology is advancing and imaging becomes more available, we have to keep pushing the animal models and imaging biomarkers forward, for better and more relevant readouts. Investing the effort in development of more translational models will bring more knowledge and will become more suitable for studies on intervention and therapeutic titrations.