Marlène Rasschaert

Brain uptake of gadolinium chelates : multimodal imaging and analysis of neurotoxicological consequences

Abstract

Gd chelates are widely used as contrast agents in magnetic resonance imaging. In 2015, the finding of T1 hyperintensities in brain structures was associated with the prior administrations of these agents in patients. This observation raised the question of the long-term tolerance of these molecules. The Gd³⁺ ionic radius is very close to that of Ca²⁺, and therefore this lanthanide interferes with numerous Ca²⁺-dependent biological processes. Its chelation by a ligand considerably improves its tolerance. Gd chelates are categorized into 2 classes: macrocylic and linear agents, differing in their thermodynamic stabilities, and therefore in their ability to dissociate. It is classically admitted that Gd chelates do not cross the healthy blood-brain-barrier. The observation of these hyperintensities, in the dentate nucleus of the cerebellum, the globus pallidus, and sometimes other structures, questioned this assumption.This thesis aimed to study the mechanism of Gd accumulation in the central nervous system: access pathways, tissue and subcellular location, Gd speciation). Potential neuro-toxicological effects associated with long term Gd presence in the brain were also researched.Using a rat model, we evidenced that the lower the thermodynamic stability of Gd chelates, the greater the cerebral Gd concertation was, thus confirming clinical observations. T1 hyperintensities exclusively appeared following administrations of linear Gd chelates. We also established that moderate renal failure potentiates Gd brain uptake in the case of linear Gd chelate. We also observed that brain structures accumulate even more Gd that their endogenous Fe concentration is high. Administration of linear Gd chelates resulted in an increased zincuria. Gadolinium vs. Zn transmetalation may be responsible for this effect.The combination of X fluorescence, transmission electronic microscopy, and NanoSIMS, showed Gd deposits at various scales and in various forms. It allowed us to document Gd pathways, and the role of endogenous metals and phosphorus in this phenomenon. X fluorescence analysis depicted, in rat deep cerebellar nuclei, that the majority of Gd was accumulated in the form of elongated and ramified structures, believed to be blood vessels where Gd would be retained in the perivascular area. By means of electron microscopy in rats, Gd insoluble deposits were observed in basal lamina of vessels, in cerebellar interstitium, and in the perivascular space. These Gd deposits, of spiny aspect, were rich in phosphorus, thus suggesting the presence of GdPO₄. Co-presence of Gd and phosphorous was also identified into glial cells, accumulated in intracellular lipofuscine pigments. No Gd deposits were found in rats treated with a macrocyclic Gd chelate.The established mechanistic hypothesis consists in the early access of Gd chelates to cerebrospinal fluid, followed by their passive diffusion into the parenchyma close to cerebral ventricles, through the ependyma. Encountering areas rich in endogenous metals and/or phosphorus, the less thermodynamically stable Gd chelates would dissociate, and Gd would bind endogenous macromolecules, or precipitate. Cerebrospinal fluid circulation along penetrating arterioles would also trap Gd at the perivascular level. Intact Gd chelates would be eliminated through perivascular glymphatic pathway, or “intramural periarterial drainage”, where probably dissociated Gd is also found.Except a non-specific hypoactivity, neurobehavioural, histopathological and neurochemical studies performed in rats did not demonstrate any obvious neurotoxicity, even at high doses. (Defended on February 11, 2019)