Relaxation by clustered ferritin: a model for ferritin-induced relaxation in vivo Yves Gossuin, 1 * Pierre Gillis, 1 Robert N. Muller 2 and Aline Hocq 1 1 Biological Physics Department, University of Mons-Hainaut, 7000 Mons, Belgium 2 NMR and Molecular Imaging Laboratory, Department of General, Organic and Biomedical Chemistry, University of Mons-Hainaut, 7000 Mons, Belgium Received 29 June 2006; Revised 13 November 2006; Accepted 28 November 2006 ABSTRACT: Ferritin, the iron-storing protein of mammals, is known to darken T 2 -weighted MR images. This darkening could be used for the non-invasive measurement of an organ’s iron content. Unexplained discrepancies exist between T 2 data obtained in ferritin-containing tissues and aqueous solutions of ferritin. The clustering of the protein induced by trypsin is used to evaluate the effect of ferritin agglomeration on the relaxation rates. Although the longitudinal relaxation is not significantly influenced by clustering, T 2 depends greatly on the stage of agglomeration: the transverse relaxation rate is higher for a clustered sample than for an unclustered sample. Moreover, the field and inter-echo time dependences of the relaxation rate indicate that the relaxation mechanism may be different between small clusters – where a linear dependence of 1/T 2 on B 0 is observed – and large clusters – where a quadratic dependence is observed. These results help to explain the relaxation induced by ferritin in tissues. Copyright # 2007 John Wiley & Sons, Ltd. KEYWORDS: ferritin; clustering; transverse relaxation; iron content of tissues INTRODUCTION Ferritin, the mammalian iron-storage protein, contains a superparamagnetic ferrihydrite (5Fe 2 O 3 ,9H 2 O) crystal (1,2) which accelerates the transverse NMR relaxation of water. T 2 -weighted MRI was proposed early on for the non-invasive quantification of ferritin-bound iron in the organs. Different MRI protocols have since been elaborated to study the distribution of ferritin in the liver (3–11) and in the brain, especially the extrapyramidal nuclei (12–19). Simultaneously, different groups have worked on the relaxation of aqueous solutions of horse spleen ferritin and hydrated iron oxide nanoparticles (20–27), in order to provide an understanding of the MRI contrast caused by ferritin. It was first believed that the relaxation induced by ferritin was provoked by the diffusion of water molecules near the magnetic crystal contained inside ferritin. This explanation, known as the outer sphere mechanism, proved to be in contradiction to the experimental relaxation results. Among other things, the unique proportionality between 1/T 2 and the applied magnetic field observed in solution (21,22,24) and in tissues (23,28–30) does not match the quadratic dependence predicted by the outer sphere theory. It was finally shown that ferritin-induced T 2 shortening arises from the binding of water protons to the surface of the ferrihydrite crystal, and an appropriate theoretical model, the proton exchange dephasing model, was subsequently developed (26). This model qualitatively and quantitatively matches the results obtained for aqueous ferritin solutions but does not seem to be sufficient to explain the results for ferritin-containing tissues. It should be noted that the effect of ferritin on in vivo MRI contrast increases with the increase in the imaging magnetic fields, as does 1/T 2 . High fields, up to 3T, are already used to follow the evolution of iron content in diseased brains. However, the MRI protocols for evaluation of iron content are very sensitive to different parameters (type of organ, iron content, degradation of the tissues, etc.), which complicates their routine use in hospitals (31). This is caused by the significant differences between ferritin-induced relaxation in aqu- eous solutions and in tissues: for the same iron concentration, at 1 T, the transverse relaxation rate is more than three times higher in tissue than in aqueous horse spleen ferritin solution. Even in tissues, the rate is significantly higher in mouse liver than in spleen (30). The reasons for these differences are not really known, even though the in vivo clustering of ferritin in organs, which has been shown to depend on the type of organ (32–34), is thought to affect transverse relaxation properties. A first indication of the influence of clustering NMR IN BIOMEDICINE NMR Biomed. 2007; 20: 749–756 Published online 2 March 2007 in Wiley InterScience (www.interscience.wiley.com) DOI:10.1002/nbm.1140 *Correspondence to: Y. Gossuin, Service de Physique Expe ´rimentale et Biologique, Faculte ´ de Me ´decine, Universite ´ de Mons-Hainaut, 24 Avenue du Champ de Mars, 7000 Mons, Belgium. E-mail: Yves.gossuin@umh.ac.be Abbreviation used: TEM, transmission electron microscopy. Copyright # 2007 John Wiley & Sons, Ltd. NMR Biomed. 2007; 20: 749–756