7T MRI detects widespread brain iron deposition in neuroferritinopathy

Abstract Neuroferritinopathy is a disorder of neurodegeneration with brain iron accumulation that has no proven disease‐modifying treatments. Clinical trials require biomarkers of iron deposition. We examined brain iron accumulation in one presymptomatic FTL mutation carrier, two individuals with neuroferritinopathy and one healthy control using ultra‐high‐field 7T MRI. There was increased magnetic susceptibility, suggestive of iron deposition, in superficial and deep gray matter in both presymptomatic and symptomatic neuroferritinopathy. Cavitation of the putamen and globus pallidus increased with disease stage and at follow up. The widespread brain iron deposition in presymptomatic and early disease provides an opportunity for monitoring disease‐modifying intervention.


Introduction
Neuroferritinopathy (OMIM #606159) is a disorder of neurodegeneration with brain iron accumulation (NBIA type 3), typically presenting in mid-adulthood with a slowly progressive movement disorder and cognitive impairment. 1 Most affected individuals have a single base insertion in exon 4 (c460InsA) 2 of the ferritin light chain (FTL) gene but six other mutations in the same gene have been identified. 3All variants likely restrict ferritin ironsequestering, causing iron deposition and eventual cell death. 4The long presymptomatic stage in neuroferritinopathy presents an opportunity for iron chelation before irreversible neurodegeneration.
Brain iron accumulation can be detected in vivo using MRI. 5,68][9] Other case reports of 3T MRI in neuroferritinopathy show cortical iron deposition, particularly in the motor cortex, and suggest this "pencil sign" may be specific to neuroferritinopathy. 10,11 Here, we report results using ultra-high field (7T) MRI in three individuals with a pathogenic FTL variant at varying disease stages and a healthy control.Our aim was to determine if 7T MRI could detect brain iron deposition in the superficial and deep gray matter in both presymptomatic and symptomatic neuroferritinopathy.

Methods
We recruited three participants with the same FTL gene mutation (c.460InsA; p.Arg154LysfsTer27): An asymptomatic male in his 30s with a positive family history; a woman in her 50s with a 1-year history of mild dysarthria, limb ataxia, and dystonia; a man in his 40s with a 10-year history of orofacial dyskinesia, dysarthria, limb tremor, and dystonia.An asymptomatic woman in her 20s with no family history of neuroferritinopathy or other neurological or psychiatric disease was used as a healthy control.All individuals gave written informed consent to participate (REC 13/YH/0310).The participant with 10 years disease duration returned for a 2-year follow up assessment and scan.We have not provided detailed demographic information to preserve participant anonymity given the rarity of neuroferritinopathy.None of these patients are included in other published studies.
Qualitative review was performed by a neuroradiologist (TM).R 2 * maps were obtained using the Auto-Regression on Linear Operations algorithm. 12QSM maps were derived using the QSMbox 2.0 Multi-Scale Dipole Inversion algorithm. 13,14Tissue segmentation of MP2RAGE images was performed using SPM12 default settings.Cortical regions were derived from the Harvard-Oxford atlas and subcortical regions from the PD25 atlas. 15The dentate region of interest was derived from the Diedrichsen template. 16We summarized the R 2 * and QSM concentrations only from voxels predicted as gray or white matter.
We used CSF fraction of the putamen and globus pallidus to measure cavitation in these regions as a postmortem study of neuroferritinopathy confirmed that the cavitation appearance on MRI is caused by loss of brain tissue. 8CSF fraction was calculated using SPM12 tissue segmentation maps and the PD25 atlas regions using the spm_summarise function in SPM and Matlab 2021b (Mathworks, USA).

Results
In participants with an FTL mutation, R 2 * maps demonstrated markedly increased signal compared to controls, consistent with iron deposition, throughout the gray matter of the supratentorial brain.This was most pronounced in the primary motor cortex.There was less involvement of the cerebellum, except in the dentate nucleus where signal was increased in neuroferritinopathy.There was high signal in deep gray matter structures including the putamen, globus pallidus, red nucleus, substantia nigra, and subthalamic nucleus (Fig. 1A).The participants with 10 years of neuroferritinopathy symptoms had higher magnetic susceptibility compared to the presymptomatic FTL carrier and individual with 1 year of symptoms.There was no consistent difference in magnetic susceptibility between the presymptomatic FTL mutation carrier and the participants with only 1 year of symptoms (Table 1).QSM demonstrated similar intensity distribution to R 2 * maps (Table 1, Fig. 1B).R 2 * and QSM maps were largely unchanged, with minor changes in most regions with formal quantification, in the single neuroferritinopathy participant with a 2-year interval scan (Table 1).
CSF replacement of brain tissue (cavitation) in the putamen and globus pallidus increased with disease duration in neuroferritinopathy (Fig. 2).There was no cavitation in the healthy control nor the presymptomatic FTL mutation carrier.Early cavitation was apparent in the patient with 1 year of symptoms and severe cavitation after 10 years disease duration which then increased over 2 years (Table 1).

Discussion
Ultra-high field MRI shows R 2 * and QSM values are increased, suggesting iron accumulation, throughout the brain in neuroferritinopathy.This was already widespread in the presymptomatic participant.More advanced disease was associated with greater cavitation of putamen and globus pallidus but there was not a clear increase in iron deposition.
We show that 7T MRI can detect the iron deposition in deep and superficial gray matter, seen at postmortem but not with lower field (1.5T) MRI. 1,3,8-107T MRI revealed involvement of widespread cortical regions, including the motor cortex, and deep gray matter structures, including the red nucleus and substantia nigra.
There was no increase in iron deposition between participants with different disease stages, or in the single case with longitudinal follow up.In contrast, there was a clear association between cell loss in the striatum and more advanced disease.The presymptomatic participant had widespread iron but no striatal cavitation while symptomatic participants had a similar extent of iron deposition but severe cavitation, which increased with follow up.This suggests that it is the cell loss which eventually follows iron deposition that causes the clinical features of neuroferritinopathy.This is supported by previous findings of iron deposition in early childhood, decades before anticipated symptom onset 7 and postmortem evidence of a presymptomatic case with widespread brain iron deposition but no neuronal loss. 8Alternatively, any increase in iron deposition may be undetected by MRI due to a ceiling effect of magnetic susceptibility.This has been suggested in other NBIAs with postmortem validation 17 but not in neuroferritinopathy. 8 The widespread iron deposition and subsequent neurodegeneration mirrors the clinical presentations of neuroferritinopathy and explains why many symptoms are not typically associated with a focal disease of the basal ganglia. 1 Limb stiffness has been attributed to dystonia but can be associated with pyramidal signs 1 which would be explained by iron accumulation in the motor cortices.The extensive cortical iron deposition explains the heterogenous cognitive impairments which include executive dysfunction, disinhibition, emotional lability, psychosis, and apraxia. 18ur study has limitations.This was a pilot study with only 3 FTL mutations carriers and a single control with  unmatched demographics.Only a single participant had longitudinal imaging.This prevents any definitive conclusions on the relationship between iron accumulation, confounders (e.g., age and sex), and clinical phenotype.The control participant was younger than any of the FTL mutation carriers which may confound our results as brain iron accumulates with age.We used magnetic susceptibility as a measure of brain iron but this can be affected by other substances including myelin.Our results are corrected for brain atrophy, but a larger study would enable more robust statistical analysis of the relationship between brain atrophy and magnetic susceptibility.
In conclusion, this pilot study using ultra-high field MRI suggests neuroferritinopathy is associated with widespread superficial and deep gray matter iron deposition before symptom onset.Neuroferritinopathy has a presymptomatic stage of decades, 7 which we show can be detected with ultra-high field MRI and provides an opportunity for disease modification before irreversible neurodegeneration and striatal cavitation.Iron chelation is a promising treatment for neuroferritinopathy, with a causative link between FTL mutation, iron accumulation, and neurodegeneration. 3,4Our pilot study supports the use of ultra-high field MRI as a trial endpoint, 1,9 similar to other NBIA subtypes. 19Larger studies are required to test the association between regional variation in iron deposition and clinical phenotype.

1360 ª 2024
The Authors.Annals of Clinical and Translational Neurology published by Wiley Periodicals LLC on behalf of American Neurological Association.

Figure 1 .
Figure 1.Brain iron accumulation in neuroferritinopathy. (A) Two representative axial R 2 * maps from each participant.Brightness and contrast widths are the same for all images.The R2* images were thresholded at 0.15 ms À1 to improve visualization.(B) Two representative axial quantitative susceptibility v maps from each participant.

Table 1 .
Quantification of magnetic susceptibility v and striatum cavitation (% of CSF in region) in neuroferritinopathy.QSM) in ppm.R 2 * values in ms-1.%CSF values are percentage of region containing CSF (a proxy for cavitation).