Introduction

Increased brain iron deposition has been repeatedly reported in Alzheimer’s disease (AD) patients. The majority of AD studies have investigated in vivo iron levels in deep gray matter. Only few studies evaluated iron deposition in AD outside the basal ganglia ^1,2, which are clinically highly relevant but challenging to assess from an MRI perspective because of discontinuities in the magnetic susceptibility resulting mainly from air-tissue boundaries in the brain.³ To our knowledge, no published study has examined the relationship between the longitudinal iron change in the neocortex and the cognitive decline in AD. In this study, we assessed brain iron in a large clinical cohort of participants with AD by R2* relaxation rate mapping using an advanced correction method which also allows for the R2* calculation in neocortical regions. We compared the global and regional iron differences between AD cases and age-matched controls, evaluated the longitudinal R2* change in patients with AD over a 17 months follow-up period and correlated the longitudinal iron data with the cognitive decline.

Methods

101 participants with AD (mean age=72.7±8.7 years) and 101 age-matched healthy controls (HC) from a community-dwelling study were included. Of the 101 participants with AD, 59 had subsequent neuropsychological testing and MRI at a mean follow-up of 1.39±0.54 years. All study participants underwent MRI at 3 Tesla including R2* relaxation rates mapping, corrected for macroscopic field variations. Anatomical structures were segmented using FreeSurfer and median R2* rates were calculated in the neocortex and cortical lobes, the basal ganglia, hippocampus and thalamus. Multivariate linear regression analysis was applied to study the differences of the R2* levels between groups and the association between longitudinal changes in R2* and cognition.

Results

Characteristics of study participants are shown in Table 1. Smoking was the only one risk factor that was different between AD and HC. Table 2 shows the comparison of R2* in total and regional grey matter between AD and HC. After correcting for smoking and regional volumes, R2* remained higher in the whole cortex including total neocortex and the deep grey matter in AD patients as compared to HC (p <0.01; p= 0.01, respectively). Regionally, AD showed higher R2* in the temporal (p <0.01) and occipital lobes (p= 0.01) and in the caudate nuclei (p <0.01). Table 3 summarizes the global and regional annualized percentage R2* rates. R2* in parietal lobe decreased whereas R2* relaxation rates of global basal ganglia, putamen, caudate nucleus and thalamus increased over time (Table 3). R2* change in the temporal and occipital lobes, after adjustment for change in brain volume over the observational period, correlated significantly with change in cognition over 17 months (ß = -0.31, p = 0.02, ß = -0.34, p = 0.01, respectively).

Discussion and Conclusion

In the present study using R2* relaxometry with a correction technique for intravoxel dephasing, we found higher R2* levels in the global basal ganglia and total neocortex in patients with AD as compared to age-matched community-dwelling individuals without overt cognitive impairment. The regional differences of iron load between AD and control groups were found in the occipital and temporal lobes and in the caudate nuclei. Our study extends previous investigations in sample size and by longitudinally exploring if effects of R2* are related to cognitive functioning. Our results demonstrate that an increase of iron in the temporal and occipital lobes correlated with decline in cognitive functions independently of longitudinal change in brain volume. We can only speculate on the cause of the association between R2* and cognition in these brain regions. Both the temporal and the occipital lobe are strongly affected by beta-amyloid deposition. Iron is particularly present in beta-amyloid plaques ⁴ and it might well be that interactive effects between beta-amyloid and iron in amyloid-rich brain areas such as the temporal and occipital lobe are responsible for our finding that already small increases of iron in these regions had substantial effects on the cognition. Given these evidences and the fact that iron is attached to beta-amyloid might explain the association between R2* and cognition, at least to a certain extent.

Acknowledgements

This study was funded by the Austrian Science Fund (FWF grant numbers: KLI523, P30134 and I2889-B31) and the German Research Foundation (DFG DU1626/1-1).