Off-resonance saturation as an MRI method to quantify ferritin-bound iron in the post-mortem brain

Purpose To employ an Off-Resonance Saturation (ORS) method to measure the ferritin-bound iron pool, which is an endogenous contrast agent which can give information on cellular iron status. Methods An ORS acquisition protocol was implemented on a 7T preclinical scanner and the contrast maps were fitted to an established analytical model. The method was validated by correlation and Bland-Altman analysis on a ferritin-containing phantom. Ferritin-iron maps were obtained from post-mortem tissue of patients with neurological diseases characterized by brain iron accumulation, i. e. Alzheimer’s disease, Huntington’s disease and aceruloplasminemia, and validated with histology. Transverse relaxation rate and magnetic susceptibility values were also obtained for comparison. Results In post-mortem tissue, the ferritin-iron contrast strongly co-localizes with histological iron staining, in all the cases. Quantitative iron values obtained via the ORS method are in agreement with literature. Conclusions Off-resonance saturation is an effective way to detect iron in grey matter structures, while mitigating for the presence of myelin. If a reference region with little iron is available in the tissue, the method can produce quantitative iron maps. This method is applicable in the study of brain diseases characterized by brain iron accumulation and complement existing iron-sensitive parametric methods.


Introduction
ORS images were derived from 2D steady state free precession (SSFP) images (FISP-FID) with 1 TE/TR=3/6ms; inter-scan delay TR =2 s; FA= 10 degrees. A 'control' magnitude image (Mcontrol), 2 without saturation pulses, was also acquired. Saturated images (Msat) were produced by a saturation 3 module consisting of six hyperbolic secant pulses of bandwidth 500 Hz (B1=1.36 µT), followed by a 4 spoiler gradient of 3 ms duration. The off-resonance saturation pulses were phase-cycled to improve 5 the spoiling. A list of Msat was obtained upon varying the frequency of the off-resonance pulse. 6 Details of the off-resonance acquisition are reported in Table 2

12
MGE-magnitude images were fitted on a pixel-by-pixel basis to a mono-exponential decay function 23 , 13 after ruling out the presence of multi-components, using the Levenberg-Marquardt curve-fitting 14 algorithm to derive the transverse relaxation rate R2 * . The noise floor was reached in the last 2 or so 15 echoes in some regions of the tissue (ACP) containing extremely high iron accumulation but did not 16 affect the overall fitting quality. To more accurately fit the data, we excluded the pixels with 17 intensities at the noise level. µm slices were used for additional staining: non-heme (mostly trivalent) iron was detected with Perl's 23 Prussian blue (Merck 1.04984.0100); and immunohistochemical detection of myelin was done with 24 anti-myelin PLP antibody (Bio-Rad, MCA 839G) with second antibody Rb-aMs/biotin (DAKO) for 25 1h at room temperature, followed by ABC (avidin-biotin-complex/HRP, ABC Elite Kit, Vector) for 26 30 min at room temperature. Immunohistochemical detection of ferritin was done with ferritin 27 antibody (Bethyl A80-140, dilution 1:1000, overnight incubation at room temperature), with as 28 second step antigoat/biotin (Betyhl A50-204B, dilution:1:1000, 1h incubation at room temperature), 29 followed by ABC incubation for 30 min at room temperature. After the ABC treatment, the tissue was 30 rinsed three times with phosphate buffered saline and incubated in 0.05% DAB (Sigma-Aldrich) with 31 15 µl 30%H2O2/100ml for 5-10 minutes. After rinsing several times with demineralized water, the 32 slices were counterstained for 30 s with Harris haematoxilin and washed for 5-10 mins with tap water. 33 Finally, the tissue sections were dehydrated with ethanol 70%, 96% ,100%, and xylene.

Correlation between histology and ORS imaging 1
To quantify the degree of agreement between the histological staining for iron with the ferritin-bound 2 iron concentration derived from the ORS method, oval regions of interest (ROI, N=30) were drawn in 3 the grey matter regions of the ferritin-bound iron maps and ORS maps. The iron staining image was 4 manually co-registered (affine transformation) to the MRI maps with the TrackEM2 plugin in ImageJ. 5 The registered image stack was converted to an 8-bit greyscale image. Oval ROIs were carefully 6 drawn and propagated between the iron histological map and the MRI map with the ROI manager 7 tool. Mean grey values from the predefined ROIs were quantified 32 , after inspecting for the precise 8 placement of the ROIs. Pearson's correlation coefficient (ρ) of the association between histological 9 staining intensity and the ferritin-bound iron map was calculated per case.   Elevated R2 * levels are seen in the myelin-rich white matter and in the more superficial cortical layers. 28 In the QSM map, negative susceptibility is predominantly detected in the white matter. The where the ROR was drawn. The histological iron staining (Meguro) appears very intense across the 22 whole slice, with a slightly higher intensity in the caudate nucleus and the globus pallidus, which is 23 also captured by the [Fe]ORS map. In contrast to the previous cases, the ferritin staining is rather weak 24 across the whole slice, as discussed below. Additional histological results on all the tissue blocks are  Alzheimer's (AD), Huntington's (HD) and aceruloplasminemia (ACP) patient. This method can be 31 translated in vivo to evaluate tissue iron load and therapeutic efficacy on a pathophysiological level. 32 Water molecules diffusing around iron particles can be selectively excited by ORS pulses, based on 33 the correspondence between frequency offset and distance from the nanoparticle: the larger the positive contrast obtained by this method is distinct from, and can be observed also in the presence of, 1 MT effects 39 . Although the ORS-method was originally introduced to quantify/detect 2 superparamagnetic iron-oxide nanoparticles (SPION), here we show that ORS can also be employed 3 to assess the presence and the concentration of ferritin-bound iron, by virtue of the super-4 paramagnetic properties of the protein and its high concentration across the brain. Our results on the 5 agarose phantom show good agreement between nominal and fitted iron concentrations and support 6 the use of the method for iron quantification. 7 Visual comparison between the parametric MRI maps acquired in this study shows some advantages 8 of the ORS method: i) iron load appears highly localized in specific brain regions (mainly the gray 9 matter); ii) the white matter structures, which are highly myelinated, are predominantly masked out in 10 the [Fe]ORS maps, except for the ACP tissue where iron appeares largely diffuse throughout the whole 11 slice; iii) the [Fe]ORS map largely mirrors the QSM map, when the latter is free from artefacts; iv) the 12 strength of the association between the [Fe]ORS maps and the iron staining was moderately high in all 13 cases (ρ>0.5), and independent on the disease type. 14 15 The appearance of the ferritin-bound iron map in the tissue block obtained from the AD patient 16 suggests that, when looking at the temporal lobe, iron preferentially accumulates in the cortical grey 17 matter. This is in agreement with previous R2 *52,53 and QSM studies 54 reporting increased cortical iron 18 levels in patients with AD or mild cognitive impairment, which was associated with AD pathological 19 hallmarks, and an increased risk of cognitive decline. Magnetometry studies carried out on both 20 sporadic and genetic types of AD have also found that ferritin-bound iron is more abundant in the AD 21 group than in age-and gender-matched healthy controls 55,56 . 22 Previous analytical studies have detected significant increase in the total iron content of the temporal 23 lobe of AD patients with respect to controls 57,58 . However, the absolute iron concentrations ranged 24 from ~30 ng/mg 58 to 120 ng/mg (wet weight) 59 , since age, gender, disease state and technique 25 sensitivity can impact the measured iron values. The highest cortical ferritin-bound iron 26 concentrations detected in our tissue block were 1.37 ± 0.25 mM (mean ± sd), corresponding to 76.45 27 ± 13.95 ng/mg, which is in agreement with the estimate of ferritin-bound iron by magnetometry in the 28 AD temporal cortex 56 , after correcting for dry weight mass-loss. 29

30
The ferritin-bound iron map of the tissue block from the HD patient shows an increase of iron in the 31 putamen and part of the atrophic caudate nucleus. Alterations in brain iron metabolism with increased 32 iron accumulation have been previously identified by MRI in the striatal nuclei of patients with 33 Huntington's disease 26,60 . Iron accumulation in these structures seems to occur early on during the 34 disease process 61 , suggesting a key role of iron in the initiation and progression of the disease 11 . In 35 fact, in vivo MRI studies at both 3T and 7T have shown that iron accumulation 62 is associated with tissue atrophy of basal ganglia structures and is already present in the premanifest phase of HD 61,63,64 . 1 Our mean fitted iron concentration in the putamen was 4.66 ± 1.35 mM, equivalent to 260.0 ± 75.33 2 ng/mg, which agrees well with iron concentrations measured by inductively coupled mass 3 spectrometry 12 . 4 Finally, the ferritin-bound iron map of the ACP tissue block shows a very diffuse iron load, with mean 5 iron values of 2.57 ± 1.45 mM, in the putamen, which is about 50% lower than our recent 6 magnetometry study which quantified ferritin-iron in the same structure of the same patient 65 . This is 7 likely due to the combination of very fast relaxation times and the lack of an appropriate ROR in the 8 slice. In contrast to AD and HD, iron accumulation in ACP is directly related to its genetic 9 background, and results from the absence of functioning ceruloplasmin 13 . The lack of ceruloplasmin-10 mediated oxidation of ferrous iron (Fe 2+ ) to ferric iron (Fe 3+ ) impairs iron efflux from astrocytes, and 11 leads to massive iron accumulation within these cells, while neurons that are mainly dependent on the 12 supply of iron from astrocytes are probably iron-starved 13,66 . Although ferritin-bound iron appears to  Different MRI methods, in addition to those already mentioned, have been used to estimate the effect 20 of ferritin nanoparticles on the MRI parameters. In fact, ferritin-bound iron has long been identified as 21 the main source of iron-driving contrast, given the large abundance of the protein in the brain and its 22 magnetic properties. The magnetic susceptibility of a single ferritin protein was estimated as χ = 520 23 ppm for a fully loaded particle 19 , although this value might be an upper limit 18,71 . 24 Ferritin-iron displays a peculiar linear inverse dependence of T2 with B0 field 72 , a trend that is retained 25 in brain tissue and is attributed to the 'fingerprint' of iron stores. This characteristic is exploited by the 26 Secondly, we assumed that the iron loading of each ferritin protein was approximately equal to half of 3 the maximum filling capacity of the homonymous protein, while lower iron loading ranges, i. e. 4 between 1500-1850 iron ions within each ferritin protein, have been reported for AD brain tissue and 5 controls 77 . However, since the magnetization of the ferritin particle would not, or only minimally, 6 depend on the iron loading, this does not significantly affect the total ferritin-iron concentrations 7 reported here. 8 Additionally, the ferritin-bound iron maps were obtained from contrast maps that were referenced to a 9 region (ideally) without iron. Therefore, the iron concentrations here displayed cannot be considered 10 as absolute. This is especially clear in the case of the ACP tissue block, where the ferritin-bound iron 11 map shows comparable iron concentrations to the HD case, despite the striking difference in the 12 Meguro staining. This is probably due to the lack of a good ROR in the ACP tissue block. 13 Finally, T2 maps (not acquired in this study) can provide additional information on the degree of iron 14 accumulation and, in combination with the T2 * maps, could offer valuable information on the effective 15 size of iron-rich compartments 78 . 16

17
In conclusion, we adapted an ORS method 46,39 to quantify the ferritin-bound iron pool in the post-18 mortem brain tissue of three patients affected by neurological diseases associated with increased brain 19 iron. This method can aid the interpretation of R2 * and QSM maps, especially when these are 20 confounded by reconstruction artifacts or the co-presence of iron and myelin. The accuracy of the 21 ferritin-bound iron map depends on the availability, within the tissue, of a region without (or with 22 little) iron content with respect to the region of interest. We foresee that this method will find use in 23 the study of the progression of neurodegenerative diseases characterized by brain iron accumulation 24 and the assessment of iron chelation therapy.