4023

A quantitative tool for the speciation of mineralized iron in the brain
Lucia Bossoni1, Laurens Boers2, Andrew G. Webb1, and Louise van der Weerd1,3

1C. J. Gorter Center for High Field MRI, Leiden University Medical Center, Leiden, Netherlands, 2Leiden University, Leiden, Netherlands, 3Human Genetics, Leiden University Medical Center, Leiden, Netherlands

Synopsis

We implemented and optimized an Off-Resonance Saturation (ORS) pulse sequence on a 7T- preclinical scanner, with the aim of quantifying iron present in brain phantoms doped with ferritin-bound and magnetite-bound iron. These mineralized iron forms are known to be involved in ageing, and also in toxic cellular pathways in neurodegenerative diseases. We show that the ORS method offers the possibility to quantify, and possibly differentiate, iron ions in brains affected by neurodegenerative diseases which are characterized by disturbed iron homeostasis. Additionally, our results are discussed in the light of susceptibility values of the same minerals.

Introduction

While iron dis-homeostasis is known to occur in the brain of patients with neurodegenerative diseases, including Alzheimer’s disease (AD), only recently has attention been paid to the presence of different forms of mineralized iron1,2,3. Experimental evidence suggests that magnetite is present in the core of amyloid plaques, where it may contribute to initiate/progress the amyloid cascade and induce toxicity4. Another crucial form of iron, in terms of MRI contrast and biological relevance, is ferritin: a nanoglobular protein responsible for iron-storage. Iron and ferritin contents are altered across the cortices, in AD5,6.

Gradient-Echo images are routinely employed to detect iron overload in tissue7 and to obtain Quantitative Susceptibility Mapping (QSM) maps, which have been used to indirectly assess iron ex-vivo and in-vivo8,9. However, these methods do not offer speciation of tissue iron.

Here we develop and optimize an Off-Resonance Saturation (ORS) pulse sequence to obtain absolute concentrations of iron contained in ferritin and magnetite nanoparticles. We discuss our results and compare them with QSM values.

Methods

1.5%-agarose brain phantoms, one containing different concentrations of 20nm-magnetite nanoparticles, a second containing different concentrations of horse spleen ferritin, and a third one containing both species, were fabricated. Iron concentrations were chosen to cover the range of physiological and pathological conditions. The ORS preparation module consisted of multiple saturation sech-pulses, with 90° phase cycling in-between pulses, followed by a spoiler gradient in the slice-selection direction. The acquisition consisted of a post-excitation SSFP sequence, with four segments. An additional 90° phase cycling in between segments was implemented to improve spoiling. The Z-spectrum was inspected to assess the presence of magnetization transfer (MT) effects, which may confound the analysis. Afterward, the contrast curve was fitted to the following expression:

$$C=S_0 \frac{N}{\pi^2}\left[ \arctan \left( \frac{\alpha}{B_{eq}[Fe]}(\omega_0 + \Delta \omega_0/2)\right)-\arctan \left( \frac{\alpha}{B_{eq}[Fe]}(\omega_0 -\Delta \omega_0/2) \right) \right]+off$$

Where S0 is the magnitude of the unsaturated image, N is the number of ORS pulses, α a numerical factor10, ω0 and Δω the frequency and bandwidth of the ORS pulse respectively, [Fe] the iron concentration and off a constant. These last two are the fitting parameters. The equatorial field, Beq, was fixed to 167 mT for magnetite10 and 1.2 mT for ferritin11. Δω was fixed to 500 Hz, and ω0 was varied. Multi-echo gradient-Echo images were acquired on the magnetite-doped phantom, for comparative analysis. STI Suite toolbox was used for QSM reconstruction12.

Results

Fig.1 shows positive contrast on the first two samples. For these, the fit to the equation above returned the nominal [Fe], within the error bar (Fig. 2). For the mixed sample, a weighted fit was used to extract [Fe] bound to ferritin and magnetite, respectively. The fit returned the concentration of magnetite-iron with a good accuracy, while it underestimated the ferritin-iron fraction (Fig. 3).

QSM maps were used to extract mean susceptibility values (χ) from the magnetite sample. A correlative slope of 0.65 ± 0.10 ppb/µg iron/ml was found (Fig. 4), in agreement with the calculated value of 0.5 ppb/µg iron/ml13.

Discussion

The ORS method is known to generate positive contrast which is sensitive to superparamagnetic particles14. The flexibility of choosing the number of ORS pulses and delay between pulses allows water diffusion to increase the fraction of saturated protons10. When the ORS results are compared to χ-values, a high degree of inhomogeneity is seen in the sample-compartments and in-between slices, in the QSM maps. Such heterogeneity may derive from clustered magnetite nanoparticles (as indeed occurs in the brain) and from a χ-discontinuity produced by the sample holder. Despite this effect, we could estimate a correlative slope which is in good agreement with the theoretical value, and about a factor of two smaller than the value for ferritin15. Thus, in a sample doped with both iron species, QSM cannot be used alone to disentangle the two minerals. On the other hand, the ORS method has the ability to select ω0 and Δω to possibly filter-out MT effects and extrinsic χ-inhomogeneities, and to single out the contribution of the nanoparticles, provided that Beq is known. While more work is needed to better characterize the contrast of the mixed-sample, our preliminary results prove that the ORS method is a promising tool to quantify iron bound to different mineral forms.

Conclusions

The development of non-invasive methods to target the different molecular species of iron in the brain would help diagnosis and therapy of neurodegenerative disease. Our preliminary results show that ORS has the potential to differentiate and quantify ferritin and magnetite-bound iron in the brain.

Acknowledgements

We thank E. Vroon, M. Bulk and C. Liu for useful discussions, and L. Hirschler and E. Suidgeest for technical support. L.B. is supported by the Netherlands Organization for Scientific Research (NWO), through a VENI fellowship (016.Veni.188.040).

References

  1. Duyn, J.H. and Schenck, J. (2017) Contributions to magnetic susceptibility of brain tissue. NMR Biomed. 30, e3546
  2. Ropele, S. and Langkammer, C. (2017) Iron quantification with susceptibility. NMR Biomed. 30, e3534
  3. Bulk, M., van der Weerd, L., Breimer, W., Lebedev, N., Webb, A., Goeman, J.J., Ward, R.J., Huber, M., Oosterkamp, T.H., Bossoni, L. (2018) Quantitative comparison of different iron forms in the temporal cortex of Alzheimer patients and control subjects. Sci. Rep. 8, 6898
  4. Germán Plascencia-Villa, Arturo Ponce, Joanna F. Collingwood, et al., (2016) High-resolution analytical imaging and electron holography of magnetite particles in amyloid cores of Alzheimer’s disease. Sci. Rep. 6, 24873
  5. Connor, J.R., Snyder, B.S., Arosio, P., Loeffler, D.A., LeWitt, P. (1995) A quantitative analysis of isoferritins in select regions of aged, Parkinsonian, and Alzheimer's diseased brains. J. Neurochem. 65, 717–724
  6. Dedman, D.J., Treffry, A., Candy, et al., (1992) Iron and aluminium in relation to brain ferritin in normal individuals and Alzheimer's-disease and chronic renal-dialysis patients. Biochem. J. 287, 509–514
  7. Langkammer, C., Krebs, N., Goessler, W., Scheurer, E., Ebner, F., Yen, K. Fazekas, F., Ropele, S. (2010) Quantitative MR imaging of brain iron: A postmortem validation study. Radiology 257, 455–462
  8. Deistung, A., Schäfer, A., Schweser, F., Biedermann, U., Turner, R., Reichenbach, J.R. (2013). Toward in vivo histology: A comparison of quantitative susceptibility mapping (QSM) with magnitude-, phase-, and -imaging at ultra-high magnetic field strength. NeuroImage 65, 299–314
  9. Ayton, S., Fazlollahi, A., Bourgeat, et. a., (2017) Cerebral quantitative susceptibility mapping predicts amyloid-β-related cognitive decline. Brain 140, 2112–2119
  10. S. Delangre, Q.L. Vuong, C. Pob, B. Gallez, Y. Gossuin (2016), Improvement of the Off-Resonance Saturation, an MRI sequence for positive contrast with SPM particles: Theoretical and experimental study, JMR 265, 99–107
  11. R. Blakemore, R. Frankel, Iron Biominerals (2013) Springer Science & Business Media
  12. Wei Li Alexandru V. Avram Bing Wu Xue Xiao Chunlei Liu (2014) Integrated Laplacian‐based phase unwrapping and background phase removal for quantitative susceptibility mapping, NMR Biomed. 27: 219–227
  13. J. F. Schenck, Health and physiological effects of human exposure to whole-body four-tesla magnetic fields during MRI (1992), Ann N Y Acad Sci. Mar 31;649:285-301
  14. O. Zurkiya and X. Hu, Off‐resonance saturation as a means of generating contrast with superparamagnetic nanoparticles (2006), Magnetic Resonance in Medicine 56:726–732
  15. Zheng W, Nichol H, Liu S, Cheng Y-CN, Haacke EM. (2013) Measuring iron in the brain using quantitative susceptibility mapping and X-ray fluorescence imaging. NeuroImage, 78:68–74.

Figures

Figure 1. Unsaturated and ORS images (saturated at ω0=300Hz and subtracted from the unsaturated one) for magnetite and ferritin-doped phantoms. The iron concentration, in mM, is written in the sample compartments. Top row: Acquisition parameters for the magnetite-phantom were TE/TR=1.634/3.268 ms and FA=10°, voxel size=0.25x0.25x1.5 mm3. ORS images were obtained by varying ω0 in a [-700, 700] Hz range. Three ORS pulses were used. Bottom row: Acquisition parameters for the ferritin-phantom were as above except for TE/TR=20/40 ms, voxel size=0.25x0.25x1mm3. Five ORS pulses were used.


Figure 2. (a) Example of contrast data (yellow) and fit (red dashed line) for the magnetite-phantom. (b) Fitted iron concentration for the magnetite sample (blue circles) and nominal iron concentration in red. The shaded areas correspond to the confidence interval/uncertainty in iron concentration, respectively. (c) Z-spectrum for the ferritin-doped sample. An NOE effect possibly originating from aliphatic protons in the gel/ferritin was found, therefore only the positive frequency axis data were used for the contrast analysis. (d) Fitted iron concentration for the ferritin sample. The R2 shows the overall performance of the method.

Figure 3. (a) Contrast versus offset frequency (blue circles) and fit (red dashed line) for the third sample, containing both magnetite and ferritin. (b) Table representing the known concentration of iron in the two mineral forms, with percentage uncertainty, and fitted (estimated) iron concentration. Numbers in parenthesis refer to the confidence intervals.

Figure 4. From the top-left panel: magnitude map, unwrapped and background-subtracted phase, QSM map and mean χ-values from the magnetite-doped phantom. Acquisition parameters were TE=2.35 ms, 8 echoes, inter-echo-spacing 2.46 ms, TR=40.15 ms, FA= 18°, voxel size=0.25x0.25x0.25 mm3. χ-values relative to the agarose compartment are plotted versus the known iron concentration (bottom-left panel). The heterogeneity of the QSM map increases with the iron concentration and reflects the same patchy-pattern of the phase map.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
4023