1150

Triple Quantum Filtered Sodium Imaging at 21.1 T Reveals Dynamic Progression in a Preclinical Migraine Model
Nastaren Abad1,2, Ghoncheh Amouzandeh2,3, Jens T Rosenberg2, Michael G Harrington4, and Samuel Colles Grant1,2

1Chemical and Biomedical Engineering, Florida State University, Tallahassee, FL, United States, 2National High Magnetic Field Laboratory, Florida State University, Tallahassee, FL, United States, 3Physics, Florida State University, Tallahassee, FL, United States, 4Neurosciences, Huntington Medical Research Institutes, Pasadena, CA, United States

Synopsis

The brain allocates >50% of its energy reserves to the regulation of sodium homeostasis, indicating the critical importance of sodium and its fluxes in normal brain as well as neurological disorders. With the goal of localizing specific changes within intra- vs. extracellular compartments, this study utilizes Triple Quantum (TQ) coherence to evaluate 23Na fluxes in a preclinical rodent analogue of migraine. At a high spatial and temporal resolution, longitudinal scanning was performed at 21.1 T to quantify alterations in bulk and bound sodium during and following the onset of central sensitization.

Introduction

Ionic instability is implicit in clinical migraineurs. Following the hypothesis that increased sodium induces sustained neuronal excitability, we have previously demonstrated that bulk sodium changes predominantly in the ventricular (CSF) and trigeminovascular system (TGVS) correlate with the onset and progression of induced migraine [1]. This study mechanistically extends bulk quantification to identify bound vs. free sodium distributions as a means of probing cellular function related to ionic homeostasis and Na+K+ATPase transport.

Triple Quantum (TQ) coherences have been used non-invasively to separate sodium (23Na) MR signals between bound and bulk contributions based on differential relaxation rates and tissue properties [3,5]. TQ experiments target the rotationally restricted bound fraction, often attributed to the intracellular compartment, as opposed to the more freely moving single quantum transition. Performed at 21.1 T, this study aims to provide high spatial and temporal discrimination of the bound sodium fraction following the initiation and progression of central sensitization over 2 h.

Materials and Methods

Animal Model: Twenty Sprague-Dawley male rats were imaged in this study. While under anesthesia in the magnet, the rats were administered in situ with an IP injection of either 10 mg/kg of nitroglycerin (NTG) to provide conditions of a migraine analogue (n=8), or saline (n=8) as a control, with naïve (n=4) to serve as baselines (naïve data not shown).

MR Acquisitions: Using the 21.1-T ultra-wide bore magnet at the US National High Magnetic Field Laboratory and a linear 1H/23Na birdcage coil, 3D 23Na images were acquired using a modified gradient recalled echo (GRE) sequence that included a TQ preparation. Following pulse calibration (α90°=0.185-ms-hard-pulse), B0/B1 maps were acquired using two single quantum datasets.

This study utilized a three-pulse coherence transfer technique with a modified 12-step phase cycling scheme for B0 correction [3,5]. Flip angles α1,2,3 were set to 90° with τ1 = 7.5 ms, τ2 = 10 ms and τ3 = 0.18 ms to maximize signal. For a FOV of 6.4x6.4x3.2 cm, BW = 130 Hz/pixel and TR = 105 ms, TQ images were acquired with an isotropic resolution of (2 mm)3 for a total acquisition time of 20 min. A SQ dataset also was acquired with the same parameters but a standard DC and quadrature correction phase cycling scheme.A total of five sequentially repeated scans were acquired, including a pre-injection scan, over 2 h.

For 1H reference and segmentation, 3D spin-echo and time of flight scans were acquired with the same FOV as the sodium images, but at higher resolution.

Data Analysis: Data was acquired as a full echo and reconstructed offline using MATLAB. Acquisition matrices were zero-padded and reconstructed to (1 mm)3; no other filtering was performed. Intracellular sodium (ISC) and total sodium concentration maps based on TQ and SQ sequences were generated (Figure 1) as recommended by Fleysher et al., [2,4]. Composite images for 23Na SQ/TQ and 1H were manually segmented with ROIs placed in brain tissue and ventricular regions (Figure 1). Mean sodium concentrations in the 3D ROIs and data are presented as percent change from baseline concentration (Figure 2-4).

Results

Sustained increases of 23Na signal in specific brain regions are evident as early as 20-min post NTG injection. Increased 23Na MR signal in the thalamus and brainstem (Figure 2-3), both serving as central TGVS processing centers, occur almost immediately after NTG injection and for both extracellular and intracellular compartments. The cisterna magna and fourth ventricle also show sustained significances following NTG administration, with increases demonstrated in the aqueduct and third ventricle (Figure 4).

Discussion

The data provides spatially unique evidence of an early change in sodium concentration, markedly in the posterior fossa CSF and TGVS. Interestingly, the brainstem shows sustained and significant increases in both ISC and TSC. The thalamus, known to be involved in migraine receives direct input from trigeminal and retina ganglion cell afferents, shows early significance for the first time point in the ISC, whereas it demonstrated sustained significances in TSC. Though the unique changes in both brain tissue and ventricular system indicate a posterior fossa located migraine generator driven by extracellular sodium changes, dynamic contributions from sodium sources maybe masked by acquisition duration or volume averaging.

Conclusions

The present study presents novel insights regarding longitudinal changes in intracellular brain sodium concentration in an in vivo acute migraine rodent model. Early sodium increases, which consistently precede behavioral changes, support the theory of sodium flux as potentially causative. Further study of fluctuations of sodium and its modulation with treatment could help elucidate the dynamic features of migraine, locate the elusive migraine generator and guide development of therapeutic measures to correct disturbances of sodium homeostasis.

Acknowledgements

This work was supported by the NIH (R01-NS072497 and RO1-NS102395) and User Collaborations Grant Program (to SCG) from the National High Magnetic Field Laboratory, which is funded by the NSF (DMR-1644779) and the State of Florida.

References

[1] Abad N, Rosenberg JT, Hike DC, Harrington MG, Grant SC. Dynamic Sodium Imaging at Ultra-High Field Reveals Progression in a Preclinical Migraine Model. Pain 2018;159(10):2058-2065.

[2] Fleysher L, Oesingmann N, Brown R, Sodickson DK, Wiggins GC, Inglese M. Noninvasive quantification of intracellular sodium in human brain using ultrahigh-field MRI. NMR Biomed 2013;26:9-19.

[3] Fleysher L, Oesingmann N, Inglese M. B-0 inhomogeneity-insensitive triple-quantum-filtered sodium imaging using a 12-step phase-cycling scheme. NMR Biomed 2010;23:1191-1198.

[4] Inglese M, Madelin G, Oesingmann N, Babb JS, Wu W, Stoeckel B, Herbert J, Johnson G. Brain tissue sodium concentration in multiple sclerosis: a sodium imaging study at 3 tesla. Brain 2010;133:847-857.

[5] Tanase C, Boada F. Triple-quantum-filtered imaging of sodium in presence of B-0 inhomogeneities. J Magn Reson 2005;174:270-278.

Figures

Figure 1. Representative proton, total and intracellular sodium MR images in the rodent brain. in vivo coronal images of (A) 1H spin echo overlaid with angiographic information using time of flight acquisition and associated segmentations; 3D 23Na acquisitions at 2-mm isotropic resolution (zero-filled to 1-mm) (B) total sodium maps (SQ), (C) intracellular sodium concentration (ISC) maps (TQ) and (D) composite images for SQ (purple) and TQ (green) images, demonstrating filtering of free sodium signal predominantly from ventricular areas.

Figure 2. Temporal bar plots for percent change in total 23Na concentration between NTG and saline rodents. Percent change in 23Na concentration (%change±SE) in the thalamus, hippocampal formation (HCF), brainstem and cerebellum as a function of time after injection. Table inserts are indicative of maximum difference (mM). Utilizing a one-way ANOVA least significant different test for post-hoc comparisons at *p<0.05 between NTG (n=8) and saline (n=8) and by †p<0.05 for comparisons between NTG and naïve (n=4) rats. Naïve rodent data is not shown, and no significances were determined between naïve and saline controls.

Figure 3. Temporal bar plots for percent change in intracellular 23Na concentration between NTG and saline rodents. Percent change in intracellular 23Na concentration (%change±SE) in the thalamus, hippocampal formation (HCF), brainstem and the cerebellum, as a function of time after injection. Table inserts are indicative of maximum difference (mM). Utilizing a one-way ANOVA with a least significant difference test for post-hoc comparisons at *p<0.05 between NTG (n=8) and saline (n=8) and by †p<0.05 for comparisons between NTG and naïve (n=4) rats. Naïve rodent data is not shown, and no significances were determined between naïve and saline controls.

Figure4. Temporal bar plots for percent change in total 23Na concentration between saline and NTG rodents in the ventricular system. Percent change in ventricular 23Na concentration (%change±SE) in the cisterna magna(cisterna), fourth ventricle (4V), aqueduct, extracerebral CSF(extraC), third ventricle(3V) and the lateral ventricle(Lat V), as a function of time post-injection. Table inserts indicate maximum difference(mM). Utilizing a one-way ANOVA with LSD test for post-hoc comparisons at *p<0.05 between NTG (n=8) and Saline (n=8) and by †p<0.05 for comparisons between NTG and naïve (n=4) rats. Naïve rodent data is not shown, and no significances were determined between naïve and saline controls.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
1150