Introduction: Hyperpolarized ¹²⁹Xe gas is a safe, inhalable contrast agent for MRI that has primarily been used to image the air spaces of the lungs. ¹²⁹Xe also dissolves through the parenchymal tissue into the blood and is then transported to distal organs such as the brain. In both human and rat brains, five different spectral dissolved-phase peaks have been identified corresponding to xenon dissolved in the red blood cells (RBC), blood plasma/cerebral spinal fluid, grey matter, white matter, and the surrounding lipids.^1,2 Conformational differences between oxy- and deoxyhaemoglobin affect the resonance frequency of local ¹²⁹Xe atoms giving rise to an oxygenation dependence of the chemical shift for ¹²⁹Xe in RBCs.³ It has been shown that the chemical shift of in vitro human blood increases with increasing blood oxygenation^4,5 whereas the chemical shift of pulmonary rat blood decreases with increasing blood oxygenation.⁶ Measurement of the RBC chemical shift has been used to explore different pulmonary diseases. For example, the RBC chemical shift has been shown to be higher in healthy control rats compared to rats exposed to a model of bronchopulmonary dysplasia⁷ and the pulsatility of the RBC chemical shift has been shown to differentiate between healthy human subjects and subjects with idiopathic pulmonary fibrosis⁸ or nonspecific interstitial pneumonia.⁹ The purpose of this study is to extend the applicability of RBC chemical shift measurements to brain disease by measuring the effect of the fraction of inhaled oxygen (F_iO₂) on the chemical shift of ¹²⁹Xe dissolved in the rat brain.
Methods: All protocols were approved by the Animal Care Committee of the Hospital for Sick Children. MR spectra were acquired on a clinical 3T scanner (Prisma, Siemens, Erlangen, Germany) using a 1.7cm radius, 5.7cm long transmit/receive birdcage coil (Morris Instruments, Ottawa, Canada) centred on the rat brain. 5 female Sprague-Dawley rats (290 ± 30g, Taconic Biosciences, Rensselaer, NY) were ventilated for 10 minutes without xenon at an F_iO₂ of either 22% (i.e. normoxia) or 14% (i.e. hypoxia) as previously described.¹⁰ After 10 minutes, ¹²⁹Xe gas was added to the ventilation mix, while the F_iO₂ was maintained. After 10 seconds, 10 FIDs were acquired using the following parameters: α = 90°, TR = 1 s, bandwidth = 15 kHz, and 2048 spectral points. The transmit and receive frequencies were centred on the RBC chemical shift, 210ppm from the gas frequency. The averaged FIDs were filtered in the time domain with a 10Hz exponential line broadening function and fit to 6 exponentials, (5 dissolved resonances and 1 gas resonance) using an open-source MATLAB toolkit.¹¹ After Fourier transformation, the gas peak was used as the reference for estimating the RBC and the grey matter chemical shifts.
Results: Figure 1 shows a representative ¹²⁹Xe spectrum acquired in the rat brain and figure 2 demonstrates the effect of hypoxia on the RBC and grey matter chemical shifts. The RBC chemical shift increased during hypoxia (F_iO₂ = 14%) by 0.77ppm ± 0.42 (p = 0.015) while no effect was seen for the grey matter chemical shift.
Discussion: The increase in RBC chemical shift at 14% O₂ is very similar to the increase measured in the lungs at the same F_iO₂ (0.79ppm ± 0.34).⁶ As the change in ¹²⁹Xe RBC chemical shift is principally dependent on conformational changes of the haemoglobin molecule,³ it may not be surprising that the effect is independent of the organ. These results also reinforce that the relationship between oxygenation and RBC chemical shift in rats is opposite to that observed in humans, where RBC chemical shift has been observed to decrease with hypoxia.^4,5 In vivo measurement of cerebral blood oxygenation (SO₂) in rats could be an important tool for evaluating the progression and treatment of pre-clinical models of brain diseases such as stroke or brain tumors. While the results presented here provide evidence of the relationship between hypoxia and RBC chemical shift in the rat brain, it does not provide an empirical relationship between SO₂ and chemical shift. SO₂ is directly dependent on F_iO₂ but there are conflating physiological effects that create variability in the relationship between SO₂ and F_iO₂. Therefore, future experimental setup should be prescriptive of SO₂ instead of F_iO₂. Alternatively, an empirical relationship between SO₂ and chemical shift could be developed through in vitro experiments as was previously done for human blood.^4,5 An empirical equation relating SO₂ to RBC chemical shift in vivo is further complicated by local changes in bulk magnetic susceptibility that affect the regionally dispersed RBC chemical shift but not the gas reference frequency. This is evidenced by the difference in RBC chemical shift at normoxia measured in the brain (209.9ppm ± 0.48) and the lungs (210.7ppm ± 0.1).⁶ To control for this, the chemical shift should be measured regionally (using CSI or single voxel spectroscopy) and measured relative to a local reference frequency (e.g. using ¹H spectroscopy).¹² Conclusions: The RBC chemical shift increased during hypoxia (F_iO₂ = 14%) by 0.77ppm ± 0.42. This relationship is the same as previously reported for the rat lungs. This provides the groundwork for future experiments that measure the effect of disease on RBC chemical shift in the brain.

Acknowledgements

This work was supported by grant funding from NSERC Discovery Grant (RGPIN-2015-03832). Y.F. was financially supported by a Queen Elizabeth II Graduate Scholarship in Science and Technology.