Influence of DNP Polarizing Agents on Biochemical Processes: TEMPOL in Transient Ischemic Stroke

Hyperpolarization of 13C by dissolution dynamic nuclear polarization (dDNP) boosts the sensitivity of magnetic resonance spectroscopy (MRS), making possible the monitoring in vivo and in real time of the biochemical reactions of exogenously infused 13C-labeled metabolic tracers. The preparation of a hyperpolarized substrate requires the use of free radicals as polarizing agents. Although added at very low doses, these radicals are not biologically inert. Here, we demonstrate that the presence of the nitroxyl radical TEMPOL influences significantly the cerebral metabolic readouts of a hyperpolarized [1-13C] lactate bolus injection in a mouse model of ischemic stroke with reperfusion. Thus, the choice of the polarizing agent in the design of dDNP hyperpolarized MRS experiments is of great importance and should be taken into account to prevent or to consider significant effects that could act as confounding factors.

H yperpolarization of 13 C in labeled small molecules temporarily increases spin polarization by up to 5 orders of magnitude.Among the different hyperpolarization techniques designed to enhance nuclear magnetization, 1 dynamic nuclear polarization (DNP) is the most versatile one.The DNP process involves the transfer of polarization from unpaired electron spins to neighboring nuclear spins through a dipolar interaction, making it possible to hyperpolarize a large variety of substrates.In a typical DNP sample, unpaired electron spins (also called polarizing agents) in the form of free radicals are homogeneously mixed with the labeled substrate of interest as frozen glassy beads.In practice, only tens of millimolar of polarizing agent are needed to hyperpolarize molars of substrate.A large variety of radicals has been proposed since the introduction of DNP, 2 including nonpersistent UV-induced radicals 3,4 and stable free radicals, with the latter being widely used in biological investigations.
In fact, the choice of polarizing agent has a great influence on the maximal polarization levels that can be achieved, as it influences the mechanism by which the DNP process occurs.As a rule, when directly polarizing low gyromagnetic ratio nuclei (such as 13 C), narrow electron spin resonance (ESR) line radicals (e.g., trityl or BDPA) are preferable, and high polarization levels can be reached within several hours.On the other hand, broad ESR line radicals (e.g., nitroxyl) usually provide better performances for high gyromagnetic ratio nuclei and especially for 1 H.Nevertheless, recent efforts for optimizing and accelerating the preparation of hyperpolarized (HP) samples while using the ubiquitous and accessible nitroxyl radicals showed that the use of microwave modulation with or without the combination of cross-polarization schemes may lead to very competitive polarization even on 13 C-labeled compounds. 5he advent of the dissolution dynamic nuclear polarization (dDNP) protocol allowed for the preparation of biocompatible solutions of HP metabolic substrates.Consequently, it paved the way for novel magnetic resonance spectroscopy (MRS) applications, by enabling in vivo acquisitions of biochemical reactions of exogenously infused solutions of HP 13 C-labeled tracers in real time.
−8 Early on, proton MRS was used to report the evolution of the neurochemical profile of endogenous metabolites like lactate after ischemic brain injury. 9Interestingly, while it is well established that large amounts of lactate are produced in brain areas subjected to hypoxia due to reduced blood supply, 10,11 it has been shown that exogenous lactate administration at reperfusion protects against ischemia-induced cell death and disability. 12,13Thus, in the context of advancing hyperpolarized MR neuroimaging and, in particular, the development of theranostic probes for ischemic stroke, we focused, as a first step, on the implementation of HP [1- 13 C] lactate as a probe for interrogating cerebral metabolism. 14,15We demonstrated that after transient hypoxia-ischemia injury, hyperpolarized [1- 13 C] lactate administered at a beneficial dose 12,13 rapidly reaches the brain and gets converted into pyruvate and CO 2 . 15t is important to highlight that in that study the polarizing agent of choice was the widely available and affordable nitroxyl radical 4-hydroxy-2,2,6,6-tetramethylpiperidine-1-oxyl (TEM-POL).Although small amounts of TEMPOL are necessary for the DNP process and its millimolar concentration is decreased upon the dilution with the dissolution solvent, even at these low doses, it may not be biologically inactive.Indeed, TEMPOL is a radical scavenger whose antioxidant activity has been used for a long time to prevent the adverse consequences of oxidative stress and inflammation in a number of pathological conditions (see Wilcox et al. for a review 16 ), which have recently incorporated SARS-and MERS-Co-Vs infections. 17−22 Hence, the aim of the present work was to investigate whether the presence of TEMPOL influences the cerebral metabolic readouts of an HP [1- 13 C] lactate bolus injection in a mouse model of ischemic stroke with reperfusion.In the dDNP protocol, the nitroxyl radical was replaced as a polarizing agent with the trityl radical OX063, which has not been reported for neuroprotection and produces higher 13 C polarization 23 in metabolic precursors.Then, the metabolic readouts of the injection of HP [1- 13 C] lactate polarized with the trityl radical were compared to that of a similar bolus spiked with TEMPOL.
The transient middle cerebral artery occlusion (MCAO) stroke procedure induces a focal ischemic lesion in the striatum, as visible in the hypersignal in representative T 2 W axial images (Figure 1A−B).Although it is only very slightly visible at 1 h post-reperfusion, the time point at which HP lactate is injected, the lesion boundaries become substantially more contrasted at 2 h post-reperfusion.
DNP requires the use of free radicals as polarizing agents.While they are filtered out in clinical investigations, this is  generally not the case in preclinical studies.Despite their low concentration in the sample after dissolution, they could still potentially interfere with the biochemical processes that are being probed.Here we show that the in vivo administration of the nitroxyl radical TEMPOL, even at the low dose used as polarizing agent, significantly altered the cerebral metabolic response to an HP [1- 13 C] lactate bolus injection following transient hypoxia-ischemia.As expected, immediately after the infusion of HP [1- 13 C] lactate at 1 h post-reperfusion, [1- 13 C] pyruvate, [1- 13 C] alanine, and [ 13 C] bicarbonate were detected in the brain compartment (Figure 1 and Figure 2).However, the metabolite ratios, computed from the sum of the MRS signals acquired in the first 120 s post-infusion, reported distinct outputs between mice receiving HP [1- 13 C] lactate polarized with trityl radical after MCAO (to be referred to as the MCAO group) and mice receiving the bolus of HP [1- 13 C] lactate with the addition of TEMPOL after MCAO (to be referred to as the MCAO+TEMPOL group, Figure 3).While the pyruvate-to-lactate ratio (PLR, Figure 3A) of the MCAO group ((7.0 ± 1.4) × 10 −3 ) was lower than that of the group  which received the co-injection of TEMPOL ((10.7 ± 2.1) × 10 −3 ), the alanine-to-lactate ratio (ALR, Figure 3B) was lower in the MCAO+TEMPOL group ((5.3 ± 1.0) × 10 −3 ) compared to the MCAO group ((8.2 ± 1.7) × 10 −3 ).The alanine-to-pyruvate ratio (APR, Figure 3D) was significantly lower in the MCAO+TEMPOL group (0.51 ± 0.11) than in the MCAO group (1.19 ± 0.20).Finally, no changes were observed in the lactate-to-bicarbonate conversion (BLR, Figure 3C).By polarizing 13 C using OX063 we were able to double the initial polarization compared to our previous results with TEMPOL. 15The increased sensitivity enabled an accurate quantification of the 13 C-bicarbonate signals.
The blood flow deficit during ischemic stroke results in oxygen and glucose depletion in the affected brain region, giving rise to excitotoxicity and ionic imbalance, nitrosative and oxidative stress, as well as apoptotic cell death. 24The generation of toxic free radicals is even exacerbated after the restoration of blood flow, in what is known as reperfusion injury.During cerebral ischemia, superoxide anion (O 2 •− ), which is mainly generated in mitochondria as a result of oneelectron reduction of oxygen, is produced at such high levels that the ability of the natural scavenger enzyme superoxide dismutase (SOD) to dispose of it is overwhelmed.TEMPOL is a cell-and BBB-permeable compound with SOD-mimic activity that can react successively with hydroperoxyl and superoxide radicals to decompose them into H 2 O 2 and O 2 , while consuming H + , acting as a self-replenishing antioxidant agent 16,25,26 (Figure 4A).Our results highlight different PLRs between both groups of MCAO animals, being higher in those that were co-injected with TEMPOL (Figure 3A).This difference could be related to the consumption of H + in the TEMPOL-mediated decomposition of reactive oxygen species, which could indirectly enhance the conversion of lactate into pyruvate by favoring a more efficient uptake of the exogenous HP [1- 13 C] lactate via the lactate/H + monocarboxylate transporter (MCT) import and/or by favoring a displacement of the lactate dehydrogenase (LDH) equilibrium toward the production of [1- 13 C] pyruvate, NADH, and H + (Figure 2, Figure 4B).The dose of TEMPOL used in our experiments (3.8 mg/kg) is about 5 to 50 times lower than that reported to protect at reperfusion 18,19,22 or in other pathological conditions. 16,17,27Nevertheless, since TEMPOL antioxidant properties stem from a catalytic reaction, small amounts of the nitroxyl radical could be sufficient to appreciate a biological effect.In contrast, trityl radicals are typically more stable 28 due to the electron delocalization conferred by their molecular structure.Although they react with the superoxide radical, 29 the low dosage and noncatalytic reaction are unlikely to substantially affect the experiment outcome.
Earlier experiments using HP [1- 13 C] pyruvate to study cerebral metabolism reported that [1- 13 C] alanine rather originates from peripheral tissues than from the brain. 30,31It is probable that the change in ALR after stroke is related to muscles around the skull, close to the coil, whose metabolism was affected by the common carotid artery ligature during surgery.However, the cerebral concentration of alanine has been reported to be steadily elevated during ischemia, an effect attenuated by TEMPOL. 21In the same study, TEMPOL decreased extracellular glutamate release, reduced the ischemic lesion size, and improved neurobehavioral outcomes.In the ischemic tissue, neural cells respond to oxygen and glucose deficits by rapidly depolarizing and massively releasing glutamate.This glutamate can be taken up by astrocytes, where it can be converted into alanine by alanine aminotransferase (ALT) using pyruvate as a cosubstrate. 32The conversion of pyruvate to alanine by ALT requiring glutamate (Figure 4C), the decrease that we observed in [1- 13 C] alanine labeling (Figure 3B and D), could thus be related to the effect of TEMPOL in decreasing the availability of glutamate to be consumed for the transamination reaction of [1-13 C] pyruvate.
In conclusion, the administration of TEMPOL at the dose commonly used as a polarizing agent for DNP results in a significantly different cerebral metabolic response to HP [1- 13 C] lactate following transient ischemic stroke on the time scale of hyperpolarized MR examination.Our results highlight that the boost in sensitivity afforded by hyperpolarized 13 C MRS made feasible the detection of the metabolic interference of TEMPOL.Additionally, they show that care should be taken when choosing the polarizing agent in DNP hyperpolarization experiments, as certain biologically active reagents like TEMPOL can meddle with the biochemical processes of interest.Even if used in small doses, they may have significant effects that could act as confounding factors.Not only could this be important for in vivo preclinical research as shown here, but it could also have some impact on new, recently developed uses of DNP HP techniques like 13 C hyperpolarization via cross-polarization from DNP-polarized 1 H spins, which is being applied to metabolomics studies and which so far prefers TEMPOL to achieve this cross-polarization. 33METHODS Animal Experimentation.All experiments involving mice were conducted according to federal and local ethical guidelines and were approved by the local regulatory authorities (Service de la Consommation et des Affaires Veteŕinaires, Canton de Vaud, Switzerland), with license numbers VD2017.5 and VD2017.6.Male C57BL/6J mice (6 to 10 weeks, Charles River, France) were maintained in an animal facility with controlled humidity and temperature, a 12 h light/dark cycle, and free access to food and water.
Transient Middle Cerebral Artery Occlusion (MCAO) Model of Stroke.A lesion in the left striatum was induced by transient 30 min focal cerebral ischemia as previously described. 34In summary, mice were kept under anesthesia with 1.5−2.0%isoflurane in 60% oxygen, and laser-Doppler flowmetry was used to monitor the regional cerebral blood flow (rCBF) through a flexible probe (Perimed AB, Sweden) glued to the skull at 1 mm posterior and 6 mm lateral from the bregma.The neck was incised, and both left common and external carotid arteries were exposed and ligated.A silicone-coated nylon monofilament (Doccol Corp., Sharon, USA) was inserted through the common carotid artery into the internal carotid artery to obstruct the left middle cerebral artery (MCA).The occluding filament was removed after 30 min to restore the blood flow.The intervention was considered successful if the rCBF remained below 20% of the baseline during occlusion and increased above 50% of the initial value within 10 min after the filament retraction.The left femoral vein was cannulated during occlusion to allow intravenous injection of the HP solution.
Magnetic Resonance Measurements.MR measurements were performed on a 9.4T/31 cm horizontal actively shielded magnet (Magnex Scientific, Abingdon, UK) connected to a Varian INOVA spectrometer (Varian, Palo Alto, USA).
Upon reperfusion, mice were transferred into the MRI scanner with a home-built 1 H quadrature/ 13 C linear surface coil above the head, whose sensitivity profile was described in a previous study. 15Using the FASTESTMAP routine, 36 static field inhomogeneity was corrected in a 3.6 mm × 6.9 mm × 4.5 mm voxel within the brain to optimize the signal quality.
Anatomical axial T 2 weighted (T 2 W) images were acquired with a fast spin−echo sequence (effective echo time TE eff = 52 ms, TR = 4000 ms, 18 mm × 9 mm FOV, 256 × 128 matrix) at the beginning of the MR scan to provide localization for the shimming voxels, as well as within 5 min of the HP injection and at 2 h post-reperfusion to assess the evolution of the striatal lesion.
At 1 h post-reperfusion, the lactate sample was dissolved in superheated D 2 O, pushed to a separator/infusion pump, and injected through the automated protocol. 35The injection volume was set to 450 μL, including 125 μL of dead volume, to reach a therapeutic dose of HP [1-13 C] lactate (1.07 ± 0.14) μmol/g or (121 ± 16) mg/kg of sodium lactate.Immediately, global 13 C HP MRS was triggered and acquired every 3 s with 30°BIR-4 adiabatic pulses.The spatial localization was provided by the coil's sensitivity profile. 15nimal Groups.Two animal groups were scanned: MCAO (n = 5) and MCAO+TEMPOL (n = 5).In the latter, a dose of 22 nmol/g or 3.8 mg/kg of 4-hydroxy-2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPOL, Sigma-Aldrich, Buchs, Switzerland) was co-injected with the HP lactate bolus by adding it to the separator/infusion pump where both substances are mixed between the injection and dissolution.The TEMPOL dose was identical as when previously used as the polarizing agent. 15etermination of the Injected Dose.250 μL of the remaining solution in the infusion pump was mixed with an equal volume of 80.0 mM [1- 13 C] acetate solution in D 2 O and doped with 1 mM of Gd-DO3A-butrol (Gadobutrol, Gadovist, Bayer AG, Zurich, Switzerland) to reduce the 13 C T 1 .1D 13 C high resolution NMR was performed on a 400 MHz NMR spectrometer (Avance NEO, Bruker BioSpin, Fallanden, Switzerland).The integrals of the [1- 13 C] lactate and [1- 13 C] acetate peaks were compared to determine the concentration of the lactate solution.

13
C MRS Data Processing.The signal from the first 120 s postinjection was summed, and then the area under the curve (AUC) of the metabolite peaks was fitted using the Bayesian Data-Analysis Software Package V4.01 (Washington University in St. Louis).The peak areas of [1-13 C] lactate, [1-13 C] alanine, [1- 13 C] pyruvate, and [ 13 C] bicarbonate were then used to compute the metabolite ratios.Across the experimental data of this study, the concentration of the lactate solution and the weight of the animals were homogeneous.
Statistical Analysis.Mann−Whitney U-test analyses were performed using Matlab R2021b (MathWorks, Natick, USA).A pvalue below 0.05 was considered statistically significant.All data are presented as the mean ± standard deviation unless otherwise stated.

■ AUTHOR INFORMATION
Corresponding Author

Figure 1 .
Figure 1.(A−B) Representative dynamic cerebral 13 C MRS acquired after a bolus infusion of [1-13 C] lactate (lb = 20 Hz).The summed signal from the first 120 s post-infusion is plotted in blue.The vertical scale was normalized to the height of the summed HP lactate peak.In both groups (A−B), the HP [1-13 C] lactate (183.5 ppm) was converted into [1-13 C] pyruvate (171.1 ppm), [1-13 C] alanine (176.7 ppm), and [ 13 C] bicarbonate (161.2 ppm).The signal observed at 177.7 ppm (*), partially overlapping with the alanine peak at 176.7 ppm, is an impurity from the stock lactate solution.Representative axial T 2 W images of the brain acquired at 1 and 2 h post-reperfusion.Images were acquired with a fast spin− echo multislice sequence (voxel size: 0.07 × 0.07 × 1 mm 3 , 4 averages).In both groups (A−B), the striatal lesion was slightly visible at 1 h postreperfusion and clearly contrasted at 2 h post-reperfusion (white arrows).

Figure 3 .
Figure 3. Metabolite ratios measured following injection of HP [1-13C] lactate.Data are displayed as the mean ± standard deviation and overlaid with individual data points (black diamonds).Pyruvate-to-lactate ratio (PLR, A), alanine-to-lactate ratio (ALR, B), bicarbonate-to-lactate ratio (BLR, C), alanine-to-pyruvate ratio (APR, D), and bicarbonate-to-pyruvate ratio (BPR, E).The PLR was significantly lower in the MCAO group compared to the MCAO+TEMPOL group (A).A lower alanine labeling (ALR) was observed in the MCAO+TEMPOL group compared to the MCAO group (B).No changes were observed in the bicarbonate-to-lactate ratio (BLR, C).

Figure 4 .
Figure 4. Schematic representations of the chemical reactions potentially involved in the interpretation of the measured 13 C NMR signals, with or without TEMPOL.