N-Acetylcysteine rapidly replenishes central nervous system glutathione measured via magnetic resonance spectroscopy in human neonates with hypoxic-ischemic encephalopathy

Abstract

Persistent oxidative stress depletes reduced glutathione (GSH), an intracellular antioxidant and an important determinant of CNS injury after hypoxia ischemia. We used standard, short echo time Stimulated Echo Acquisition Mode (STEAM) to detect GSH by magnetic resonance spectroscopy (MRS) in 24 term neonates with hypoxic-ischemic encephalopathy (HIE), on day of life 5–6, after rewarming from therapeutic hypothermia. MRS demonstrated reliable, consistent GSH of 1·64 ± 0·20 mM in the basal ganglia immediately before intravenous infusion of N-acetylcysteine. N-acetylcysteine resulted in a rapid and significant GSH increase to 1·93 ± 0.23 mM within 12–30 min after completion of infusion (n = 21, p < 0.0001, paired t-test), compared with those who did not receive N-acetylcysteine (n = 3, GSH = 1.66 ± 0.06 mM and 1.64 ± 0.09 mM). In one perinatal stroke patient, GSH in the diffusion-restricted stroke area was 1.0 mM, indicating significant compromise of intracellular redox potential, which also improved after N-acetylcysteine. For comparison, GSH in healthy term neonates has been reported at 2.5 ± 0.9 mM in the thalamus. This is the first report to show persistent oxidative stress reflected in GSH during the subacute phase in neonates with HIE and rapid response to N-acetylcysteine, using a short echo MRS sequence that is available on all clinical scanners without spectral editing.

Keywords

Glutathione hypoxic-ischemic encephalopathy magnetic resonance spectroscopy N-acetylcysteine vitamin D

Introduction

Oxidative stress is a common mediator of acute and chronic neuroinflammatory diseases. Numerous animal and human research studies indicate that oxidative stress correlates with disease state, and reduction in oxidative stress is associated with better outcomes.^1–5 CNS oxidative stress is an important therapeutic target, but measures are difficult to obtain in many clinical situations, particularly involving acute CNS injury. Characteristics of an ideal biomarker of oxidative stress are one that can be obtained non-invasively and rapidly in critically ill patients, reflect real-time conditions in CNS, and respond to treatment. Validation of such a marker would be a significant advancement in the field and apply to many disease states in diverse populations.

Research studies have begun to investigate metabolic profiles obtained by magnetic resonance spectroscopy (MRS) in elucidating on-going oxidative stress in specific regions of the CNS. Standard clinically available MRS sequences include point resolved spectroscopy (PRESS) and stimulated echo acquisition mode (STEAM) sequences. Molecules which may be detected using these sequences include lactate, which reflects severe acute injury,¹ but is difficult to reliably distinguish and quantify due to co-localized lipid and macromolecular peaks. Other major metabolites obtained under standard MRS protocols are N-acetylaspartate (NAA), a known neuronal marker, and total creatine (tCr), a measure of cellular energetics that may decrease with significant neural compromise in stroke, but may not be altered in more moderate injury or in more chronic disease states. Currently, these standard MRS metabolites do not serve as sensitive MRS biomarkers with a direct link to oxidative stress.

The most abundant intracellular antioxidant, GSH, would satisfy the requirements for a CNS biomarker that is depleted both acutely and chronically under oxidative stress,⁶ is related to degree of injury,^3,8 and may respond to therapeutic intervention.^7,9,10 Intracellular GSH reduces reactive oxygen species (ROS) via glutathione (GSH) peroxidase, forming oxidized, glutathione disulfide (GSSG).¹ With ongoing oxidative stress, GSH is depleted and ROS accumulate, causing cellular and mitochondrial damage. Decreased GSH content is also complicit in other mechanisms of CNS injury, including potentiation of glutamate toxicity and fatty acid synthase (FAS)/cell death receptor activated apoptosis.^7,11 Sufficient intracellular GSH is essential for cell survival and is therefore an important therapeutic target.¹² [GSH] is reported at 2 mM in healthy adults and in term neonates in basal ganglia (BG) and thalami,^2,9,13 as quantified by special research sequences that are not currently available on clinical scanners.^9,13

Using standard STEAM sequence (echo time, TE = 20 ms) for quantifying GSH, we investigated whether short echo time STEAM has adequate sensitivity to detect a persistent [GSH] decrease in BG of term neonates with encephalopathy due to hypoxic-ischemic injury at birth (HIE). The BG exhibit high metabolic activity and are among the most vulnerable regions in term neonatal brain to injury.¹⁴ Our premise was that [GSH] quantification in the BG by MRS was technically feasible and would provide reliable and reproducible measures of acute oxidative stress after neonatal HIE. We hypothesized that [GSH] would be <2 mM in BG in HIE newborns on day of life (DOL) 5 and exhibit a rapid, measureable, and clinically significant increase after infusion of N-acetylcysteine (NAC) and 1,25 -(OH)₂Vitamin D₃ (calcitriol). NAC and vitamin D may increase GSH via complimentary mechanisms and act synergistically to mitigate oxidative stress.¹⁵ NAC restores GSH rapidly by providing the rate-limiting, cysteine precursor, and vitamin D increases GSH production more slowly with induction of enzymes involved in GSH synthesis.^15,16 We previously demonstrated that NAC and calcitriol, added to hypothermia, provided long-term neuroprotection in a neonatal animal model of severe HIE.¹⁷ Here we report our initial results that CNS GSH in vivo by MRS can be measured reliably on a clinical scanner in human neonates with HIE, using standard sequence without spectral editing, in minimal time. Furthermore, GSH is decreased on day 5 after HIE birth, indicating persistent oxidative stress, and responds rapidly to antioxidant neuroprotective therapy.

Materials and methods

Validation of [GSH] quantification using phantoms and human reproducibility

MRS quantification of GSH was validated via phantom solutions containing concentrations of GSH (0.5–5 mM [GSH]), 5 mM DTT, 10 mM choline, 25 mM creatinine in phosphate buffered saline, pH 7.1, prepared immediately prior to MRS scanning at TE 20 ms using a STEAM sequence on a 3T Siemens TIM Trio. The relationship of GSH values generated from spectral fitting with LCModel¹⁶ versus the known GSH concentrations of the phantom solutions is presented in Figure 1. To improve standard fitting algorithms provided with LCModel software, we constructed a specialized simulated basis set created with VeSPA that included all standard metabolites (i.e. PCh + GPC, Cr + PCr, NAA + NAAG, etc.) as well as GSH with a cysteinyl peak at 2.95 ppm. The spectral fitting with this basis set was significantly improved over standard basis set, shown in the right panel of Figure 2, and was used to fit all spectra acquired for GSH. Reproducibility of this STEAM TE 20 ms sequence was verified on adult human volunteers (n = 2) with repeated scans in the occipital cortex on two volunteers on different days prior to starting the study. Three scans were performed, and then the subject was removed from the magnet bore (pre-) and then sent back in (post-) for three additional scans. The resulting [GSH] values for both subjects (pre- and post-bore removal) in the occipital lobe, respectively, were 1.35 ± 0.15 mM and 1.43 ± 0.24 mM; 1.32 ± 0.03 mM and 1.46 ± 0.09 mM, comparable to Terpstra’s values in nine healthy adults using the same voxel location and MEGAPRESS sequences at TE = 68 ms.¹⁸

Figure 1.

LCModel quantification of [GSH] in phantom solutions.

Figure 2.

MR spectra using STEAM 20 ms, with peaks for GSH, NAA, Cr, and Cho noted (left). Spectral fitting with LCModel with (blue) and without (red) GSH cysteinyl peak at ∼2.95 ppm included in basis set (right).

Standard of care hypothermia treatment

All infants presumed to have suffered HIE, presented with the following criteria to qualify for cooling per standard protocol: Gestational age >34 weeks, cord or initial neonatal pH < 7.0, base deficit worse than −13, Apgar score at 5 min <5, need for continued resuscitation after 10 min, and two signs of stage 2 or 3 neonatal encephalopathy on clinical neurological examination (abnormality of tone, reflexes, state of arousal, posture, autonomic system or seizures).¹⁹ Infants were then cooled by Criticool™ blanket (Mennen Medical Corporation, Feasterville-Trevose, PA) to a rectal temperature of 33℃ for 72 h, and rewarmed at 0.2℃/h over 16 h, per standard clinical protocol.

NAC and calcitriol infusion (NVD)

This pilot study was approved by the Medical University of South Carolina’s Institutional Review Board (#31254) and conducted under the ethical principles of the Declaration of Helsinki. We obtained written informed consent from the parents of 30 neonates with moderate to severe HIE who qualified for hypothermia treatment, prior to enrollment within 6 h of HI birth. HIE neonates then received daily intravenous NAC + calcitriol infusions, as NAC 25–40 mg/kg every 12 h (Acetadote®, Cumberland Pharmaceuticals, Nashville, TN) and 1,25(OH)2Vitamin D₃ 0.03–0.1µg/kg/day (Calcitriol Injection USP, Akron Inc., Lake Forest, IL) from 6 h of life to DOL 10 or discharge. Calcitriol doses were withheld for elevated free serum (ionized) calcium levels >1.3 mmol/L, and the dosing interval increased from 12 to 24 h for NVD subjects 11–30. Both study drugs were withheld if the infant was clinically unstable and progressing to extracorporeal membrane oxygenation support. Study drug was resumed when neonates were stable on ECMO.

Magnetic resonance imaging protocol for HIE infants

More than 24 h after rewarming from hypothermia on DOL 5–6, MRI and MRS were performed as part of routine clinical care for prognostication after HIE, on a Siemens 3T Skyra system. We included a research MRS scan for GSH immediately before and after this standard of care MRI. For GSH quantification, a single voxel (20 × 20 × 20 mm³) was placed in the left BG (Figure 3); spectra were then acquired with a STEAM sequence (TR = 2000 ms; TE = 20 ms; NS = 176) with and without water suppression for a duration of 12 min with localization and shimming.

Figure 3.

Representative voxel placement in the basal ganglia of near-term and term neonates with HIE.

Study drugs were administered by intravenous push (calcitriol) and 45-min infusion (NAC) during the standard of care MRI for HIE, after which we acquired a second STEAM TE 20 ms in the BG within 12–30 min following completion of the NAC infusion. Neonates had cardiorespiratory and oximetry monitoring during the scan, and sedation with standard of care low-dose morphine (0.02 mg/kg/dose) or lorazepam (0.05 mg/kg/dose) which are given for comfort during cooling/rewarming, as needed. The principal investigator (neonatologist) and neonatal intensive care unit nurse were present during the entire scan.

LCModel data fitting

Spectra were fit with LCModel²⁰ using the simulated basis set created with VeSPA²¹ that included all standard metabolites (i.e. PCh + GPC, Cr + PCr, NAA + NAAG, etc.) as well as GSH with a cysteinyl peak at 2.95 ppm. Water normalization by means of the unsuppressed water scan allowed determination of GSH concentrations. Eddy current correction was also performed on the raw data before the final model fit. Phantom GSH scans were done to confirm proper basis set calibration (Figure 1). A representative 3T MRS in the BG shows GSH peak at 2.95 ppm and improved fit with basis set optimized for GSH (Figure 2). The investigator responsible for automated LCModel outputs was blinded to sequence timing.

Statistical analysis

A two-tailed paired t-test compared GSH changes in HIE neonates pre- and post-infusion with p < 0.05 being designated as statistically significant (MATLAB version R2016b).

Results

Clinical demographic data on the 30 HIE neonates are presented in Table 1. Infants were approximately equally divided between moderate and severe stages of encephalopathy. Infants received NAC via 45-min infusion twice daily, and Vitamin D as intravenous push, starting 9 h after birth, then daily for 10 days. Therefore, the pre-dosing scan for [GSH] on DOL 5–6 represents a trough value just before the next scheduled dose, while the 12 to 30-min scan post-dosing represents a peak [GSH] within 72–100 min of start of infusion.

Table 1.

Demographics.

	N = 30
Mean gestational age (weeks)	38.2 ± 1.5
Mean birth weight (g)	3248 ± 545
Number of males/females	20/10
pH of cord or neonatal blood	6.9 ± 0.2
Base deficit of cord or neonatal blood (mmol/L)	16.9 ± 7.5
Mean serum lactate <6 h after birth (mmol/L)	8.9 ± 4.6
Mean Apgar score at 1 min	1.4 ± 1.3
Mean Apgar score at 5 min	3.2 ± 2.0
Mean Apgar score at 10 min	5.4 ± 2.1
Number of HIE stage 2/3	14/16

HIE: hypoxic-ischemic encephalopathy.

Four infants were not able to be scanned due to unstable oxygenation with persistent pulmonary hypertension (NVD 4, 6, 18, 22); scans for two infants suffered inadvertent technical errors when the non-water suppressed scan was omitted, prohibiting [GSH] quantification (NVD 1), and the scan was delayed outside of dosing window due to urgent clinical cases (NVD 27). In three infants, NAC infusion was interrupted for an extended period or not completed prior to the second scan due to intravenous line or infusion pump malfunction (NVD 3, 12, 24), and these scans are noted separately in Figure 4(b). For NVD 3, pump malfunction was remedied with rapid intravenous infusion and a third scan was obtained 18 min after NAC infusion. GSH concentrations for the remaining 22 infants are presented in Figure 4(a), including one who only had an adequate post-infusion scan (NVD 8). One infant also had perinatal stroke in the left parietal area, and an additional scan was performed with a voxel in the diffusion restricted stroke area.

Figure 4.

(a) [GSH] in the basal ganglia (except as noted) immediately before and after NAC* alone or NAC + calcitriol in HIE neonates on DOL 5–6 (n = 21). NVD 14, 17, 29 had NAC infusions paused due to pump malfunction, but did receive NAC prior to second MRS. An additional ROI was obtained in NVD 26 with voxel placement in left parietal stroke area, which was adjacent to the BG voxel placement. (b) GSH concentration in basal ganglia of HIE neonates, on DOL 5–6, for whom infusion of *NAC alone or NAC + calcitriol could not be completed prior to second MRS. Little change between pre- and post-infusion [GSH] values is noted. Mean and SD of individual GSH concentrations before and after NAC + calcitriol/NAC infusion are noted with black diamonds and error bars.

Using short TE STEAM MRS sequence with LCModel quantification, we found reliable and consistent GSH concentrations in the BG before the 10–11th study dose on DOL 5–6, with Cramer Rao lower bounds <10% in all patients. All 24 HIE infants with adequate pre-infusion scans had [GSH] ≤2.0 mM in the BG, with a mean (SD) of 1.64 ± 0.20 mM (range 1.2–2.0 mM, 95%CI: 1.51, 1.71). In the three infants without NAC infusion between the first and second scans (NVD 3, 12, 24; Figure 4(b)), little to no change in [GSH] is evident in scans 83–92 min apart. This time period is similar to subjects who received the NAC infusion prior to the second scan.

There was a significant increase in BG [GSH] within 12–30 min after NAC infusion, with or without vitamin D, for those who received the full NAC dose prior to second/third MRS (Figure 4(a), n = 21, p < 0.0001, by two-tailed paired t-test, excluding NVD 8 who had an inadequate pre-scan). Mean absolute [GSH] increase was 0.30 ± 0.21 mM, or 18%, to 1.93 ± 0.23 mM (95%CI: 1.83, 2.02). Using intent to treat, including all infants with infusion problems, there was still a significant increase in pre- to post-dosing [GSH] (mean 1.64 ± 0.20 mM pre-dosing, 1.90 ± 0.22 mM post-dosing, n = 24, p < 0.0001). Only three out of the 21 HIE neonates (14%) with paired data did not show an increase in [GSH] within the time frame of the MR studies, and all had pump infusion problems (Figure 4(a)). The uniformity of [GSH] response to the relatively low doses of NAC is evident with 86% of HIE infants showing increased [GSH] post-dosing.

The range of pre-dosing [GSH] demonstrates the expected variability based on injury severity. The parietal cortical stroke area in NVD 26 has the lowest [GSH] of all spectra at 1.0 mM before dosing (Figures 4(a) and (5)). [GSH] increases by 13.5% after NVD infusion in this diffusion-restricted area (Figure 5). With GSH concentration approximately half normal values for a term neonate in white or grey matter,² our data support the likelihood of severely compromised ability to handle oxidative stress and or cell death in the stroke area. At the same time, as seen in Figure 5, the voxel placement in the BG immediately adjacent to the stroke area in NVD 26 demonstrated [GSH] of 1.64 mM pre-dosing and 2.3 mM post-dosing, respectively.

Figure 5.

Diffusion-weighted images showing ADC (a) and trace (b) parametric maps with clear stroke lesion delineation, (c) voxel placement in left parietal stroke area.

We also determined if calcitriol had any acute effect on the measured GSH increase before and after infusion. We compared GSH changes in infants in whom NAC alone or NAC plus calcitriol were administered. We found no difference in GSH changes between the NAC only (n = 9) and NAC plus calcitriol (n = 14) groups. The values for the NAC only pre- and post-infusion were 1.61 ± 0.21 mM and 1.86 ± 0.21 mM; for the NAC plus calcitriol group, pre- and post-infusion values were 1.62 ± 0.27 mM and 1.90 ± 0.32 mM, respectively.

Discussion

This is the first report of pre- and post-NAC response of GSH as a biomarker of intracellular oxidative stress in neonatal HIE using reliable MRS quantification via sequence available on all clinical scanners. We demonstrated depletion of GSH in a region of the neonatal CNS that is highly susceptible to oxidative stress in the subacute phase after hypoxic-ischemic injury. HIE results from global oxygen deprivation, leading to CNS and multi-organ system injury, and long-term cognitive and/or motor deficits, such as cerebral palsy. Hypothermia to 33℃ rectal temperature for 72 h is standard of care; however, up to 50% of HIE neonates still have significant developmental compromise.²² To combat persistent oxidative stress in spite of hypothermia treatment, we have investigated NAC and Vitamin D in animal studies,^17,23 and in this study of HIE neonates on DOL 5–6, after rewarming. Our robust in vivo data show that low-dose NAC infusion results in a rapid, significant increase in BG [GSH], substantiating GSH as a responsive biomarker of injury and therapeutic effect. Our findings also agree with preliminary data in three Gaucher’s and Parkinson’s patients,⁹ and the known mechanism of action: NAC provides rate-limiting cysteine for GSH synthesis, which in turn increases neuroprotection by eliminating ROS in the damaged tissue. Due to infusion problems, infants in the study who did not complete the NAC infusion showed no change in [GSH], in contrast to those who completed the drug infusion without problems. In addition, in the stroke area adjacent to the BG in one patient, [GSH] at 1 mM was much lower than previously reported in either white or grey matter in term neonates on a 1.5T MRI.² Thus, MRS provides a direct, quantitative biomarker of oxidative stress in the subacute phase in neonates after HIE.

Administration of NAC alone increased GSH to the same extent as NAC plus calcitriol in our HIE neonates. 1,25(OH)₂D is not a direct ROS scavenger and is thought to increase GSH via the slower process of inducing expression of synthetic enzymes, cysteine glutamate ligase and GSH reductase.^15,16,24 Therefore, we would not expect scans performed before and 72–100 min after the start of NAC plus calcitriol infusion to show significant differences in GSH compared with NAC alone. Rather, we postulate that calcitriol might work in a more sustained, supportive role for GSH synthesis, which we did not specifically test in this study.

Hypothermia itself decreases oxidative stress, but in HIE neonates who do not improve with hypothermia, pre-existing neuroinflammation and persistent oxidative stress is thought to play a role in the incomplete neuroprotection.^25–27 In addition, the increased metabolic activity and consumption of oxygen associated with rewarming should result in re-emergence or augmentation of oxidative stress, if antioxidants remain depleted and mitochondrial pathways are not fully restored. The GSH concentrations immediately before NAC or NAC plus calcitriol infusion should reflect the individual patient’s state of oxidative stress, including the effects of rewarming.

The cysteinyl-GSH MRS peak at 2.95 ppm has been quantified and validated in phantoms, animals and humans in research scans¹ (normal [GSH] = 1.97 ± 0.37 mM in parietal cortex in adults on a 3T MR with MEGAPRESS TE = 80 ms²⁸; 1.6 ± 0.2 mM in the precuneus in healthy adults measured on a 4T MR, with MEGAPRESS TE = 68 ms²⁹; 2.69 ± 0.13 µmol/g and 2.16 ± 0.23 µmol/g anterior and posterior cingulate in healthy adults¹³; and most pertinent to our study, 2.5 ± 0.8 mM in BG healthy neonates at term,² primarily measured with editing and subtraction sequences such as MEGA-PRESS, or ultra-short echo times such as phase rotation 6 ms STEAM, which are not available on most clinical scanners.^13,30–32 The use of STEAM allows shorter echo time sampling than with the common MRS technique of PRESS and enhances signal from small molecules in lower concentrations. STEAM, however, carries a signal-to-noise ratio penalty of a factor of 2 compared with PRESS sequences. Unlike other methods of quantification of reduced intracellular GSH in the CNS, the short echo STEAM sequence used in this study has the advantage of not requiring extensive post-processing spectra editing to obtain a reliable measure of GSH. Some investigators have deemed spectral editing too complicated for routine clinical use,³⁰ particularly in critically ill patients in whom timely information is essential to treatment. Other investigators used spectral editing to quantify GSH in acute stroke patients using higher TE of 131 ms,³² but found no asymmetry between stroke area and contralateral hemisphere.³¹ These data may be explained by poor resolution of MRS at long echo times for small molecules, such as GSH. In contrast, the STEAM sequence at 20 ms TE in HIE neonates is extremely sensitive, reproducible, and responsive to treatment, even in a diffusion-restricted area of stroke. Furthermore, the STEAM sequence used in this study is available on all clinical 3T scanners and requires only 12 min including shimming and localization, which is consistent with other fast MRI protocols. As such, this GSH scan could be adapted to many clinical situations of acute and chronic stroke or neuroinflammation to measure oxidative stress.

Our data verify one of the key aims of neuroprotective trials, that the drug has crossed the blood–brain barrier and had an effect on the target molecule. Therefore, with this report, we have illustrated that the effect of therapies on intracellular CNS redox status can be measured non-invasively using MRS in clinical situations. Other metabolites such as NAA, glutamate + glutamine, choline and creatine may also be quantified at short echo times. An indicator of poor tissue perfusion and anaerobic metabolism, lactate is more difficult to quantify due to overlapping lipid and macromolecular peaks. In addition, serum and CNS lactate increase to normal values within 6–12 h in many clinical HIE cases, such that after reperfusion, lactate remains markedly elevated only in the most severe cases. Therefore, quantification of CNS GSH may be more sensitive than lactate to mild to moderate oxidative stress and provide valuable diagnostic and therapeutic monitoring of cellular and mitochondrial recovery after hypoxia-ischemia.

In spite of promising preclinical data, translation of antioxidant therapies in neuroinflammatory diseases such as stroke and Alzheimer’s disease has been hampered by lack of in vivo markers of therapeutic drug effect, and inability to measure and adjust for the basal redox status of trial participants.¹⁰ Our data show variation in the basal or trough level of GSH. This variation may be a crucial factor in absolute response to treatment, if the goal is to reach a normal concentration of GSH. This hints at one of the most powerful applications of MRS-based GSH quantification as a CNS marker of oxidative stress, that of enabling dose adjustments during treatment against a benchmark GSH concentration. Therapy may then be based on CNS pharmacodynamics of oxidative stress (GSH) instead of peripheral oxidative stress markers. Moreover, a non-invasive intracellular oxidative stress measurement would afford a biomarker useful in both clinical trials and clinical practice, leading to immediate translation and generalizability of findings, if standard MRS sequences are used.

The use of MRS for GSH measurement using STEAM fulfills the pragmatic clinical requirements for rapid assessment with a sequence that is standard on all clinical scanners. At this time, the research community and manufacturers are working to refine sensitive methods for quantification of small, less abundant molecules such as GSH and GABA, which are critical for understanding disease processes. Off-line processing is straightforward with LCModel software and our specialized basis set. To facilitate the clinical translation of GSH as valuable biomarker of oxidative stress, we will make our basis set optimized for GSH quantification available via a GitHub repository link. Methods that utilize as much on-line acquisition and processing as possible ultimately will be important for clinical implementation. Potentially, a GSH optimized basis set may be incorporated into clinical scanner software to provide immediate quantification of GSH. Our results support further development of GSH processing, so that accurate measurements from MRI embedded software programs are available on-line. We demonstrate that [GSH] responds rapidly to antioxidant treatment with NAC, confirming that NAC crosses the blood–brain barrier rapidly, even in the subacute phase of HI injury, when the blood–brain barrier may have greater integrity than in the immediate post-asphyxia phase. This is the first report of pre- and post-NAC dosing response of GSH as a biomarker of intracellular oxidative stress using reliable MRS quantification with a standard STEAM sequence available on all clinical scanners.

Innovation

Biomarkers of oxidative stress are not currently available clinically in patients with acute CNS injury. We report that GSH may be measured rapidly, reliably and repeatedly in critically ill patients and demonstrate both the degree of oxidative stress and immediate response to therapy in susceptible regions of the CNS. Repeated scans around drug administration can inform pharmacodynamics of CNS effects, provide novel therapeutic guidance linking drug dosing choices to a specific target, and relate therapies to mechanisms of action in mitigating oxidative stress.

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Medical University of South Carolina Neuroscience Institute grant.

Acknowledgements

We wish to acknowledge Donnie Beason and James Purl for their dedication and expertise in performing all of the MR scans.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ contributions

HGM and TRB constructed the specialized basis set for GSH determination, tested reliability and reproducibility, analyzed phantom data, designed the MRS protocol, verified quality of spectra, performed blinded LCModel analysis and interpretation of MRS data. DDJ and DBW designed and executed the study, supervised administration of NAC and calcitriol in HIE neonates. DJ consented and enrolled all neonates shortly after HIE birth, attended all patients in MRI, monitored adverse events and communicated with the IRB. DJ and HM wrote the manuscript draft, and TB an DBW edited the manuscript.

References

Rae

Williams

. Glutathione in the human brain: review of its roles and measurement by magnetic resonance spectroscopy. Anal Biochem 2017; 529: 127–143.

Kreis

Hofmann

Kuhlmann

et al.

Brain metabolite composition during early human brain development as measured by quantitative in vivo 1H magnetic resonance spectroscopy. Magn Reson Med 2002; 48: 949–958.

Filibian

Frasca

Maggioni

et al.

In vivo imaging of glia activation using 1H-magnetic resonance spectroscopy to detect putative biomarkers of tissue epileptogenicity. Epilepsia 2012; 53: 1907–1916.

Monin

Baumann

Griffa

et al.

Glutathione deficit impairs myelin maturation: relevance for white matter integrity in schizophrenia patients. Mol Psychiatry 2015; 20: 827–838.

Srinivasan

Ratiney

Hammond-Rosenbluth

et al.

MR spectroscopic imaging of glutathione in the white and gray matter at 7 T with an application to multiple sclerosis. Magn Reson Imaging 2010; 28: 163–170.

Fan

Chen

et al.

Hypoxia-ischemic encephalopathy in full-term neonate: correlation proton MR spectroscopy with MR imaging. Eur J Radiol 2003; 45: 91–98.

Choy

Dean

Berk

et al.

Effects of N-acetyl-cysteine treatment on glutathione depletion and a short-term spatial memory deficit in 2-cyclohexene-1-one-treated rats. Eur J Pharmacol 2010; 649: 224–228.

Franco

Panayiotidis

Cidlowski

. Glutathione depletion is necessary for apoptosis in lymphoid cells independent of reactive oxygen species formation. J Biol Chem 2007; 282: 30452–30465.

Holmay

Terpstra

Coles

et al.

N-Acetylcysteine boosts brain and blood glutathione in Gaucher and Parkinson diseases. Clin Neuropharmacol 2013; 36: 103–106.

10.

Pocernich

Butterfield

. Elevation of glutathione as a therapeutic strategy in Alzheimer disease. Biochim Biophys Acta 2012; 1822: 625–630.

11.

Ibi

Sawada

Kume

et al.

Depletion of intracellular glutathione increases susceptibility to nitric oxide in mesencephalic dopaminergic neurons. J Neurochem 1999; 73: 1696–1703.

12.

Franco

Cidlowski

. Apoptosis and glutathione: beyond an antioxidant. Cell Death Differ 2009; 16: 1303–1314.

13.

Wijtenburg

Gaston

Spieker

et al.

Reproducibility of phase rotation STEAM at 3T: focus on glutathione. Magn Reson Med 2014; 72: 603–609.

14.

Johnston

Nakajima

Hagberg

. Mechanisms of hypoxic neurodegeneration in the developing brain. Neuroscientist 2002; 8: 212–220.

15.

Garcion

Sindji

Leblondel

et al.

1,25-dihydroxyvitamin D3 regulates the synthesis of gamma-glutamyl transpeptidase and glutathione levels in rat primary astrocytes. J Neurochem 1999; 73: 859–866.

16.

Jain

Micinski

. Vitamin D upregulates glutamate cysteine ligase and glutathione reductase, and GSH formation, and decreases ROS and MCP-1 and IL-8 secretion in high-glucose exposed U937 monocytes. Biochem Biophys Res Commun 2013; 437: 7–11.

17.

Lowe

Fraser

Rollins

et al.

Vitamin D improves functional outcomes in neonatal hypoxic ischemic male rats treated with N-acetylcysteine and hypothermia. Neuropharmacology 2017; 123: 186–200.

18.

Terpstra

Henry

Gruetter

. Measurement of reduced glutathione (GSH) in human brain using LCModel analysis of difference-edited spectra. Magn Reson Med 2003; 50: 19–23.

19.

Eicher

Wagner

Katikaneni

et al.

Moderate hypothermia in neonatal encephalopathy: efficacy outcomes. Pediatr Neurol 2005; 32: 11–17.

20.

Provencher

. Automatic quantitation of localized in vivo 1H spectra with LCModel. NMR Biomed 2001; 14: 260–264.

21.

Soher

Semanchuk

Todd

et al.

VeSPA: Integrated applications for RF pulse design, spectral simulation and MRS data analysis. Proc Intl Soc Mag Reson Med 2011; 19: 1410.

22.

Edwards

Brocklehurst

Gunn

et al.

Neurological outcomes at 18 months of age after moderate hypothermia for perinatal hypoxic ischaemic encephalopathy: synthesis and meta-analysis of trial data. BMJ 2010; 340: c363.

23.

Nie

Lowe

Rollins

et al.

Sex-specific effects of N-acetylcysteine in neonatal rats treated with hypothermia after severe hypoxia-ischemia. Neurosci Res 2016; 108: 24–33.

24.

Ibi

Sawada

Nakanishi

et al.

Protective effects of 1 alpha,25-(OH)(2)D(3) against the neurotoxicity of glutamate and reactive oxygen species in mesencephalic culture. Neuropharmacology 2001; 40: 761–771.

25.

Osredkar

Thoresen

Maes

et al.

Hypothermia is not neuroprotective after infection-sensitized neonatal hypoxic-ischemic brain injury. Resuscitation 2014; 85: 567–572.

26.

Osredkar

Sabir

Falck

et al.

Hypothermia does not reverse cellular responses caused by lipopolysaccharide in neonatal hypoxic-ischaemic brain injury. Dev Neurosci 2015; 37: 390–397.

27.

Jacobs

Berg

Hunt

et al.

Cooling for newborns with hypoxic ischaemic encephalopathy. Cochrane Database Syst Rev 2017; October: CD003311.

28.

Satoh

Yoshioka

. Contribution of reduced and oxidized glutathione to signals detected by magnetic resonance spectroscopy as indicators of local brain redox state. Neurosci Res 2006; 55: 34–39.

29.

Kaiser

Marjanska

Matson

et al.

(1)H MRS detection of glycine residue of reduced glutathione in vivo. J Magn Reson 2010; 202: 259–266.

30.

Matsuzawa

Hashimoto

. Magnetic resonance spectroscopy study of the antioxidant defense system in schizophrenia. Antioxid Redox Signal 2011; 15: 2057–2065.

31.

Dani

Shen

et al.

Pilot results of in vivo brain glutathione measurements in stroke patients. J Cereb Blood Flow Metab 2012; 32: 2118–2121.

32.

Zhang

Thomasson

et al.

Measurement of glutathione in normal volunteers and stroke patients at 3T using J-difference spectroscopy with minimized subtraction errors. J Magn Reson Imag 2009; 30: 263–270.