
==== Front
Transl Psychiatry
Transl Psychiatry
Translational Psychiatry
2158-3188
Nature Publishing Group UK London

39313523
3086
10.1038/s41398-024-03086-5
Article
Making bloodwork work: the impact of sample collection, processing, and storage on plasma glutathione measurement, and implications for translation
http://orcid.org/0000-0002-1622-7424
Coden Kendall M. kcoden@umich.edu

12
Nguyen Duyen K. K. 3
Moorhead Roberta 1
Stix-Brunell Beatriz E. 1
http://orcid.org/0009-0008-7954-5017
Baker Joanna N. 1
http://orcid.org/0000-0002-6836-6338
Parker Karen J. 13
http://orcid.org/0000-0002-3209-9389
Garner Joseph P. jgarner@stanford.edu

13
1 https://ror.org/00f54p054 grid.168010.e 0000 0004 1936 8956 Department of Comparative Medicine, Stanford University, Stanford, CA 94305 USA
2 https://ror.org/00jmfr291 grid.214458.e 0000 0004 1936 7347 Neuroscience Graduate Program, University of Michigan, Ann Arbor, MI 48109 USA
3 https://ror.org/00f54p054 grid.168010.e 0000 0004 1936 8956 Department of Psychiatry and Behavioral Sciences, Stanford University, Stanford, CA 94305 USA
23 9 2024
23 9 2024
2024
14 3856 9 2023
28 8 2024
2 9 2024
© The Author(s) 2024
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Psychiatry has traditionally focused on the study of neurons and neurotransmitter physiology in the pathophysiology and treatment of psychiatric disorders. A growing literature highlights REDOX imbalance (a state in which demand for antioxidants surpasses their bioavailability) as a common pathophysiology for a diverse array of brain conditions (e.g., trichotillomania, schizophrenia, autism, Parkinson’s disease). REDOX imbalance is typically measured via plasma glutathione, as glutathione is critical to the adaptive antioxidant response in the brain. Accordingly, glutathione, its precursors, and/or metabolites serve as biomarkers of disease risk, therapeutic targets, and measures of treatment response. However, as with any emerging field, there are currently several different methods for collection, processing, storage, and calculation of summary measures of plasma glutathione metabolism, within and between preclinical and clinical research. The lack of evidence-based best-practice methodology hampers reproducibility (preclinical or clinical), and translation (between preclinical and clinical work). To address this methodological need, here we used a repeated measures within-subject design to investigate how sample preparation (type of anticoagulant used during blood collection, deproteinization status, and storage temperature) affects plasma glutathione levels. Accordingly, we collected whole blood from mice (N = 13), and then, using a commercially available kit, quantified glutathione in plasma stored in four different ways. Presuming that these preparation conditions and post-processing calculations are unimportant, we would expect to see no difference in glutathione levels and summary measures from the same sample. However, we found each of these variables to significantly alter quantified glutathione levels. Accordingly, we propose a vital, gold-standard methodology for both sample collection, processing, and storage of plasma used for glutathione quantification and for summary calculations of glutathione that can be used preclinically and clinically, thus yielding more streamlined translation.

Subject terms

Psychiatric disorders
Molecular neuroscience
Biomarkers
Institutional award granted to early-stage research projects that could exponentially accelerate our understanding of human biology—and our ability to predict, prevent, and cure disease.issue-copyright-statement© Springer Nature Limited 2024
==== Body
pmcIntroduction

Radical oxygen species (ROS), which are cytotoxic in large quantities [1, 2], are produced as byproducts of general cellular function (e.g., metabolism of glucose [3]) and degradation of other chemicals in the brain, such as metabolism of the neurotransmitter dopamine [4, 5]. The brain has several mechanisms to neutralize ROS, of which antioxidant scavenging is one of the final lines of defense [6]. However, when the physiological demand for antioxidants surpasses their bioavailability, the brain enters a state of REDOX imbalance in which ROS accumulate, leading to neuronal damage [6, 7]. Accordingly, production of the endogenous antioxidant glutathione (GSH) by glial cells within the brain [8], particularly astrocytes [8–10], is a critical function, as most antioxidants are unable to cross the blood brain barrier.

GSH exists in two forms, the biologically available “free GSH”, and as a biologically inactive dimer, glutathione disulfide (GSSG). Free GSH neutralizes ROS via a REDOX reaction in which glutathione peroxidase oxidizes free GSH into GSSG [11], thereby producing a stable oxygen species and water as a byproduct. Thus, “free GSH” is the reduced, or antioxidant form, and GSSG is the oxidized form, of GSH. There is compelling evidence, including low plasma levels of GSH [12, 13], and effective antioxidant therapies (both with GSH [14] and its precursor, N-acetylcysteine [14, 15]), of REDOX imbalance contributing to the pathophysiology of several brain conditions such as obsessive-compulsive disorder, schizophrenia, bipolar disorder, autism, trichotillomania, and neurodegenerative disorders (see refs. [15–18]). Despite this, the implications of REDOX balance, and more specifically GSH, in the development and treatment of brain disorders remains largely understudied.

Given the ease in accessing blood samples, blood-based biomarkers are frequently used as measures of REDOX state (e.g., refs. [15, 19, 20]). However, the process of collecting and storing blood samples involves several decisions that can alter profiles of biological analytes, thus serving as potential confounds in analysis. To obtain plasma, typically whole blood is collected into tubes pre-coated with an anticoagulant—heparin and EDTA being the most common—and then transported to the laboratory on wet ice where the whole blood is centrifuged to fraction the sample into its component parts (red blood cells, white blood cells, and plasma). Additional steps may be taken (e.g., deproteinization and/or storage at −80 oC) to prevent a variety of processes from altering GSH levels (e.g., enzymatic degradation of GSH; or spontaneous oxidation of free GSH to GSSG) if plasma is not analyzed on the same day as collection. Furthermore, some biological analytes, including GSH [21], have a short half-life and are unstable at room temperature. Thus, the accuracy of analyte quantification can be significantly impacted by the duration of time required to collect and process samples. This is of particular importance in translational psychiatry because preclinical and clinical studies typically differ in sample collection procedures. Specifically, preclinical studies tend to batch collect and process samples, whereas clinical studies tend to collect and process samples one at a time.

Within the GSH literature, there are inconsistencies in sample collection methodologies, processing techniques, and final summary measures for data analysis. Given the potential impact of these factors on the quantification of GSH, here we examined if the type of anticoagulant used during sample collection (heparin or EDTA), deproteinization of plasma prior to storage (yes or no), storage temperature (−80 oC or 4 oC), and processing time (duration of time between collection of whole blood and deproteinization of plasma) impact levels of free GSH and GSSG, in plasma. (See Table 1 for descriptions of preparation conditions.) In addition to differences in sample collection, processing, and storage methodology, studies also differ in final summary measure reporting. In particular, some studies report the ratio of free GSH to GSSG [19, 22–26], whereas others quantify and report total GSH (i.e., free GSH + 2xGSSG) [12, 19, 24, 26–30]. These differences in reporting can also significantly alter reproducibility and translatability as two samples can have significantly different levels of total GSH, yet the same ratio of free GSH to GSSG. Therefore, we sought to assess how post-processing summaries are affected by sample preparation methods.Table 1 Plasma GSH measurement: preparation conditions and comparisons.

Plasma Preparation Conditions	
Preparation Condition	Anticoagulant used	Deproteinized prior to storage?	Storage Temperature	
1	Heparin	Yes	−80 oC	
2	EDTA	Yes	−80 oC	
3	EDTA	No	−80 oC	
4	EDTA	No	4 oC	
Hypothesis Testing Comparisons	
Preparation Condition 1

(heparin → deproteinization → −80 oC)

vs.

Preparation Condition 2

(EDTA → deproteinization → −80 oC)

	Do EDTA and heparin differentially stabilize GSH?	
Preparation Condition 2

(EDTA → deproteinization → −80 oC)

vs.

Preparation Condition 3

(EDTA → −80 oC → deproteinization)

	Does deproteinization prior to storage prevent spontaneous oxidation of free GSH in plasma?	
Preparation Condition 3

(EDTA → −80 oC → deproteinization)

vs.

Preparation Condition 4

(EDTA → 4 oC → deproteinization)

	Does storage of plasma at −80 oC prevent enzymatic activity in the sample?	
See Fig. 1 for schematic overview of sample collection and experimental conditions.

To do so, we employed a repeated measures within-subject design in which a single blood draw from each animal was processed in four different ways (see Fig. 1). Plasma levels of total GSH, free GSH, GSSG, and the ratio of free GSH to GSSG were then independently analyzed to produce 16 measures per animal. The power of this approach is that baseline levels of all GSH measures are consistent across all preparation conditions from the same animal. Thus, each animal acts as its own control, and any differences in measures obtained must be due to preparation condition. Additionally, this technique allows us to determine if the different measures (e.g., total GSH vs. ratio of free GSH to GSSG) are more or less robust in the face of differences in processing technique.Fig. 1 Schematic overview of the blood sample collection, processing, and storage methods used for subsequent GSH quantification.

Whole blood was collected from mice (N = 13) and centrifuged to obtain plasma. Plasma was then prepared in four different ways (See Table 1 for further description of preparation conditions) before levels of GSH were quantified using the DetectX Glutathione assay kit. This figure was created using Biorender.com.

Our approach enabled us to test four different hypotheses (Table 1) regarding the impact of preparation conditions on plasma GSH measures. Hypothesis 1: Heparin and EDTA equally stabilize GSH in plasma. Presuming this is true, we would expect to observe no differences in measured free GSH between the deproteinized heparinized plasma and deproteinized EDTA-treated plasma stored at −80 oC from the same subjects. However, if they differentially stabilize oxidation of free GSH to GSSG, then we would expect to see differences in the level of free GSH measured in the heparinized versus EDTA-treated plasma. Hypothesis 2: Deproteinization of samples will prevent spontaneous oxidation of free GSH. If this is true, we would expect to observe the same level of total GSH, but reduced free GSH and elevated GSSG, in the non-deproteinized EDTA-treated plasma stored at −80 oC compared to the deproteinized EDTA-treated plasma also stored at −80 oC. As a positive control for the impact of storage temperature, spontaneous oxidation of free GSH, and enzymatic processes, we stored an aliquot of non-deproteinized EDTA-treated plasma at 4 oC. Hypothesis 3: Storage temperature will significantly impact levels of total GSH, free GSH, GSSG, and the ratio of free GSH to GSSG. If there is presence of enzymatic activity and/or spontaneous oxidation in storage, we would expect to observe lower levels of total GSH and free GSH, and elevated levels of GSSG, in the EDTA-treated sample stored at 4 oC compared to the non-deproteinized EDTA-treated plasma stored at −80 oC. Hypothesis 4: Processing time will impact levels of free GSH, GSSG, and the ratio of free GSH to GSSG. If processing time does not impact sample integrity, we would expect to see no differences in levels of free GSH, GSSG, or the ratio of free GSH to GSSG as a function of processing time.

Methods

Ethics approval

All procedures were approved by the Stanford University Institutional Animal Care and Use Committee. Stanford is an AAALAC accredited institution.

Animals and housing

Following laboratory animal welfare (“3Rs”: Replacement, Reduction, and Refinement) best practice [31], we acquired 13 adult female mice (C57BL/6 background) from another lab which were excess to experimental requirements, experimentally naïve, and slated for euthanasia. Individual mice ranged from 51 to 191 days of age. For the duration of the study, animals were singly housed in standard individually ventilated cages (Innovive, San Diego, CA) with nesting material enrichment (ALPHA-Twist; Shepard Specialty Papers, Watertown, TN) in a climate and temperature-controlled room (72 ± 2 °F) on a 12 hr/12 hr light/dark cycle (lights on at 13:15). Cages were changed weekly. Food (Teklad 2018 mouse diet, Envigo, Haslett, MI) and deionized water were available ad libitum for the duration of the study. Subjects were acclimated to housing conditions for at least one week prior to sample collection.

Procedures

Blood collection and centrifugation

Animals were anesthetized with isoflurane and whole blood collected via cardiac puncture. (Animals were secondarily euthanized via cervical dislocation.) Whole blood was transferred into two separate micro-collection tubes (heparin-coated micro-collection tube: BD Microtainer item 13-680-62; EDTA-coated micro-collection tube: BD microtainer item 365972-1). Collection tubes were inverted 10x to ensure proper mixing of blood and anticoagulant and then placed on wet ice for transportation to the laboratory. Samples were then centrifuged at 15,000 rpm for 5 min at 4 oC to obtain plasma. Due to hemolysis of red blood cells during sample collection, one subject was excluded from all analyses (N = 12 final subjects).

Plasma preparation

We separated the resulting plasma into four preparation conditions that differed in anticoagulant used during sample collection, storage temperature, and time to deproteinization. (See Fig. 1 for schematic overview of sample collection and processing procedures.) The roles of these preparation conditions and their respective hypothesis-testing comparisons are detailed in Table 1.

We deproteinized the heparinized plasma by mixing it with an equal volume of cold 5% aqueous 5-Sulfosalicylic Acid (SSA; Sigma Aldrich Item s2130). The mixture was then incubated for 10 min at 4 oC (obtained by nesting the mixture in wet ice). Following the 10 min incubation period, we centrifuged the 5% SSA plasma mixture for 10 min at 14,000 rpm in a refrigerated centrifuge at 4 oC. The resulting 2.5% SSA plasma supernatant was collected into 1.5 mL microcentrifuge tubes, flash frozen on dry ice, and stored at −80 oC for future analysis.

One third of the EDTA-treated plasma was processed and stored in the same manner as the heparinized plasma—deproteinized and stored at −80 oC. An additional third of the EDTA-treated plasma was flash frozen on dry ice and stored at −80 oC without prior deproteinization but was deproteinized upon thawing. The remaining third of the EDTA-treated plasma was stored at 4 oC without prior deproteinization and was deproteinized immediately prior to analysis.

Quantification of GSH

We used the Arbor Assay DetectX Glutathione Detection Kit (#K006, Arbor Assay, Ann Arbor, MI, USA), with which we have experience (e.g., ref. [15]), to measure GSH levels within our samples. This kit, which is compatible with both heparinized and EDTA-treated samples, uniquely allows for measurement of total GSH, free GSH, and GSSG from the same well, rather than requiring the use of two plates to measure free GSH and GSSG. Due to their small body mass, mice provide small sample volumes of plasma. Thus, being able to measure all forms of GSH in one well significantly reduced the sample volume required for our experiment and mitigated inter-plate variation as a confound in data analysis. Based on prior experience with this kit, to ensure proper mixing of reagents, we placed the 96-well plate on an orbital plate shaker during incubation steps. Other than this modification, all manufacturer instructions were followed.

Based on prior experience, regardless of analyte, manufacturer, or species, we assume the sensitivity of kits will change between lot numbers. We therefore routinely pilot dilution factors prior to an experiment, even for kits we know well. Accordingly, based on such data, and manufacturer instructions, we determined that a final sample dilution of 1:5 was best for measurement of GSH within plasma. As such, plasma which was deproteinized prior to storage at −80 oC was thawed and further diluted 1:2.5 with assay buffer to produce a final dilution of 1:5 before assaying. Plasma which was not deproteinized prior to storage was removed from storage (−80 oC and 4 oC, respectively) and mixed with an equal volume of cold 5% SSA and incubated for 10 min at 4 oC (on ice). Following this incubation period, the 5% SSA plasma mixture was centrifuged at 4 oC for 10 min at 14,000 rpm. The resulting 2.5% SSA plasma supernatant was collected and mixed 1:2.5 with assay buffer to produce a final dilution of 1:5 before assaying.

Statistical analyses

Analyses were conducted using JMP16 for Windows. Repeated measures REML Mixed Models were used to reflect the design and properly compare preparation condition within each subject before amalgamating these comparisons across subjects. All GSH measures and time to deproteinization were log-transformed to meet assumptions of linear models (homogeneity of variance, normality of error, and linearity) [32]. Sample size was determined following Mead’s Rule [33].

Using individual variation in GSH combined with a repeated measures approach is a highly effective form of randomization. Sample collection was conducted blind to preparation condition and randomized by subject identification. Sample processing was conducted by one individual. Data analysis was predetermined and performed blind to expected outcomes.

Is there an effect of preparation condition on measured levels of GSH?

We used the same REML Mixed Model separately for each outcome (total GSH, free GSH, GSSG, or the ratio of free GSH to GSSG). The between-subject random effect was mouse, and the within-subject repeated measure was preparation condition. Significant results were examined with Tukey-corrected post-hoc comparisons.

Is there an effect of time to deproteinization on measured levels of GSH within storage method and/or anticoagulant type used?

According to manufacturer instructions for our assay kit, best practice is to deproteinize plasma prior to storage. Thus, to answer this question, we conducted analyses using only heparinized plasma (Preparation Condition 1) and EDTA-treated plasma (Preparation Condition 2) which were both deproteinized prior to storage at −80 oC. We used the same REML Mixed Model as above, but now included the time to deproteinization, as well as the interaction between anticoagulant and time to deproteinization. Nonsignificant interactions were retained in the model and marginal main effects tested at the mean of other variables, to give the overall effect of anticoagulant and time to deproteinization controlling for all other variables.

Results

Plasma processing condition significantly affects measured plasma levels of total GSH, free GSH, GSSG, and the ratio of free GSH to GSSG in plasma

We found significant differences in total GSH levels (F3,33 = 88.8411; P < 0.0001), free GSH levels (F3,33 = 19.4471; P < 0.0001), GSSG (F3,33 = 66.2382; P < 0.0001), and the ratio of free GSH to GSSG (F3,33 = 17.5240; P < 0.0001) across the preparation conditions (see Fig. 2). Post-hoc Tukey tests revealed patterns of differences relevant to our hypotheses, as described below.Fig. 2 Plasma preparation conditions result in differing levels of GSH.

There is no effect of anticoagulant on total GSH when samples are stored at −80 °C. Free GSH, GSSG, and the ratio of free GSH: GSSG vary as an effect of anticoagulant used during sample collection, whereas total GSH is robust in samples stored at −80 °C. All data are presented as least squares mean +/− standard error. Blue bars depict samples which were deproteinized prior to storage; orange bars depict samples which were not deproteinized prior to storage. Within each graph (graph A: total GSH; graph B: free GSH; graph C: GSSG; graph D: the ratio of free GSH to GSSG), preparation conditions with shared letters do not differ significantly (P < 0.05), whereas treatments without shared letters do significantly differ. Refer to Table 1 for descriptions of preparation conditions and hypothesis testing comparisons.

Preparation Condition 1 (heparin → deproteinization → −80 oC) vs. Preparation Condition 2 (EDTA → deproteinization → −80 oC): Heparinization and EDTA-treatment of plasma result in different levels of free GSH, GSSG, and their relative ratio, but does not affect stability of total GSH

Suggesting that total GSH is a valid measure in both heparin and EDTA-treated plasma, we found no difference in total GSH between the heparinized plasma (Preparation Condition 1) and EDTA-treated plasma (Preparation Condition 2) stored at −80 oC (Fig. 2A). However, suggesting that free GSH, GSSG, and their relative ratio is less valid than total GSH, we found the level of free GSH to be reduced (Fig. 2B) and GSSG to be elevated (Fig. 2C) resulting in a decreased relative ratio of free GSH to GSSG (Fig. 2D) in the heparinized plasma compared to the EDTA-treated plasma, both of which were deproteinized and stored at −80 oC prior to quantification.

Preparation Condition 2 (EDTA → deproteinization → −80 oC) vs. Preparation Condition 3 (EDTA → −80 oC → deproteinization): Deproteinization status does not affect measurement of GSH in EDTA-treated plasma

We found no differences in total GSH (Fig. 2A), free GSH (Fig. 2B), GSSG (Fig. 2C), or the ratio of free GSH to GSSG (Fig. 2D) between the EDTA-treated samples stored at −80 oC regardless of deproteinization status.

Preparation Condition 3 (EDTA → −80 oC → deproteinization) vs. Preparation Condition 4 (EDTA → 4 oC → deproteinization): Storage temperature affects total GSH, free GSH, and GSSG, but not the ratio of free GSH to GSSG in plasma

We included EDTA-treated plasma stored at 4 oC (Preparation Condition 4) as a positive control for enzymatic degradation of glutathione which may occur during sample storage. We found reduced levels of total GSH (Fig. 2A), free GSH (Fig. 2B), and GSSG (Fig. 2C) in the non-deproteinized EDTA-treated plasma stored 4 oC (Preparation Condition 4) compared to the non-deproteinized EDTA-treated plasma stored at −80 oC (Preparation Condition 3). However, the differences in free GSH and GSSG did not affect their relative ratio (Fig. 2D) between the preparation conditions, suggesting that storage of plasma at −80 oC is sufficient to prevent enzymatic degradation of GSH in plasma.

Time to deproteinization influences some measures of GSH with additional differences between deproteinized heparinized and EDTA-treated plasma

Noting the short half-life of free GSH within plasma and the possibility that heparin and EDTA differentially stabilize REDOX processes of GSH, we examined the impact of time to deproteinization and its interaction with type of anticoagulant used during sample collection on the measured levels of total GSH, free GSH, GSSG, or the ratio of free GSH to GSSG within plasma. We did not find any significant interactions between the type of anticoagulant used during sample collection and time to deproteinization (total GSH: F1,10.71 = 0.2112; P = 0.6550; free GSH: F1,14.39 = 0.1873; P = 0.6716; GSSG: F1,11.71 = 3.7005; P = 0.0790; free GSH:GSSG: F1,11.23 = 1.2756; P = 0.2823). For the marginal main effects we saw different results across the different GSH outcome measures. Specifically, we found no impact of anticoagulant (F1,11 = 0.2288; P = 0.6418) or time to deproteinization (F1,18.72 = 1.0654; P = 0.3151) on measurement of total GSH (Fig. 3A). Similarly, we found no impact of anticoagulant (F1,11.76 = 3.0991; P = 0.1043) or time to deproteinization (F1,19 = 2.8008; p = 0.1106) on measured level of GSSG (Fig. 3C). In contrast, we found the level of free GSH to degrade over time (F1,14.39 = 42.9577; P < 0.0001) in both anticoagulants, with the level of free GSH being higher in EDTA-treated plasma compared to heparinized plasma (F1,10.8 = 10.4686; P = 0.0081; Fig. 3B). Additionally, we found significant differences in the measured ratio of free GSH to GSSG as a function of time (F1,18.96 = 21.2362; P = 0.0002), and a significant difference in anticoagulant (F1,11.21 = 17.1690; P = 0.0016; Fig. 3D). These data suggest that the anticoagulant used for sample collection and time to deproteinization, both independently and additively impact measured levels of free GSH (and thus the ratio of free GSH to GSSG) in samples of plasma which are deproteinized and stored at −80 oC.Fig. 3 Anticoagulant type used during sample collection and time to deproteinization affect some measures of GSH within EDTA-treated and heparinized plasma samples stored at −80 oC.

Time to deproteinization ranged from 30 min to 132 min for heparinized plasma (Preparation Condition 1) and from 39 to 89 min for EDTA-treated plasma (Preparation Condition 2). Circles on the graph depict individual samples (data are corrected for blocking factors) with lines depicting linear regressions between data points (Blue = Preparation Condition 2: EDTA → deproteinization → −80 oC; Orange = Preparation Condition 1: heparin → deproteinization → −80 oC. Neither total GSH (A) nor GSSG (C) differ as a function of time to deproteinization, whereas both free GSH (B) and the ratio of free GSH to GSSG (D) decrease significantly in both samples as time to deproteinization increased. There was no difference in level of total GSH (A) nor GSSG (C) between EDTA-treated or heparinized plasma. Whereas both levels of free GSH (B) and the ratio of free GSH to GSSG (D) were significantly higher in the EDTA-treated plasma compared to the heparinized plasma. These data explain patterns in Fig. 2—In short, total GSH is a more robust measure than free GSH, GSSG, or their ratio as total GSH is not affected by spontaneous oxidation. Individual p-values for marginal main effect of anticoagulant type and time to deproteinization, respectively, are included on each graph.

Discussion

REDOX biology is a novel and promising area of research for understanding the etiology and development of therapeutic interventions for several brain conditions. As with nearly any emerging field of study, there are methodological inconsistencies which can significantly impact measures of analyte(s). Here, we investigated whether anticoagulant type used in blood collection, deproteinization of plasma, and/or storage temperature impacts plasma GSH levels. Additionally, we tested whether the time between sample collection and deproteinization impacted plasma GSH in its various forms (total GSH, free GSH, GSSG, and the ratio of free GSH to GSSG) in heparinized and EDTA-treated plasma samples. Furthermore, we tested if these methodological differences impact post-processing summaries of GSH.

EDTA treatment is preferred to heparin treatment of whole blood for quantification of GSH in plasma

Preparation Condition 1 (heparin → deproteinization → −80 oC) vs. Preparation Condition 2 (EDTA → deproteinization → −80 oC): Anticoagulants used during collection impacts measures of GSH

Suggesting the anticoagulant used during sample collection significantly alters plasma GSH, we found lower levels of free GSH, higher levels of GSSG, and a lower ratio of free GSH to GSSG in the heparinized plasma (Preparation Condition 1) compared to the EDTA-treated plasma (Preparation Condition 2). The observed differences in free GSH, GSSG, and their respective ratio is likely due to spontaneous oxidation of free GSH in the heparinized plasma rather than a matrix effect of anticoagulant type used during sample collection. If this difference was due to a matrix effect, we would expect to observe different levels of total GSH between the two anticoagulants. However, we found no difference between level of total GSH between heparinized (Preparation Condition 1) or EDTA-treated (Preparation Condition 2) plasma which were both deproteinized and stored at −80 oC. Thus, although heparin and EDTA are both capable of preventing coagulation of blood, they differentially stabilize REDOX reactions. As such, careful consideration of anticoagulant used during collection is necessary when conducting research on REDOX related processes.

Preparation Condition 2 (EDTA → deproteinization → −80 oC) vs. Preparation Condition 3 (EDTA → −80 oC → deproteinization): Deproteinization may not be necessary to prevent spontaneous oxidation of free GSH

Unexpectedly, we found no difference in the levels of total GSH, free GSH, GSSG, or the ratio of free GSH to GSSG between the EDTA-treated plasma samples which were stored at −80 oC (Preparation Condition 2 and Preparation Condition 3). This suggests that deproteinization of EDTA-treated plasma prior to storage is not necessary to prevent spontaneous oxidation of free GSH to GSSG. If deproteinization of plasma prior to storage is necessary for prevention of spontaneous oxidation, then there would have been no difference in levels of GSH between Preparation Condition 1 and Preparation Condition 2, and we would have found no difference in level of total GSH and differences in levels of free GSH, GSSG, and their relative ratio between Preparation Condition 3 and Preparation Condition 4. However, we found the opposite to be the case. Therefore, noting the lack of spontaneous oxidation of free GSH between the two EDTA-treated samples stored at −80 oC, this result also suggests that the differences in GSH between the heparinized plasma (Preparation Condition 1) and deproteinized EDTA-treated plasma (Preparation Condition 2) are due to spontaneous oxidation of free GSH occurring in the heparinized plasma prior to the deproteinization process.

Preparation Condition 3 (EDTA → −80 oC → deproteinization) vs. Preparation Condition 4 (EDTA → 4 oC → deproteinization): Storage at −80 oC prevents enzymatic degradation of GSH

We included EDTA-treated plasma stored at 4 oC (Preparation Condition 4) as a positive control for enzymatic processes that may lead to degradation of GSH during sample storage. As hypothesized, we found the level of total GSH, free GSH, and GSSG to be significantly reduced in the EDTA-treated plasma stored at 4 oC compared to the non-deproteinized EDTA-treated plasma stored at −80 oC (Preparation Condition 3). We did not find a difference in the ratio of free GSH to GSSG between any of the EDTA-treated plasma samples. Thus, suggesting that the low level of total GSH found in the EDTA-treated plasma stored at 4 oC is due to enzymatic degradation of total GSH rather than other biological processes [34]. In agreement with prior literature [35], this result suggests that GSH is not stable in plasma at 4 oC, and thus samples should not be stored at this temperature. However, in the case that plasma samples must be stored at 4 oC, only the ratio of free GSH to GSSG should be analyzed.

Time to deproteinization influences levels of GSH in plasma

Given the likelihood that spontaneous oxidation of free GSH to GSSG occurred prior to deproteinization in the heparinized plasma, we assessed whether the duration of time to deproteinization of plasma influenced levels of GSH in Preparation Condition 1 and Preparation Condition 2. We found no differences between the preparation conditions tested nor an effect of time to deproteinization on stability of total GSH, suggesting that use of either heparin or EDTA during sample collection can prevent enzymatic degradation of GSH when samples are stored at −80 oC. As evidenced by a similar negative slope in their respective regression lines, neither of the anticoagulants perfectly prevent spontaneous oxidation of free GSH. Given that there was no difference in total GSH between Preparation Condition 1 and Preparation Condition 2, and that free GSH is significantly higher in Preparation Condition 1 than in Preparation Condition 2, these results suggest that EDTA-treatment more effectively prevents spontaneous oxidation of free GSH in plasma than does heparinization of samples. The noted degradation of free GSH as time to deproteinization increases also suggests that batch processing samples should be avoided. When this is not possible, then time to deproteinization should be controlled for statistically.

General commentary

There are many reasons as to why a preclinical study may fail to translate, particularly, poor experimental design and statistical analysis (for detailed discussions see refs. [36–38]). Here, we highlight another factor, lack of standardized technique, which may also contribute to translational failure. Although several compounds successfully prevent coagulation of blood, each type of anticoagulant works by mediating different aspects of the coagulation pathway (i.e., heparin prevents coagulation by inhibiting the formation of thrombin, whereas EDTA prevents coagulation by chelating calcium in blood [39]). Therefore, as evidenced by our results, different types of anticoagulants may differentially stabilize and prevent spontaneous oxidation of free GSH in plasma (with suggestive evidence that EDTA better prevents spontaneous oxidation than heparin). Differential ability to stabilize free GSH may also explain why some studies report results as the ratio of free GSH to GSSG (e.g., [19, 22–26]). whereas others report outcomes as total GSH (e.g., [12, 19, 24, 26–30]).

It is important to note that we are not the only group to show that the type of anticoagulant used during sample collection impacts measurement of analytes within blood. For example, other studies show similar effects on: plasma measurement of DNA [40, 41], thiols [42], and creatinine [43]. Additionally, while it is common to see total GSH, free GSH, GSSG, and the ratio of free GSH to GSSG, interpreted in terms of a reference range, the reference ranges stated are highly variable (e.g., refs. [44–47]). Accordingly, we are unaware of any formally reported reference ranges in the literature. The growing size of the literature highlighting the impact of research methodologies on quantification of blood-based biological analytes further supports the need for standardized research techniques when it comes to blood-based biomarker collection, preparation, and storage.

In an ideal world, preclinical and clinical studies would always use the same techniques (e.g., the same anticoagulant for sample collection, process samples from one individual at a time, control for environmental confounds, etc.). However, this is not always possible. For example, clinical studies allow for collection of larger volumes of whole blood than the volume provided by mice. Accordingly, clinical studies are afforded the ability to collect whole blood into several different anticoagulant-coated tubes, which is not typically feasible in preclinical studies. Thus, a study conducted in mice, which requires quantification of several analytes, typically collects samples using only one type of anticoagulant-coated tube (most frequently heparin e.g., refs. [15, 24, 28, 48]). However, convenience should never trump scientific best practice, and in the case of GSH, EDTA is more robust and allows for wider post-processing measures than heparin.

Limitations

As with any study, ours has limitations. The first being our selection of animals slated for euthanasia for the study, which resulted in our use of several sub-strains of mice (all on a C57BL/6 background), and limited our ability to control early life experiences. To account for this, the repeated measures approach controlled for individual variation in our analyses. Therefore, the use of several sub-strains of mice is a strength of this study [49, 50], and suggests that our findings are more likely to generalize to other mouse sub-strains than they would have if the study were conducted in only one strain of mice. Another potential limitation of this study is that we only used one assay kit to quantify levels of GSH in our plasma samples. This decision was made based on the benefits provided by the assay kit (ability to measure total GSH, free GSH, and GSSG within a single well) compared to other commercially available kits, and our prior positive results with this kit [15]. A third limitation of our study is that samples were only kept in storage for one to three weeks. As such, it is possible that our results may have been different if samples were in storage for a greater duration of time.

Conclusions and future directions

In conclusion, we report that anticoagulant, sample processing, and sample storage methods significantly affect levels of GSH within plasma samples and that these methodological differences impact post-processing summary reporting. Furthermore, our findings highlight the importance of using standardized techniques in preclinical and clinical research, particularly in the case of REDOX biology as shown here. REDOX imbalance is emerging as an important, underappreciated mechanism across neurodevelopmental, neuropsychiatric, and neurological conditions. To facilitate effective translation, it is vital that there be standardized practices in plasma-based GSH quantification both preclinically and clinically. Based on our results, for plasma-based GSH quantification, we recommend that:Whole blood should be collected into EDTA-coated collection tubes; if this is not possible, then only total GSH should be analyzed.

To prevent enzymatic activity, plasma samples should be stored at −80 oC. Although storage at 4 oC is highly discouraged, should it be necessary, only the ratio of free GSH to GSSG should be analyzed.

Despite being non-significant here, we suggest plasma samples be deproteinized prior to storage, as spontaneous oxidation of free GSH to GSSG may occur if samples are stored for long periods of time.

To better streamline translation, studies reporting the signal of free GSH to GSSG should also report the same analyses with total GSH.

Given the impact of time to deproteinization on the stability of free GSH, studies should avoid batch collecting and processing of blood. In cases where this is unavoidable, processing time (time between collection of whole blood and plasma deproteinization) should be included as a control variable in analyses.

Future directions include: establishing formal reference ranges for plasma GSH; investigating if our results hold true in other species; determining if differences in selection of anticoagulant type negatively impacts translation of REDOX related findings to clinical populations; and examining the impact of long-term storage on sample integrity.

Supplementary information

Data and Analysis

Supplementary information

The online version contains supplementary material available at 10.1038/s41398-024-03086-5.

Acknowledgements

The authors would like to thank the Stanford University Veterinary Service Center for their assistance and care of animals used in this study, and Ben Franco and Kaleigh Beacham for their support with collection of blood. This project was supported by a Stanford Medicine Discovery Innovation Award to Dr. Joseph P. Garner.

Author contributions

Conceptualization: KMC, KJP, JPG; Sample Collection/ Processing: KMC, BESB, JNB; Assay kit optimization: KMC, RM, JPG; Sample analysis: KMC, DN; Statistical Analysis: KMC, JPG; Writing—original draft: KMC; Writing—reviewing and editing: DN, RM, BESB, JNB, KJP, JPG; Data visualization: KMC, JPG; Funding: JPG

Data availability

Data and analysis code are available in supplementary information.

Competing interests

The authors declare no competing interests.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References

1. Carney JM Starke-Reed PE Oliver CN Landum RW Cheng MS Wu JF Reversal of age-related increase in brain protein oxidation, decrease in enzyme activity, and loss in temporal and spatial memory by chronic administration of the spin-trapping compound N-tert-butyl-alpha-phenylnitrone Proc Natl Acad Sci USA 1991 88 3633 6 10.1073/pnas.88.9.3633 1673789
Carney JM, Starke-Reed PE, Oliver CN, Landum RW, Cheng MS, Wu JF, et al. Reversal of age-related increase in brain protein oxidation, decrease in enzyme activity, and loss in temporal and spatial memory by chronic administration of the spin-trapping compound N-tert-butyl-alpha-phenylnitrone. Proc Natl Acad Sci USA 1991;88:3633–6.1673789
2. Floyd RA Role of oxygen free radicals in carcinogenesis and brain ischemia FASEB J 1990 4 2587 97 10.1096/fasebj.4.9.2189775 2189775
Floyd RA. Role of oxygen free radicals in carcinogenesis and brain ischemia. FASEB J. 1990;4:2587–97.2189775
3. Bonnefont-Rousselot D Glucose and reactive oxygen species Curr Opin Clin Nutr Metab Care 2002 5 561 8 10.1097/00075197-200209000-00016 12172481
Bonnefont-Rousselot D. Glucose and reactive oxygen species. Curr Opin Clin Nutr Metab Care. 2002;5:561–8.12172481
4. Slivka A Cohen G Hydroxyl radical attack on dopamine J Biol Chem 1985 260 15466 72 10.1016/S0021-9258(17)36277-4 2999117
Slivka A, Cohen G. Hydroxyl radical attack on dopamine. J Biol Chem. 1985;260:15466–72.2999117
5. Gabby M Tauber M Porat S Simantov R Selective role of glutathione in protecting human neuronal cells from dopamine-induced apoptosis Neuropharmacology 1996 35 571 8 10.1016/0028-3908(96)84626-0 8887964
Gabby M, Tauber M, Porat S, Simantov R. Selective role of glutathione in protecting human neuronal cells from dopamine-induced apoptosis. Neuropharmacology. 1996;35:571–8.8887964
6. Ighodaro OM Akinloye OA First line defence antioxidants-superoxide dismutase (SOD), catalase (CAT) and glutathione peroxidase (GPX): Their fundamental role in the entire antioxidant defence grid Alex J Med 2018 54 287 93
Ighodaro OM, Akinloye OA. First line defence antioxidants-superoxide dismutase (SOD), catalase (CAT) and glutathione peroxidase (GPX): Their fundamental role in the entire antioxidant defence grid. Alex J Med. 2018;54:287–93.
7. Salim S Oxidative Stress and the Central Nervous System J Pharm Exp Ther 2017 360 201 5 10.1124/jpet.116.237503
Salim S. Oxidative Stress and the Central Nervous System. J Pharm Exp Ther. 2017;360:201–5.
8. Aoyama K Glutathione in the Brain Int J Mol Sci 2021 22 5010 10.3390/ijms22095010 34065042
Aoyama K. Glutathione in the Brain. Int J Mol Sci. 2021;22:5010.34065042
9. Dringen R Kussmaul L Gutterer JM Hirrlinger J Hamprecht B The glutathione system of peroxide detoxification is less efficient in neurons than in astroglial cells J Neurochem 1999 72 2523 30 10.1046/j.1471-4159.1999.0722523.x 10349863
Dringen R, Kussmaul L, Gutterer JM, Hirrlinger J, Hamprecht B. The glutathione system of peroxide detoxification is less efficient in neurons than in astroglial cells. J Neurochem. 1999;72:2523–30.10349863
10. Dringen R Pfeiffer B Hamprecht B Synthesis of the antioxidant glutathione in neurons: supply by astrocytes of CysGly as precursor for neuronal glutathione J Neurosci 1999 19 562 9 10.1523/JNEUROSCI.19-02-00562.1999 9880576
Dringen R, Pfeiffer B, Hamprecht B. Synthesis of the antioxidant glutathione in neurons: supply by astrocytes of CysGly as precursor for neuronal glutathione. J Neurosci. 1999;19:562–9.9880576
11. Iskusnykh IY Zakharova AA Pathak D Glutathione in Brain Disorders and Aging Molecules 2022 27 324 10.3390/molecules27010324 35011559
Iskusnykh IY, Zakharova AA, Pathak D. Glutathione in Brain Disorders and Aging. Molecules. 2022;27:324.35011559
12. Nucifora LG Tanaka T Hayes LN Kim M Lee BJ Matsuda T Reduction of plasma glutathione in psychosis associated with schizophrenia and bipolar disorder in translational psychiatry Transl Psychiatry 2017 7 e1215 10.1038/tp.2017.178 28892069
Nucifora LG, Tanaka T, Hayes LN, Kim M, Lee BJ, Matsuda T, et al. Reduction of plasma glutathione in psychosis associated with schizophrenia and bipolar disorder in translational psychiatry. Transl Psychiatry. 2017;7:e1215.28892069
13. James SJ Cutler P Melnyk S Jernigan S Janak L Gaylor DW Metabolic biomarkers of increased oxidative stress and impaired methylation capacity in children with autism Am J Clin Nutr 2004 80 1611 7 10.1093/ajcn/80.6.1611 15585776
James SJ, Cutler P, Melnyk S, Jernigan S, Janak L, Gaylor DW, et al. Metabolic biomarkers of increased oxidative stress and impaired methylation capacity in children with autism. Am J Clin Nutr. 2004;80:1611–7.15585776
14. George NM Whitaker J Vieira G Geronimo JT Bellinger DA Fletcher CA Antioxidant Therapies for Ulcerative Dermatitis: A Potential Model for Skin Picking Disorder PLoS One 2015 10 e0132092 10.1371/journal.pone.0132092 26167859
George NM, Whitaker J, Vieira G, Geronimo JT, Bellinger DA, Fletcher CA, et al. Antioxidant Therapies for Ulcerative Dermatitis: A Potential Model for Skin Picking Disorder. PLoS One. 2015;10:e0132092.26167859
15. Vieira GLT Lossie AC Lay DC Jr Radcliffe JS Garner JP Preventing, treating, and predicting barbering: A fundamental role for biomarkers of oxidative stress in a mouse model of Trichotillomania PLoS One 2017 12 e0175222 10.1371/journal.pone.0175222 28426681
Vieira GLT, Lossie AC, Lay DC Jr, Radcliffe JS, Garner JP. Preventing, treating, and predicting barbering: A fundamental role for biomarkers of oxidative stress in a mouse model of Trichotillomania. PLoS One. 2017;12:e0175222.28426681
16. Ng F Berk M Dean O Bush AI Oxidative stress in psychiatric disorders: evidence base and therapeutic implications Int J Neuropsychopharmacol 2008 11 851 76 10.1017/S1461145707008401 18205981
Ng F, Berk M, Dean O, Bush AI. Oxidative stress in psychiatric disorders: evidence base and therapeutic implications. Int J Neuropsychopharmacol. 2008;11:851–76.18205981
17. Curpan AS Luca A-C Ciobica A Potential Novel Therapies for Neurodevelopmental Diseases Targeting Oxidative Stress Oxid Med Cell Longev 2021 2021 6640206 10.1155/2021/6640206 34336109
Curpan AS, Luca A-C, Ciobica A. Potential Novel Therapies for Neurodevelopmental Diseases Targeting Oxidative Stress. Oxid Med Cell Longev. 2021;2021:6640206.34336109
18. Dean O Giorlando F Berk M N-acetylcysteine in psychiatry: current therapeutic evidence and potential mechanisms of action J Psychiatry Neurosci 2011 36 78 86 10.1503/jpn.100057 21118657
Dean O, Giorlando F, Berk M. N-acetylcysteine in psychiatry: current therapeutic evidence and potential mechanisms of action. J Psychiatry Neurosci. 2011;36:78–86.21118657
19. Schmitt B Vicenzi M Garrel C Denis FM Effects of N-acetylcysteine, oral glutathione (GSH) and a novel sublingual form of GSH on oxidative stress markers: A comparative crossover study Redox Biol 2015 6 198 205 10.1016/j.redox.2015.07.012 26262996
Schmitt B, Vicenzi M, Garrel C, Denis FM. Effects of N-acetylcysteine, oral glutathione (GSH) and a novel sublingual form of GSH on oxidative stress markers: A comparative crossover study. Redox Biol. 2015;6:198–205.26262996
20. Pastore A Piemonte F Locatelli M Lo Russo A Gaeta LM Tozzi G Determination of Blood Total, Reduced, and Oxidized Glutathione in Pediatric Subjects Clin Chem 2001 47 1467 9 10.1093/clinchem/47.8.1467 11468240
Pastore A, Piemonte F, Locatelli M, Lo Russo A, Gaeta LM, Tozzi G, et al. Determination of Blood Total, Reduced, and Oxidized Glutathione in Pediatric Subjects. Clin Chem. 2001;47:1467–9.11468240
21. Tomin T Bordag N Zügner E Al-Baghdadi A Schinagl M Birner-Gruenberger R Blood Plasma Quality Control by Plasma Glutathione Status Antioxidants 2021 10 864 10.3390/antiox10060864 34072235
Tomin T, Bordag N, Zügner E, Al-Baghdadi A, Schinagl M, Birner-Gruenberger R, et al. Blood Plasma Quality Control by Plasma Glutathione Status. Antioxidants. 2021;10:864.34072235
22. Navarro J Obrador E Pellicer JA Aseni M Viña J Estrela JM Blood glutathione as an index of radiation-induced oxidative stress in mice and humans Free Radic Biol Med 1997 22 1203 9 10.1016/S0891-5849(96)00554-0 9098094
Navarro J, Obrador E, Pellicer JA, Aseni M, Viña J, Estrela JM. Blood glutathione as an index of radiation-induced oxidative stress in mice and humans. Free Radic Biol Med. 1997;22:1203–9.9098094
23. Wink LK Adams R Wang Z Klaunig JE Plawecki MH Posey DJ A randomized placebo-controlled pilot study of N-acetylcysteine in youth with autism spectrum disorder Mol Autism 2016 7 26 10.1186/s13229-016-0088-6 27103982
Wink LK, Adams R, Wang Z, Klaunig JE, Plawecki MH, Posey DJ, et al. A randomized placebo-controlled pilot study of N-acetylcysteine in youth with autism spectrum disorder. Mol Autism. 2016;7:26.27103982
24. Leeuwenburgh C Ji LL Glutathone and Glutathione Ethyl Ester Supplementation of Mice Alter Glutathione Homeostasis during Exercise12 J Nutr 1998 128 2420 6 10.1093/jn/128.12.2420 9868190
Leeuwenburgh C, Ji LL. Glutathone and Glutathione Ethyl Ester Supplementation of Mice Alter Glutathione Homeostasis during Exercise12. J Nutr. 1998;128:2420–6.9868190
25. Gyuraszova M Kovalcikova A Jansakova K Sebekova K Celec P Tothova L Markers of oxidative stress and antioxidant status in the plasma, urine and saliva of healthy mice Physiol Res 2018 67 921 34 10.33549/physiolres.933866 30204460
Gyuraszova M, Kovalcikova A, Jansakova K, Sebekova K, Celec P, Tothova L. Markers of oxidative stress and antioxidant status in the plasma, urine and saliva of healthy mice. Physiol Res. 2018;67:921–34.30204460
26. Ammon HPT Klumpp S Fuß A Verspohl EJ Jaeschke H Wendel A A possible role of plasma glutathione in glucose-mediated insulin secretion: in vitro and in vivo studies in rats Diabetologia 1989 32 797 800 10.1007/BF00264910 2687063
Ammon HPT, Klumpp S, Fuß A, Verspohl EJ, Jaeschke H, Wendel A, et al. A possible role of plasma glutathione in glucose-mediated insulin secretion: in vitro and in vivo studies in rats. Diabetologia. 1989;32:797–800.2687063
27. McAllister MJ Basham SA Waldman HS Smith JW Mettler JA Butawan MB Effects of psychological stress during exercise on markers of oxidative stress in young healthy, trained men Physiol Behav 2019 198 90 5 10.1016/j.physbeh.2018.10.013 30342058
McAllister MJ, Basham SA, Waldman HS, Smith JW, Mettler JA, Butawan MB, et al. Effects of psychological stress during exercise on markers of oxidative stress in young healthy, trained men. Physiol Behav. 2019;198:90–5.30342058
28. Ono H Sakamoto A Sakura N Plasma total glutathione concentrations in healthy pediatric and adult subjects Clin Chim Acta 2001 312 227 9 10.1016/S0009-8981(01)00596-4 11580931
Ono H, Sakamoto A, Sakura N. Plasma total glutathione concentrations in healthy pediatric and adult subjects. Clin Chim Acta. 2001;312:227–9.11580931
29. Lavoie S Murray MM Deppen P Knyazeva MG Berk M Boulat O Glutathione Precursor, N-Acetyl-Cysteine, Improves Mismatch Negativity in Schizophrenia Patients Neuropsychopharmacology 2008 33 2187 99 10.1038/sj.npp.1301624 18004285
Lavoie S, Murray MM, Deppen P, Knyazeva MG, Berk M, Boulat O, et al. Glutathione Precursor, N-Acetyl-Cysteine, Improves Mismatch Negativity in Schizophrenia Patients. Neuropsychopharmacology. 2008;33:2187–99.18004285
30. LEE J Clarkson PM Plasma Creatine Kinase Activity and Glutathione after Eccentric Exercise Med Sci Sports Exerc 2003 35 930 6 10.1249/01.MSS.0000069553.47739.36 12783040
LEE J, Clarkson PM. Plasma Creatine Kinase Activity and Glutathione after Eccentric Exercise. Med Sci Sports Exerc. 2003;35:930–6.12783040
31. Kovalcsik R Devlin T Loux S Martinek M May J Pickering T Animal reuse: balancing scientific integrity and animal welfare Lab Anim 2006 35 49 53 10.1038/laban1006-49
Kovalcsik R, Devlin T, Loux S, Martinek M, May J, Pickering T, et al. Animal reuse: balancing scientific integrity and animal welfare. Lab Anim. 2006;35:49–53.
32. Grafen, AHR, Modern Statistics for the life sciences. 2002: Oxford University Press.
33. Gaskill BN Garner JP Power to the People: Power, Negative Results and Sample Size J Am Assoc Lab Anim Sci 2020 59 9 16 10.30802/AALAS-JAALAS-19-000042 31852563
Gaskill BN, Garner JP. Power to the People: Power, Negative Results and Sample Size. J Am Assoc Lab Anim Sci. 2020;59:9–16.31852563
34. Lin SK Tsai SM Huang JC Lee SC Wu SH Wu SH Effects of storage time and temperature on the stability of glutathione in deproteinized blood sample J Food Drug Anal 2006 14 621 863
Lin SK, Tsai SM, Huang JC, Lee SC, Wu SH, Wu SH, et al. Effects of storage time and temperature on the stability of glutathione in deproteinized blood sample. J Food Drug Anal. 2006;14:621–863.
35. Wahdati A Lindner A Sommer H [The effect of storage time and temperature on the activity of glutathione peroxidase in plasma and whole blood of horses] Zentralbl Veterinarmed A 1992 39 81 8 10.1111/j.1439-0442.1992.tb00159.x 1590039
Wahdati A, Lindner A, Sommer H. [The effect of storage time and temperature on the activity of glutathione peroxidase in plasma and whole blood of horses]. Zentralbl Veterinarmed A. 1992;39:81–8.1590039
36. Begley CG Six red flags for suspect work Nature 2013 497 433 4 10.1038/497433a 23698428
Begley CG. Six red flags for suspect work. Nature. 2013;497:433–4.23698428
37. Garner JP Gaskill BN Weber EM Ahloy-Dallaire J Pritchett-Corning KR Introducing Therioepistemology: the study of how knowledge is gained from animal research Lab Anim (NY) 2017 46 103 13 10.1038/laban.1224 28328885
Garner JP, Gaskill BN, Weber EM, Ahloy-Dallaire J, Pritchett-Corning KR. Introducing Therioepistemology: the study of how knowledge is gained from animal research. Lab Anim (NY). 2017;46:103–13.28328885
38. Würbel H, Garner JP. Refinement of rodent research through environmental enrichment and systematic randomization. Available at NC3Rs.org.uk. 2007.
39. Bayot ML, Tadi P. Laboratory Test Tube. 2022: StatPearls.
40. Lam NYL Rainer TH Chiu RWK Lo YMD EDTA Is a Better Anticoagulant than Heparin or Citrate for Delayed Blood Processing for Plasma DNA Analysis Clin Chem 2004 50 256 7 10.1373/clinchem.2003.026013 14709670
Lam NYL, Rainer TH, Chiu RWK, Lo YMD. EDTA Is a Better Anticoagulant than Heparin or Citrate for Delayed Blood Processing for Plasma DNA Analysis. Clin Chem. 2004;50:256–7.14709670
41. Sotelo-Orozco J Chen S-Y Hertz-Picciotto I Slupsky CM A Comparison of Serum and Plasma Blood Collection Tubes for the Integration of Epidemiological and Metabolomics Data Front Mol Biosci 2021 8 682134 10.3389/fmolb.2021.682134 34307452
Sotelo-Orozco J, Chen S-Y, Hertz-Picciotto I, Slupsky CM. A Comparison of Serum and Plasma Blood Collection Tubes for the Integration of Epidemiological and Metabolomics Data. Front Mol Biosci. 2021;8:682134.34307452
42. Salazar JF Herbeth B Siest G Leroy P Leroy, Stability of Blood Homocysteine and Other Thiols: EDTA or Acidic Citrate? Clin Chem 1999 45 2016 9 10.1093/clinchem/45.11.2016 10545077
Salazar JF, Herbeth B, Siest G, Leroy P. Leroy, Stability of Blood Homocysteine and Other Thiols: EDTA or Acidic Citrate? Clin Chem. 1999;45:2016–9.10545077
43. Mohri M Shakeri H Lotfollah Zadeh S Effects of common anticoagulants (heparin, citrate and EDTA) on routine plasma biochemistry of cattle Comp Clin Pathol 2007 16 207 9 10.1007/s00580-006-0664-9
Mohri M, Shakeri H, Lotfollah Zadeh S. Effects of common anticoagulants (heparin, citrate and EDTA) on routine plasma biochemistry of cattle. Comp Clin Pathol. 2007;16:207–9.
44. Hwang C Sinskey AJ Lodish HF Oxidized Redox State of Glutathione in the Endoplasmic Reticulum Science 1992 257 1496 502 10.1126/science.1523409 1523409
Hwang C, Sinskey AJ, Lodish HF. Oxidized Redox State of Glutathione in the Endoplasmic Reticulum. Science. 1992;257:1496–502.1523409
45. Majumder P Duchen MR Gale JE Cellular glutathione content in the organ of Corti and its role during ototoxicity Front Cell Neurosci 2015 9 143 10.3389/fncel.2015.00143 25972783
Majumder P, Duchen MR, Gale JE. Cellular glutathione content in the organ of Corti and its role during ototoxicity. Front Cell Neurosci. 2015;9:143.25972783
46. Meister A Glutathione metabilist and its selective modification J Biol Chem 1988 263 17205 8 10.1016/S0021-9258(19)77815-6 3053703
Meister A. Glutathione metabilist and its selective modification. J Biol Chem. 1988;263:17205–8.3053703
47. Michelet F Gueguen R Leroy P Wellman M Nicolas A Siest G Blood and plasma glutathione measured in healthy subjects by HPLC: relation to sex, aging, biological variables, and life habits Clin Chem 1995 41 1509 17 10.1093/clinchem/41.10.1509 7586526
Michelet F, Gueguen R, Leroy P, Wellman M, Nicolas A, Siest G. Blood and plasma glutathione measured in healthy subjects by HPLC: relation to sex, aging, biological variables, and life habits. Clin Chem. 1995;41:1509–17.7586526
48. Christensen LL Selman C Blount JD Pilkington JG Watt KA Pemberton JM Plasma markers of oxidative stress are uncorrelated in a wild mammal Ecol Evol 2015 5 5096 108 10.1002/ece3.1771 26640685
Christensen LL, Selman C, Blount JD, Pilkington JG, Watt KA, Pemberton JM, et al. Plasma markers of oxidative stress are uncorrelated in a wild mammal. Ecol Evol. 2015;5:5096–108.26640685
49. Walker M Fureix C Palme R Newman JA Ahloy Dallaire J Mason G Mixed-strain housing for female C57BL/6, DBA/2, and BALB/c mice: validating a split-plot design that promotes refinement and reduction BMC Med Res Methodol 2016 16 11 10.1186/s12874-016-0113-7 26817696
Walker M, Fureix C, Palme R, Newman JA, Ahloy Dallaire J, Mason G. Mixed-strain housing for female C57BL/6, DBA/2, and BALB/c mice: validating a split-plot design that promotes refinement and reduction. BMC Med Res Methodol. 2016;16:11.26817696
50. Festing MFW Randomized Block Experimental Designs Can Increase the Power and Reproducibility of Laboratory Animal Experiments ILAR J 2014 55 472 6 10.1093/ilar/ilu045 25541548
Festing MFW. Randomized Block Experimental Designs Can Increase the Power and Reproducibility of Laboratory Animal Experiments. ILAR J. 2014;55:472–6.25541548
