Mitochondrial Glutathione Transferase Zeta 1 Is Inactivated More Rapidly by Dichloroacetate than the Cytosolic Enzyme in Adult and Juvenile Rat Liver

Abstract

Dichloroacetate (DCA) has potential for treating mitochondrial disorders and cancer by activating the mitochondrial pyruvate dehydrogenase complex. Repeated dosing of DCA results in reduced drug clearance due to inactivation of glutathione transferase ζ1 (GSTZ1), its metabolizing enzyme. We investigated the time-course of inactivation of GSTZ1 in hepatic cytosol and mitochondria after one oral dose of 100 mg/kg DCA to female Sprague–Dawley rats aged 4 weeks (young) and 52 weeks (adult) as models for children and adults, respectively. GSTZ1 activity with both DCA and an endogenous substrate, maleylacetone (MA), as well as GSTZ1 protein expression were rapidly reduced in cytosol from both ages following DCA treatment. In mitochondria, loss of GSTZ1 protein and activity with DCA were even more rapid. The cytosolic in vivo half-lives of the loss of GSTZ1 activity with DCA were 1.05 ± 0.03 and 0.82 ± 0.02 h (mean ± S.D., n = 6) for young and adult rats, respectively, with inactivation significantly more rapid in adult rats, p < 0.001. The mitochondrial inactivation half-lives were similar in young (0.57 ± 0.02 h) and adult rats (0.54 ± 0.02 h) and were significantly (p < 0.0001) shorter than cytosolic inactivation half-lives. By 24 h after DCA administration, activity and expression remained at 10% or less than control values. The in vitro GSTZ1 inactivation half-lives following incubation with 2 mM DCA in the presence of physiological chloride (Cl⁻) concentrations (cytosol = 44 mM, mitochondria = 1–2 mM) exhibited marked differences between subcellular fractions, being 3 times longer in the cytosol than in the mitochondria, regardless of age, suggesting that the lower Cl⁻ concentration in mitochondria explained the faster degradation of GSTZ1. These results demonstrate for the first time that rat mitochondrial GSTZ1 is more readily inactivated by DCA than cytosolic GSTZ1, and cytosolic GSTZ1 is inactivated more rapidly in adult than young rats.

Graphical Abstract

INTRODUCTION

Over the last half century, dichloroacetate (DCA) has been investigated for both its intriguing pharmacological properties and its presence as an environmental contaminant.¹ Studies in rodents chronically exposed to high levels of DCA (>1g/L in drinking water for 2 years) have shown multiorgan toxicities, particularly to the liver.² DCA has been measured in groundwater and surface water at concentrations less than 100 μg/L in the United States³ and is found in chlorinated drinking water. Because of toxicity concerns, the United States Environmental Protection Agency has set a limit in drinking water for five halogenated acetic acids combined, including DCA, to be 60 μg/L.⁴ The World Health Organization has set a provisional guideline for levels of DCA alone in drinking water of 50 μg/L.³ These exposure levels are considerably lower than doses linked to adverse effects observed in rats. In contrast, daily DCA doses between 10 and 50 mg/kg have been administered as an investigational drug to children and adults with various disorders of intermediary metabolism, including congenital and acquired forms of lactic acidosis, diabetes, and cancer, with generally good tolerability and safety.^5–10 To date, the main adverse effect of therapeutic use of DCA is a reversible peripheral neuropathy, which affects adults more frequently than children.¹¹

The therapeutic mechanism of action for DCA involves the pyruvate dehydrogenase complex (PDC), a mitochondrial matrix complex that converts glycolysis-derived pyruvate to acetyl-CoA.¹² The PDC is reversibly phosphorylated and inhibited by pyruvate dehydrogenase kinase (PDK).^13,14 DCA activates the PDC both by inhibiting PDK and by decreasing the rate of turnover of the complex.¹ Consequently, the mitochondrion is regarded as the most essential site of DCA’s pharmacodynamics.

The only known primary metabolite of DCA is glyoxylate, which is inactive toward PDK. Dechlorination of DCA to glyoxylate is catalyzed by glutathione transferase ζ1 (GSTZ1), a phase II enzyme found in hepatic cytosol and mitochondrial matrix. GSTZ1 is also known as maleylacetoacetate isomerase (MAAI) for its role in tyrosine catabolism.^15–17 GSTZ1 catalyzes the penultimate step in the tyrosine catabolic pathway, isomerizing maleylacetoacetate (MAA) and maleylacetone (MA) to fumarylacetoacetate and fumarylacetone (FA), respectively, around a carbon double bond. Reactions with DCA, MAA, and MA require, but do not consume, glutathione (GSH). Due to the selective nature of MAAI/GSTZ1, the efficiency of isomerization of the tyrosine catabolites is very high, having a specific activity hundreds of times greater than with DCA or other xenobiotic substrates.¹⁸ In dose- and time-dependent manners, treatment of rats with DCA has been shown to reduce the ability of hepatic cytosolic GSTZ1 to metabolize DCA.^19–22 DCA is both a substrate of GSTZ1 and a mechanism-based inactivator. It has been shown that DCA can form adducts with GSTZ1, and it is thought that adduct formation triggers the observed loss of the GSTZ1 protein and subsequent reduction of the DCA metabolism,¹¹ though the mechanism of loss of the adducted protein is not yet known.

In vitro studies in rat and human liver cytosol as well as in vivo studies in rats and people have shown that numerous factors affect the rate of DCA-induced GSTZ1 inactivation, including age and physiological chloride (Cl⁻) concentration. Age has been shown to be directly related to the clearance rate of DCA and related pharmacotoxicology after chronic administration, where increasing age is more often associated with adverse outcomes, primarily the prevalence of reversible peripheral neuropathy.^1,11,23 Studies showed that in vitro inactivation of GSTZ1 in human liver cytosol was rapid in the absence of Cl⁻ but slowed in a concentration-dependent manner in the presence of Cl⁻, and the protective effect was haplotype-dependent.²⁴ The mechanism of this effect of Cl⁻ and other anions is not yet determined.²⁴ Related to the presence of GSTZ1 in cytosol and mitochondria, we determined Cl⁻ concentrations in human liver from donors aged 1 day to 84 years were more than 25-fold higher in the cytosol (mean = 105 mM) than in the mitochondria (mean = 4.2 mM) and that cytosolic Cl⁻ concentrations were slightly reduced with increased age.²⁵ In 4 week old rats, the cytosolic [Cl⁻] was 44.9 mM, similar to the 43.0 mM found in adult rats. This was considerably higher than the mitochondrial [Cl⁻] in the young rats, 0.8 mM, or that in the adult rats, 1.9 mM.²⁶ We hypothesized that the much lower Cl⁻ level in the mitochondria would lead to mitochondrial GSTZ1 being more rapidly inactivated than cytosolic GSTZ1 upon DCA administration.

DCA exerts its principal dynamic effects in mitochondria,¹² and its metabolite, glyoxylate, is inactive. However, the properties of mitochondrial GSTZ1 are less studied than those of the cytosolic enzyme.²⁷ It is not known if the mitochondrial enzyme is similar to the cytosolic enzyme with respect to inactivation by DCA. There are reports that the biotransformation of DCA is inversely proportional to age in mice, rats, and humans.^19,23,28 Moreover, drug-induced neuropathy can be recapitulated in female Sprague–Dawley (S-D) rats,²⁹ making these animals a useful model for investigating DCA metabolism, since similarities in the fate of DCA between rats and humans have been noted.²⁹ Here, we examined the time-course of cytosolic and mitochondrial hepatic GSTZ1 inactivation after a single administration of DCA to young and adult female S-D rats to determine the response of mitochondrial GSTZ1 to DCA. Accordingly, we tested the hypothesis that GSTZ1 protein and related drug-metabolizing activity would be lost more rapidly from its location in the mitochondrial matrix than from the cytosol as a function of age.

MATERIALS AND METHODS

Chemicals and Reagents.

¹⁴C-DCA (specific radioactivity = 56 mCi/mmol, 99% purity) was obtained from American Radiolabeled Chemicals (St. Louis, MO), converted to its sodium salt by addition of NaHCO₃, and diluted with unlabeled sodium DCA to make a 2 mM substrate solution that contains 10–20 μCi/mL. Unlabeled clinical grade sodium DCA was purchased from TCI America (Portland, OR) and was >99% pure by GC–MS analysis. MA and FA were synthesized by the method of Fowler and Seltzer,³⁰ where the butenolide intermediate was purified by column chromatography in a solvent system of heptane:ethyl acetate (3:1). ¹H NMR and UV spectroscopy confirmed the identity of MA and FA. An E. coli expression vector (pET21a) containing DNA for N-terminal His-tagged rat GSTZ1 (NP_0011029515.1) was produced by Bio Basic Inc. (Markham, ON), and the expressed enzyme was purified by Ni-agarose affinity chromatography. ¹³C-GSH was purchased from Cambridge Isotope Laboratories, Inc. (Cambridge, MA). All other chemicals used in this study were of high purity and purchased from commercial suppliers.

Animals and DCA Administration.

These studies were approved by the University of Florida’s Institutional Animal Care and Use Committee. To model children, we used rats aged 4 weeks, and to model midlife adults, we studied rats aged 52 weeks, as the normal rat lifespan is 2 years. Female S-D rats were purchased from Charles River Laboratories (Wilmington, MA). Rats for study at 4 weeks (young) were obtained at 3 weeks of age, and two animals were housed per cage for 1 week before use. Rats for study at 52 weeks (adult) were retired breeders, purchased at approximately 32 weeks of age, and were housed one (>500 g body weight) or two (<500 g body weight) per cage until approximately 52 weeks of age for use. All rats were held in secure animal facilities under a 12 h light/dark cycle (7 AM to 7 PM, Eastern Standard Time throughout the year) with constant temperature and humidity conditions. Animals had free access to water and rat chow during the study’s duration. At the time of study, the body weight of young rats was 94.3 ± 12.0 g, n = 90 (mean ± S.D.) and 413 ± 53.8 g, n = 88, for the adult rats.

Sodium DCA (100 mg/kg) was administered by oral gavage from a 100 mg/mL water solution, a dose we previously found reduced hepatic GSTZ1 activity 90% by 24 h.²¹ Considering the FDA-recommended scaling of drug doses between humans and rats, 100 mg/kg to the rat corresponds to a human dose of 16.1 mg/kg.³¹ As noted in the Introduction, this dose is within the range administered as investigational therapy to humans. Doses were given at 8:00 AM Eastern Standard Time, and rats were euthanized at 0.25, 0.5, 1, 2, 4, 8, 12, and 24 h after dosing (n = 6, per time point and age). Control rats (n = 6, per time point and age) dosed orally with sodium acetate (100 mg/kg; dose of 100 mg/mL of water), were subjected to the same dosing time-course, except that no separate control group was used for the 0.25 h time point. Controls were studied at each time point, as there are reports of diurnal variation in liver to body weight ratios and activities of other GST enzymes.^32,33

Subcellular Fractionation of Rat Livers.

After the rats were euthanized by CO₂ inhalation at the specified time intervals, livers were removed and rinsed with buffer pH 7.4 containing 1.15% potassium chloride and 0.05 M potassium phosphate to remove excess blood. Portions of the liver (3 g) were homogenized in 4 volumes of homogenizing buffer (0.25 M sucrose, 0.02 M HEPES–NaOH, pH 7.4, 0.1 mM phenylmethanesulfonylfluoride) to prepare subcellular fractions, i.e., the cytosol, washed mitochondria, and washed microsomes, as previously reported.¹⁷ The washed mitochondria and microsomes were suspended in 3 mL of resuspension buffer (0.01 M HEPES–NaOH, pH 7.4, 0.1 mM dithiothreitol, 0.1 mM EDTA, 0.1 mM phenylmethanesulfonylfluoride, and 5% glycerol). Individual sample fractions were flushed with nitrogen and stored in aliquots at −80 °C until use.

Prior to assay of activity, samples of washed mitochondria were sonicated on ice twice for 15 s, with 25 s between bursts to ensure disruption of the mitochondrial membrane, as GSTZ1 is localized in the matrix. Buffer exchange of cytosolic and sonicated, washed mitochondrial samples was completed by ultrafiltration for four cycles at 5000g for 30 min with 10 kDa molecular weight cutoff Amicon centrifugal filters (Millipore Corporation, Billerica, MA) against 1.15% potassium chloride and 0.05 M potassium phosphate buffer (pH 7.4). This washing process removed small molecules before use in assays, as sucrose has been shown to inhibit GSTZ1 activity.¹⁷ Determination of the protein concentration of each sample was by the bicinchoninic acid protein assay (Thermo Fisher Scientific, Waltham, MA) with bovine serum albumin as the protein standard.

Cytosolic and Mitochondrial GSTZ1 Activity Measurement with DCA as the Substrate.

An established method with ¹⁴C-DCA as the substrate was used to measure GSTZ1 activity.³⁴ To determine the rate of transformation of DCA to glyoxylate, the assay tubes contained 0.1 mg or 0.3 mg of dialyzed protein (cytosol or mitochondria, respectively) with a saturating concentration of ¹⁴C-DCA of 0.2, 1, or 5 mM GSH (cytosol or mitochondria) and 0.1 M HEPES–NaOH (pH 7.6), in an assay volume of 0.1 mL. Blanks contained no protein. The reaction was initiated with the addition of protein, incubated at 37 °C for 10 or 30 min (cytosol or mitochondria), and terminated with 100 μL of ice-cold methanol. Precipitated protein was sedimented by centrifugation, and the supernatant was filtered through a 0.22 μm Costar Spin-X centrifuge nylon filter (Corning Incorporated, Corning, NY) before analysis by HPLC coupled with in-line flow-through liquid scintillation radio-chemical detection using an INUS β-Ram detector (Lab-Logic, Tampa, FL).³⁴ Analysis of enzyme activity was performed in duplicate for cytosol and mitochondria from each control and DCA-treated rat and computed as nmol of glyoxylate formed per min of incubation per mg of protein. Substrate consumption did not exceed 15%. The limit of quantitation was 0.15 nmol of glyoxylate. Samples with product formation less than this limit of quantitation were transformed to a value equal to the limit of quantitation divided by the square root of 2 (0.16 nmol).³⁵ The method of Tzeng et al.²² was used to calculate the in vivo half-life of loss of GSTZ1 activity from the linear portion of natural log–linear plots of the ratio of mean GSTZ1 activities of DCA-treated cytosol and mitochondria to respective control activities at each time point and age.

Cytosolic GSTZ1 Activity Measurement with MA as the Substrate.

GSTZ1 activity was determined by measuring the formation of FA from MA, with modifications to a published method.³⁶ Activities were determined at 25 °C using a reaction mixture of rat hepatic cytosolic protein (0.1–2 mg), 1 mM GSH, 0.01 M potassium phosphate buffer, pH 7.4, and 1 mM MA in acetonitrile (made fresh daily), in a final volume of 0.25 mL. Acetonitrile accounted for 0.025 mL of the reaction mixture volume. The isomerization reaction was initiated with 1 mM MA and quenched after 30 s with 50 μL of 250 μM 7-ethoxycoumarin in acetonitrile, the internal standard. Tubes were immediately vortex-mixed, placed on ice, and then centrifuged to sediment protein, and the supernatant was transferred to a clean tube for analysis. Previous assessment of GSTZ1 activity with DCA as the substrate guided the amount of rat cytosolic protein used in the assay with MA. For each of the DCA treatment time points (0.25, 0.5, 1, 2, 4, 8, 24 h), activity with MA was determined in hepatic cytosol from three rats of both ages, with three controls for each age. The samples selected for study with MA had activities with DCA that were closest to the mean for each DCA-treated time point or for all control samples.

All samples were filtered through a 0.45 μm filter and then injected onto a C₁₈ column (4.6 mm × 25 cm, 5 μm particle size, Supelco Analytical/Sigma-Aldrich) attached to a Shimadzu HPLC with flow-through UV detection. The column was eluted with a 0–100% acetonitrile gradient at a flow rate of 0.75 mL/min over 31 min. Solvent A was water with 0.075% acetic acid, and solvent B was acetonitrile with 0.075% acetic acid. The multistep gradient was 0 to 15% B over 3 min before increasing to 45% B over 17 min and then to 100% B over 4 min, holding at 100% B for 3.5 min, and then returning to 0% B and equilibrating over 3 min. The absorbance of FA was monitored at 312 nm, its known UV max.³⁰ The substrate (MA) and internal standard, 7-ethoxycoumarin, also showed absorbance at 312 nm.

The concentration of FA formed in reaction mixtures was quantified with a standard curve prepared with known concentrations of FA in acetonitrile (0–25 nmol) plotted against the ratio of the peak area of FA to 7-ethoxycoumarin. Peaks monitored included MA (8.9 min), FA (15.6 min), and 7-ethoxycoumarin (25.7 min). The nonenzymatic rate of conversion of MA to FA was quantitated with 100 μg of bovine serum albumin in the reaction mixture (average of 10.13 ± 0.77 nmol/min/mg of protein, mean ± SD, n = 3). The rate of the nonenzymatic formation of FA was subtracted from each measured sample with rat protein to give a rate of nmol of FA per min of incubation per mg of protein.

To determine the effect of acetonitrile on GSTZ1 activity, acetonitrile (10% v/v) was included in the reaction assay mixture with ¹⁴C-DCA as the substrate in separate studies. The inclusion of acetonitrile resulted in the GSTZ1 activity of adult rat liver control cytosol with DCA as the substrate being inhibited by 10.24 ± 2.57% (mean ± SD), n = 3. As a result, a correction factor of 1.1024 was included in the measured activities with MA as the substrate to account for the slight inhibitory effect of acetonitrile.

Western Blot Analysis of Cytosolic and Mitochondrial GSTZ1 Expression.

Rabbit polyclonal antirat GSTZ1 was custom-produced using His-tagged full-length rat GSTZ1 as the antigen (Cocalico Biologicals, Inc., Reamstown, PA). The antiserum was purified using the Pierce Protein A Antibody Purification Kit (Thermo Fisher Scientific, Waltham, MA) before use. Known amounts of cytosolic and mitochondrial protein (5–200 μg) were separated by sodium dodecyl sulfate polyacrylamide gel electro-phoresis using 12% polyacrylamide gels and transferred by electro-phoresis onto nitrocellulose membranes (Bio-Rad Laboratories, Hercules, CA) as previously documented.³⁵ The measured GSTZ1 activity guided the amount of protein used to stay within the standard curve. The membrane was incubated with this custom GSTZ1 antibody diluted 1:1000 and then with horseradish-peroxidase-labeled goat anti-rabbit secondary antibody diluted 1:5000 (GE Healthcare, Chalfont St. Giles, UK). Bands were visualized using ECL Plus Western Blotting detection reagents (Thermo Fisher Scientific, Waltham, MA), and protein signals were quantitated on the ChemiDoc MP System (Bio-Rad Laboratories, Hercules, CA).

A standard curve using purified rat GSTZ1 was linear from 0.50–10 ng of GSTZ1. To quantify samples, 5 μg of a designated young rat liver cytosol sample was included as a reference control on each blot. Linear regression was used to calculate the GSTZ1 protein expression as ng of GSTZ1 per μg of reference protein. The limit of quantitation was determined to be 0.66 ng of GSTZ1. Samples with protein bands below this limit of quantitation were transformed to a value equal to the limit of quantitation divided by the square root of 2 (0.47 ng).

GSTZ1 Inactivation Half-Life in the Presence of Physiological Concentrations of Cl⁻.

We quantitated Cl⁻ concentrations in rat samples by monitoring the conversion of pentafluorobenzyl bromide to pentafluorobenzyl chloride via HPLC as previously described.²⁵ The detailed results for the rats used in the present study were published separately.²⁶ GSTZ1 inactivation was determined by incubating cytosolic or mitochondrial protein in the presence of a physiologically measured Cl⁻ concentration (n = 3 per subcellular fraction and age), similar to a published procedure.²⁴ In short, replicate samples of dialyzed cytosol and mitochondrial fractions from control rat liver (0.5–1.8 mg of protein) were incubated in the presence of KCl (44 mM in young and adult cytosol, 1 mM in young mitochondria, and 2 mM in adult mitochondria) with 2 mM DCA, 5 mM GSH, and 0.1 M potassium phosphate buffer (pH 7.4) in a final volume of 0.5 mL for 0, 0.5, 1, 1.5, 2, 2.5, or 3 h (cytosol) or 0, 10, 20, 30, 40, 50, or 60 min (mitochondria). Recovered protein from each sample was freed of small molecules by centrifugal filters and assayed for GSTZ1 activity with ¹⁴C-DCA as described above.

Attenuating DCA-Induced GSTZ1 Inactivation by Cl⁻.

The effective Cl⁻ concentration (EC₅₀) providing 50% protection of GSTZ1 activity from inactivation was identified after incubation with a range of KCl concentrations for 45 min. For cytosol and mitochondria, we studied young and adult control rats (n = 3/group). The reaction assay included dialyzed protein (0.5–1 mg) with 2 mM DCA, 5 mM GSH, and several KCl concentrations (0–2 M in cytosol; 0–0.5 M in mitochondria) in a final volume of 0.5 mL. After the 45 min incubation, GSTZ1 activity was determined with recovered protein.²⁴

Reduced (GSH) and Oxidized Glutathione (GSSG) Levels in Liver and Hepatic Subcellular Fractions.

GSH and GSSG concentrations were measured by a modification of the method of Squellerio et al.³⁷ Weighed samples of whole rat liver (75–100 mg) were minced and homogenized with 400 μL of a solution containing 10% trichloroacetic acid (TCA), 1 mM ethylene diamine tetra-acetic acid, and 1 mM butylated hydroxytoluene in a 1.7 mL microcentrifuge tube. Samples of cytosolic and washed mitochondrial fractions (100 μL) were similarly homogenized with 400 μL of the TCA solution. The internal standard ¹³C-GSH (GSH-IS), 10 μL, 1 μM final concentration, was added to each tube. The tubes were centrifuged for 10 min to sediment protein, and 200 μL of the supernatant was transferred to a microcentrifuge 0.45 μm filter tube. The samples were centrifuged for an additional 10 min, and the filtrates were transferred to amber 1.5 mL autosampler vials fitted with 400 μL glass inserts.

LC–MS analysis was performed on a Thermo Scientific TSQ Quantum Access Max mass spectrometer equipped with electrospray ionization and operated in the multiple reaction monitoring mode with positive ion detection. GSH (molecular weight = 307.32) ions monitored were m/z 76.314, 84.235, and 161.90. GSH-IS (molecular weight = 310.3) monitored product ions were m/z 118.2 and 165. Ions monitored from oxidized GSH (GSSG, molecular weight = 612.63) were of the m/z 234.6 and 355.1. Collison energies were optimized for each transition.

GSH, GSH-IS, and GSSG were separated on a Phenomenex Luna 5 μm PFP(2) 50 × 2.00 mm column (Phenomenex Incorporated, Torrance, CA) with a mobile phase of 99% 0.75 mM ammonium formate/formic acid (pH = 3.5) and 1% methanol (isocratic) and a flow rate of 200 μL/min. GSH and GSH-IS coeluted (retention time = 4.1 min) under these conditions, but unique product ions allowed for independent quantitation. GSSG eluted at 9.1 min.

DCA Analysis.

Samples of whole liver, liver cytosol, and liver mitochondria were analyzed for DCA, as described previously.³⁸ Briefly, known amounts of liver tissue, cytosol, or mitochondria were homogenized with 0.9 mL of 5% (w/v) meta-phosphoric acid in a 1.7 mL microcentrifuge tube. Trichloroacetic acid was not used to denature proteins due to the possibility of trace levels of DCA, which could be present in the concentrated trichloroacetic acid solution. The tubes were centrifuged at low speed for 10 min, and 100 μL of the supernatant was transferred to a 7 mL glass culture tube lined with a PFTE cap for derivatization to the methyl ester. GC–MS analysis and quantitation were as described.³⁸

Data Analysis.

Various parameters from young and adult control rats were plotted against time to examine data for possible diurnal effects, and points were connected by cubic spline using GraphPad v6 software (GraphPad Software Inc., San Diego, CA). The in vivo and in vitro half-lives of GSTZ1 activity loss/inactivation and the Cl⁻ concentration that protected against 50% of DCA-induced GSTZ1 inactivation were determined using previous methods.^22,24 The chloride EC₅₀ calculation was performed by normalizing GSTZ1 activity as a percent against the GSTZ1 activity determined with the highest KCl concentration. Treatment-related comparisons of GSH and GSSG levels and the ratio of GSH:GSSG were analyzed by Student’s t-test for young and adult rat whole liver, cytosol, and mitochondria.

Means, standard deviations, and statistical significance were calculated and analyzed by Excel (Microsoft Office, Redmond, WA) and GraphPad Prism v6. Comparisons between age groups and subcellular fractions used either one-way ANOVA with Tukey’s multiple comparison tests or a two-tailed Student’s t-test. A p-value of less than 0.05 was considered to be statistically significant.

RESULTS

DCA Rapidly Inactivates GSTZ1 Activity in Rats.

Oral administration of DCA, 100 mg/kg, to young or adult rats led to a rapid depletion of both cytosolic and mitochondrial GSTZ1 activity with DCA as substrate. Activity in control rats was 5- to 7-fold higher in the cytosol than in the mitochondria (Tables S1 and S2). The mean cytosolic activity in all control rats did not differ between the young and adult animals when comparing activity in young and adult control rats. Mitochondrial activity was significantly higher in adult than young control rats, p < 0.0001. Activity in adult mitochondria was 0.49 ± 0.08 nmol of glyoxylate/min/mg of protein (mean ± S.D., n = 40), and activity in young mitochondria was 0.30 ± 0.07 nmol of glyoxylate/min/mg of protein (n = 42).

To examine the rate of loss of GSTZ1 activity following oral DCA administration, enzyme-specific activities in cytosol and mitochondria for individual young and adult rats were expressed as the percentage of GSTZ1 activity remaining, compared to the average activities in control rats at the pertinent time point, for each subcellular fraction (Figure 1). GSTZ1 activity decreased over 90% in the cytosol and mitochondria of both young and adult animals within 4 h of DCA administration. After 4 h, most mitochondrial activities in both young and adult rats were below limits of quantitation, so data were transformed as described in Materials and Methods. No significant recovery of GSTZ1 activity occurred by 24 h in the cytosol or mitochondria of either young or adult rats.

Figure 1.

Residual GSTZ1 activity after a single dose of DCA in young (A) (cytosol, black closed circle; mitochondria, red open circle) and adult (B) rats (cytosol, blue closed square; mitochondria, open gray square). Data are expressed as the mean ± SD percentage of the average control activity per time point in each subcellular fraction (n = 6/group).

In Vivo Half-Life of Loss of GSTZ1 Activity with DCA Is More Rapid in Mitochondria than in Cytosol.

In DCA-treated rats, log–linear loss of GSTZ1 activity occurred through 4 h in the cytosol and 2 h in the mitochondria. Activity data were transformed by computing the fraction of control activity based on the total average control activity within each age group’s subcellular fraction to a natural logarithmic ratio. The time (t) at which 50% activity remained was determined by fitting the data to the equation ln(A/A_o) = −k_obst, where A is the measured activity at each time point, A_o is the average control activity per age and subcellular fraction, and k_obs is the slope of the line (Figure 2). The half-life (t_1/2) was determined from the equation t_1/2 = ln 2/k_obs. The half-lives of the loss of GSTZ1 activity in young rat cytosol and mitochondria were 1.05 ± 0.03 h (mean ± SD, n = 6) and 0.57 ± 0.06 h, respectively. From adult rats, the cytosolic half-life was 0.82 ± 0.02 h, and the mitochondrial half-life was 0.54 ± 0.05 h. One-way ANOVA with Tukey’s post-test showed the mitochondrial half-lives were significantly shorter than the cytosolic half-lives for young and adult rats, (p < 0.0001). The half-life of loss of cytosolic GSTZ1 was shorter in adult rats compared to young rats (p < 0.05); however, the rates of inactivation of mitochondrial GSTZ1 were similar in rats of each age.

Figure 2.

In vivo half-life of loss of rat GSTZ1 activity following treatment with DCA. 50 percent of GSTZ1 activity remaining occurs when ln(A/A_o) equals ln(0.5), depicted by the black dashed line ln(A/A₀) = −0.693. Data for young and adult rat cytosol are shown in black closed circles and blue closed squares, respectively. Results for mitochondria are shown with red open circles (young) and gray open squares (adult). Values are shown as mean ± SD, n = 6.

DCA Inhibits Cytosolic GSTZ1/MAAI Activity with MA as the Substrate.

MA was rapidly isomerized to FA in control rat liver cytosol. Rates in young control rats were 212 ± 7 nmol of FA formed/min/mg of protein (mean ± S.D., n = 3) and in adult controls were 275 ± 21 nmol of FA formed/ min/mg of protein, compared with rates with DCA of 2.33 ± 0.19 (young) and 2.28 ± 0.16 nmol of glyoxylate/min/mg of protein (adult) in the same three samples. DCA treatment of young and adult rats led to a rapid loss of cytosolic GSTZ1 activity with the endogenous substrate, MA, over the time-course studied. Three of the six liver cytosol samples at each time point were studied with MA, due to the time-intensive synthesis and low yield of the MA. Guided by our results with DCA as the substrate, we used more protein in assays with MA at times when activity with DCA was greatly reduced. Time points after 4 h in both ages required up to 2 mg of protein for detectable formation of FA.

Specific activities for each young and adult cytosolic sample showed rapid reduction with increasing time after the DCA dose. For rat cytosolic activities with MA as the substrate, less than 10% of activity remained by 4 h compared to controls, similar to when DCA was the substrate (Figure 3).

Figure 3.

Specific GSTZ1 activity with MA as the substrate following treatment with DCA was determined in young (black closed circle) and adult (blue closed square) rat cytosol. Values are mean ± SD, n = 3.

GSTZ1 Protein Expression Is Higher in Cytosol than in Mitochondria.

To quantitate the protein expression levels of rat GSTZ1, a standard curve was prepared with 0.50 to 10 ng of purified rat GSTZ1, and a young rat reference control cytosol sample, 5 μg, was included in the standard curve and each analyzed blot. The standard curve was repeated three times with similar results. Linear regression analysis resulted in an R² of 0.99, and the relationship was used in computing the cytosolic and mitochondrial GSTZ1 protein levels of DCA-treated and control rats.

Relative protein levels of GSTZ1 were at least 5 times higher in the cytosol than mitochondria for both DCA-treated and control rats, p < 0.0001 (Tables S3 and S4). After DCA treatment, cytosolic and mitochondrial GSTZ1 protein expression was quickly lost in both young and adult rats (Figure 4). Similar to the measured GSTZ1 activities, little protein expression was found after 4 h in the cytosol and mitochondria and was at least 75% less than the average control GSTZ1 expression levels in respective subcellular locations (Figure 4). No recovery of GSTZ1 expression was observed by 24 h in the cytosol or mitochondria for either age.

Figure 4.

GSTZ1 expression remaining over a 24 h period after a single dose of DCA. (A) Young rat GSTZ1 expression loss (cytosol, black closed circle; mitochondria, red open circle). (B) GSTZ1 expression loss in adult rats (cytosol, blue closed square; mitochondria, open gray square). All values are shown as a percentage of the average control protein expression level per time point in each subcellular fraction, mean ± SD, n = 6.

Hepatic GSTZ1 Activity and Protein Expression Are Highly Correlated.

As expected, GSTZ1 activity and protein expression correlated well for each of the subcellular fractions and ages (Figure S1). Generally, with increasing GSTZ1 expression, there was an increase in the measured activity. The computed Pearson r-values for the young rat cytosol and mitochondria were 0.91 and 0.90, respectively. For the adult rat hepatic fractions, Pearson r-values were 0.91 for cytosol and 0.86 for mitochondria.

Cl⁻ Reduces the in Vitro Rate of GSTZ1 Inactivation by DCA.

Using the measured physiological concentrations of Cl⁻ in young and adult rat cytosol (44 mM) and mitochondria (1 mM for young and 2 mM for adult),²⁶ we determined the in vitro half-life of GSTZ1 inactivation with incubations of control rat samples for both age groups and subcellular fractions with 2 mM DCA. Previous reports have determined that the half-life of GSTZ1 inactivation in the absence of Cl⁻ was 5.44 min in the cytosol.²² We also found the half-life of GSTZ1 inactivation in the absence of Cl⁻ in rat liver cytosol and mitochondria to be very short, 5.94 ± 0.09 min in cytosol and 7.90 ± 0.15 min in mitochondria, mean ± SD, n = 3. In the presence of physiologically relevant Cl⁻ concentrations, cytosolic half-lives of inactivation were determined to be similar for the two rat age groups, 1.74 ± 0.13 and 1.60 ± 0.07 h (mean ± SD, n = 3) for the young and adult rats, respectively (Figure 5). The inactivation half-lives in the mitochondria were similar between ages, being 0.46 ± 0.03 h in the young rats and 0.51 ± 0.03 h in the adult rats. ANOVA with Tukey’s multiple comparisons showed the mitochondrial inactivation half-lives were significantly shorter than those for the cytosol, p < 0.0001.

Figure 5.

In vitro half-life of rat GSTZ1 inactivation following incubation with 2 mM DCA was determined in the presence of physiologically relevant Cl⁻. 50% of GSTZ1 inactivation after DCA treatment occurs when ln(A/A_o) equals ln(0.5), depicted by the black dashed line. Data for cytosol are black circles (young) and blue squares (adult), and data for mitochondria are red open circles (young) and gray open squares (adult). Values shown as mean ± SD, n = 3 per age and subcellular fraction.

In a concentration-dependent manner, Cl⁻ reduced the extent of GSTZ1 inactivation when cytosol was incubated with 2 mM DCA for 45 min prior to measurement of activity (Figure 6). With KCl concentrations of at least 1 M, full protection of GSTZ1 activity from inactivation by DCA occurred. Stability studies over a wide range of KCl concentrations showed the cytosolic Cl⁻ EC₅₀ was similar for young and adult rats, 81.8 ± 13.5 mM (mean ± SD, n = 3) and 84.0 ± 4.6 mM, respectively. Similar EC₅₀ values were found in mitochondria (data not shown).

Figure 6.

Effect of Cl⁻ concentration on the stability of rat hepatic cytosolic GSTZ1 in the presence of 2 mM DCA. The concentration range of KCl was 0–2 M. The cytosolic EC₅₀ values were not different between young (black squares) and adult (blue circles) rats. The Cl⁻ EC₅₀ to protect against 50% of GSTZ1 inactivation was determined for young and adult cytosol (n = 3 per age), where data points are shown as mean values of duplicate determinations for each sample, and lines are curves fit for each individual sample. The dashed black line represents when the Cl⁻ concentration prevented 50% of GSTZ1 activity from being lost from DCA treatment.

Single Dose of DCA Does Not Alter GSH or GSSG Concentrations in Whole Liver, Hepatic Cytosol, or Hepatic Mitochondria.

Oxidized and reduced GSH levels were determined for rat cytosol, mitochondria, and whole liver, and GSH:GSSG levels were computed. Compared with controls, the single dose DCA treatment did not result in significant differences in GSH or GSSG concentrations in the liver or subcellular fractions (Figure 7). There was no diurnal pattern for GSH or GSSG levels in livers, cytosol, or mitochondria of adult control rats. In young rats, GSH but not GSSG levels in liver varied with time, being significantly higher at 8:30 am (0.5 h time point) than all other sampling times except 4 pm (8 h time point). Table 1 summarizes the results for all control and all DCA-treated young and adult rats. Mean GSH concentrations in young rat liver were higher than in adult rat liver. Similarly, young rats had higher levels of GSH in the cytosol and mitochondria than adult rats, whereas adult rats had higher levels of GSSG in cytosol and mitochondria than young rats. These differences led to age-related changes in the ratio of GSH to GSSG, which was higher in the cytosol and mitochondria of treated and control young rats compared to adult rats (p < 0.0001). Another notable difference was that the mitochondrial GSH to GSSG ratio generally was at least 4-fold lower than the cytosolic ratio for both age groups (p < 0.0001).

Figure 7.

Hepatic GSH and GSSG measurements in whole liver, cytosol, and mitochondria of control and DCA-treated rats. Data are mean ± SD, n = 6 per treatment and time point. All panels show GSH in control (blue squares) or DCA-treated (red circles) rats and GSSG in control (blue inverted triangles) or DCA-treated (red triangles) rats. Data for young rats are shown in panels (A) (whole liver), (C) (cytosol), and (E) (mitochondria), and data for adult rats are shown in panels (B) (whole liver), (C) (cytosol), and (F) (mitochondria).

Table 1.

Measured GSH and GSSG Levels in Rat Liver Cytosol and Mitochondria

age/fraction treatment		GSH (nmol/mg)	GSSG (nmol/mg)	GSH:GSSG
young cytosol	DCA-treated control	82.0 ± 17.1****	1.85 ± 0.59	48.7 ± 15.1****
		79.9 ± 18.9****	2.05 ± 0.54	40.5 ± 11.2*
adult cytosol	DCA-treated control	47.9 ± 12.5	2.53 ± 0.36	19.3 ± 5.9
		57.4 ± 11.5	2.40 ± 0.44	24.9 ± 7.6
young mitochondria	DCA-treated control	4.31 ± 1.25*	0.40 ± 0.09*	10.5 ± 2.6****
		4.34 ± 1.18*	0.39 ± 0.10*	11.1 ± 1.4****
adult mitochondria	DCA-treated control	3.74 ± 1.29	1.01 ± 0.29	3.9 ± 1.3
		3.84 ± 0.89	0.94 ± 0.25	4.3 ± 1.6
young whole liver	DCA-treated control	8.63 ± 2.97****	0.23 ± 0.11*	44.9 ± 25.8
		8.17 ± 3.04****	0.19 ± 0.07*	51.0 ± 34.3
adult whole liver	DCA-treated control	3.91 ± 0.63	0.08 ± 0.02	52.6 ± 13.2
		3.95 ± 0.64	0.08 ± 0.02	51.9 ± 11.1

^aValues presented are mean ± SD, n = 44 for young rat DCA-treated measurements, n = 42 for young rat control measurements, n = 48 for adult rat DCA-treated measurements, and n = 40 for adult rat control measurements. The detailed time-point data that are summarized in Table 1 are shown in Figure 7. The data were analyzed to test for age-related differences between fraction treatments by Student’s t-test,

^*p < 0.05,

^****p <0.0001.

DCA Concentrations in Liver Are Inversely Associated with Age.

DCA was detected in samples of whole liver, cytosol, and mitochondria up to 4 h after the dose (Figure S2), with maximal concentrations attained in the whole liver of both young (11.2 ± 3.0 ng/mg liver, mean ± S.D., n = 6) and adult (16.5 ± 4.0 ng/mg liver) rats at 1 h after the dose. At 2 h after the dose, young rats retained significantly less DCA in the liver than adult rats. DCA was present in the mitochondria as well as the cytosol.

Diurnal Variation of Studied Parameters.

In control rats, the possibility of a diurnal effect was considered in the liver to body weight ratios and the young and adult rat cytosolic and mitochondrial GSTZ1 activity and expression levels (Figures S3–S5). No apparent diurnal variation was noted with cubic spline analysis for the liver to body weight ratio in either the young or adult rats (Figure S3). The liver was a significantly larger percentage of body weight in young rats (5.53 ± 0.49%, mean ± S.D., n = 42) than in adult rats (3.50 ± 0.35%; p < 0.0001). Plots over the 24 h period for GSTZ1 activity with DCA and protein expression levels in cytosol and mitochondria in young and adult control rats are shown in Figures S4 and S5. In adult rats, activity and expression rose between 8:30 am (0.5 h after the dose) and 10 am (2 h after the dose); otherwise, no striking variation with time was evident.

DISCUSSION

Because of its location at the site of DCA’s therapeutic action, gaining further knowledge of the properties of hepatic mitochondrial GSTZ1 is important to understand the biotransformation and therapeutic benefits of DCA. In this study, adult and juvenile female S-D rats were used as an experimental model to investigate the inactivation of hepatic cytosolic and mitochondrial GSTZ1 in the 24 h following after a single dose of a therapeutic concentration of DCA.

In control rats of both ages, GSTZ1-specific activities and measured GSTZ1 protein expression were greater in the cytosol (at least 80% of total GSTZ1 activity) than mitochondria, as observed previously.¹⁷ Mitochondrial GSTZ1-specific activities of controls were higher in adult than young rats, whereas cytosolic GSTZ1-specific activities were similar at the two ages. In humans, cytosolic GSTZ1 attained adult levels by age 7 years, whereas mitochondrial GSTZ1 continued to increase up to 21 years.^27,35 The higher mitochondrial GSTZ1 activity in adult rats may reflect a similar developmental control. Protein yields for mitochondria and cytosol were not age-dependent (data not shown); hence, the larger size of young rat liver, relative to body weight, meant that the total hepatic mitochondrial GSTZ1 activities were similar at the two ages, whereas cytosolic activities were higher in young rats, on a body weight basis. This suggests that an oral dose of DCA would be metabolized to a greater extent in hepatic cytosol by young rats or in children compared with adult rats or humans.

After DCA treatment, GSTZ1 activity rapidly decreased in both subcellular fractions to a nadir of less than 5% of controls by 8 h after dosing, regardless of age. With increasing time after dose, enzyme activity remained low and did not show any statistically significant recovery by 24 h. A similar pattern was found for GSTZ1 protein expression levels after DCA treatment. This is expected, since the measured GSTZ1 activity with DCA as the substrate is strongly related to the quantitated expression levels (Figure S1). A prior study of 200 to 225 g male Fischer rats administered 38.7 mg/kg intraperitoneally showed that hepatic cytosolic activity and expression were reduced to 35% of control at 12 h after the dose but did not recover to predose values until 8 days after the single dose.²⁰ Mitochondrial activity was not measured. Another study of 2 month old male S-D rats indicated that hepatic cytosolic GSTZ1 activity was significantly decreased in a dose-dependent manner by DCA in the drinking water, and at the highest dose (50 mg/kg for 8 weeks), cytosolic activity had not recovered to control levels 1 week after ceasing the DCA.³⁹ Modeling rat pharmacokinetic data confirmed that DCA rapidly reduced its own clearance and predicted the maximum inactivation rate constant of GSTZ1 was 0.96 per h.⁴⁰

GSTZ1 activity with an endogenous substrate, MA, also was reduced after treatment with DCA. By 8 h after a single dose of DCA, cytosolic and mitochondrial activities with MA as the substrate were less than 1% of controls. Previous work with male Fischer rat liver cytosol and with human expressed recombinant GSTZ1 variants reported that specific activity for the isomerization of MA is 100- to 1000-fold greater than dechlorination of DCA or of chlorofluoroacetic acid.^18,41 In this study, specific activity with MA as the substrate was approximately 100-fold higher than with DCA as the substrate in rat cytosol samples, presumably because GSTZ1 evolved as an efficient catalyst for isomerization of MA and MAA. Loss of isomerase activity would be predicted to result in buildup of MA and MAA, as has been observed in DCA-treated animals and people.^21,23,42 It has been speculated that buildup of MA and MAA may be responsible, at least in part, for DCA-associated peripheral neuropathy.¹¹

A key finding of this study was that in rats of both ages, the half-lives of the loss of cytosolic and mitochondrial GSTZ1 activity after DCA treatment were markedly different. While activity declined in both fractions over the first 4 h, the loss of GSTZ1 activity in the mitochondria was more rapid, having a half-life of loss of GSTZ1 activity approximately half that in the cytosol (Figure 2). It does not appear that the more rapid loss in mitochondria results from higher DCA concentrations in mitochondria than cytosol, as GC–MS analysis showed that DCA concentrations in mitochondria were similar to or slightly less than those in cytosol (Figure S2). Considering only the influence of GSTZ1 activities, it might be hypothesized that DCA concentrations would be higher in mitochondria than cytosol and higher in mitochondria of young than adult rats; however, DCA concentrations in mitochondria were similar to or lower than in cytosol and were lower in young than adult rats. DCA concentrations in both liver cells and mitochondrial organelles are likely to be influenced not just by the rate of metabolism but also by the actions of transporter proteins, which were not studied here.

The factor most likely to influence the rate of DCA-induced GSTZ1 inactivation is physiological Cl⁻ concentration. Previous studies with human liver cytosol showed that Cl⁻ protected GSTZ1 from inactivation by DCA with an EC₅₀ of 15 to 35 mM depending on the human GSTZ1 haplotype.²⁴ Human liver mitochondrial Cl⁻ concentration (mean of 4.2 mM) is about 25-fold lower than in the cytosol.²⁵ In the rats used in this study, mitochondrial Cl⁻ levels (0.4–2.2 mM) were considerably less than measured cytosolic Cl⁻ (43–45 mM).²⁶ We hypothesized that the observed differences between subcellular fractions in the rate of loss of GSTZ1 expression following DCA are likely to be due to the lower [Cl⁻] in rat liver mitochondria, making GSTZ1 more susceptible to rapid inactivation.

We tested this hypothesis by conducting in vitro studies with rat liver fractions. We found that the absence of Cl⁻ in cytosolic and mitochondrial samples led to very short half-lives of only a few minutes, as shown previously for rat liver cytosol.²⁰ Upon adding physiologically relevant Cl⁻ concentrations to assay mixtures in vitro, the half-life of loss of GSTZ1 activity was prolonged to a greater extent in cytosol than mitochondria (Figure 5). Comparing these in vitro half-lives to the estimated in vivo half-lives for loss of GSTZ1 activity in DCA-dosed rats, mitochondrial t_1/2 values were approximately 0.5 h in both circumstances. The cytosolic half-lives determined in vitro were somewhat longer with the addition of 44 mM Cl⁻ than determined following dosing in vivo, particularly for the adult rat. This could possibly be due to local fluctuations of Cl⁻ in liver cytosol that could affect the rate of GSTZ1 inactivation. The finding that mitochondrial GSTZ1 is more rapidly inactivated than cytosolic GSTZ1 is consistent with the explanation that the much lower physiological levels of Cl⁻ in the mitochondria (0.4–2.2 mM) hastened the rate of DCA-induced GSTZ1 inactivation.

The in vitro studies with liver cytosol from young and adult rats showed that Cl⁻ protected GSTZ1 from DCA-induced inactivation with EC₅₀ values of 81.8 and 84 mM, respectively (Figure 6). These values are higher than found for human liver cytosol, which gave EC₅₀ values of 15–35 mM. Thus, the rat GSTZ1 requires more Cl⁻ for protection from inactivation than the human GSTZ1. This property suggests that the rat enzyme is more rapidly inactivated by DCA than the human under normal physiological conditions, as we have shown that rat liver cytosolic Cl- concentration is 42–45 mM, compared with 105 mM in human liver cytosol.²⁵

These subcellular fraction differences in inactivation may involve the cosubstrate of the interaction between DCA and GSTZ1, GSH. Li et al.¹⁷ noted that GSTZ1 in the mitochondrial matrix of rat hepatic cells exhibited similar K_m values for DCA to that in the cytosol, but there was a larger apparent K_m for GSH in the mitochondria compared to in the cytosol (0.19 mM in cytosol, 0.50 mM in mitochondria, p < 0.01 by one-tailed t-test). However, in both cytosol and mitochondria, GSH is present at concentrations far above that of the measured K_m value (Table 1), so the rate of DCA biotransformation should not be limited by GSH concentration in either subcellular fraction under physiological conditions.

To gain further insights into the effects of DCA, we measured reduced and oxidized levels of GSH in rat whole liver, cytosol, and mitochondria. It has been shown that GSTZ1 knockout male mice have different GSH and GSSG contents and ratios than normal mice,⁴³ and because DCA chemically “knocks out” GSTZ1, we sought to determine if a similar effect would be observed after a single drug dose. We measured lower GSH levels in rat liver mitochondria than in cytosol. In rats, about 10% of total hepatic GSH is found in the mitochondria, with about 90% of that accounting for reduced GSH.⁴⁴ One study showed that the rat mitochondrial GSH pool is about 5 nmol/mg of protein, while the cytosolic pool for GSH is more than 20 nmol/mg of protein.⁴⁵ Our study also showed that mitochondrial GSH levels were considerably lower than cytosolic levels (Table 1).

It has been hypothesized that the lower GSH levels found in GSTZ1 knockout mice are linked to greater oxidative stress and susceptibility to acetaminophen toxicity.⁴³ Interestingly, treatment of GSTZ1 knockout mice with DCA increased the expression of glutamate-cysteine ligase, an enzyme of GSH synthesis.⁴⁶ Therefore, it has been suggested that DCA can provide a protective role from hepatotoxicity produced by chemicals that cause oxidative stress or consume GSH.⁴⁶ We found no difference in rat GSH levels in control or DCA-treated rats (Table 1 and Figure 7). Others reported that treatment of female S-D rats with DCA (2.45 mmol/kg) before chloroform treatment had no impact on hepatic GSH levels.⁴⁷ The studies with male GSTZ1 knockout mice are not directly comparable to the single dose studies reported here, as it is likely that prolonged loss of GSTZ1 is needed to influence GSH or GSSG levels. Alternatively, this may be a species-specific response or controlled by other unknown factors. This study showed there were age-related differences in GSH concentrations in cytosol and whole liver, where young rats had higher mean concentrations of GSH than adults. This agrees with reports in the literature that aging and disease can reduce GSH levels.^45,48 Our results showed a drop in GSH concentrations between 8:30 am and 12 pm followed by a rise at 4 and 8 pm in the liver of female rats aged 4 weeks but not 52 weeks (Figure 7). This differs from a report with male Fischer rats aged 3 to 4 months, which showed lower GSH concentrations at 5 and 9 pm than at 9 am.⁴⁹ The reasons for these differences are not clear.

Our results showed that age can be a factor in the rate of GSTZ1 inactivation by DCA in rat liver. In this study, we utilized two groups of female S-D rats, aged 4 and 52 weesk, to model children and midlife adults, respectively. Differences in the in vivo half-life of loss of GSTZ1 activity in the cytosol, liver to body weight ratios, mitochondrial Cl⁻ levels, and the GSH content of liver and cytosol were found between young and adult rats. In humans, age is an important determinant of DCA pharmacokinetics, particularly for the clearance of DCA after chronic exposure in patients suffering from a variety of genetic mitochondrial diseases.²³ In adults, DCA administration causes a large decrease in DCA plasma clearance and a 10-fold increase in the elimination half-life, while children only had a 3- and 4-fold increase in the same pharmacokinetic parameters after DCA administration, respectively. These findings most directly relate to the much shorter half-life of loss of cytosolic GSTZ1 activity in the liver of adult rats: if, as seems likely, similar differences occur in people, cytosolic GSTZ1 will be lost more rapidly in older patients treated with DCA. There is more GSTZ1 in cytosol than mitochondria, thus its more rapid inactivation in adults will lead to greater accumulation of DCA as well as the tyrosine catabolites MA and MAA if similar doses are used, consistent with observations in people.²³ As mitochondrial GSTZ1 is inactivated very rapidly by DCA in both young and adult rats, it likely plays a smaller role in the age-related differences in DCA pharmacokinetics and toxicity observed in people. The difference in GSTZ1 inactivation could possibly explain the variability in the frequency and severity of peripheral neuropathy severity with age, where the slower clearance of DCA from adults could be an underlying cause of this reversible neurological effect.

In conclusion, this study examined the time-course for loss of rat GSTZ1 activity and expression after a single dose of DCA and showed that the rate of GSTZ1 inactivation differs in rat liver cytosol and mitochondria. Rates of inactivation in cytosol were also significantly influenced by age. Our most striking finding was that mitochondrial GSTZ1 was lost more rapidly than cytosolic GSTZ1 in livers of young and adult female S-D rats. We propose this result is due in large part to the much lower concentration of Cl⁻ in the mitochondria than the cytosol.²⁶ Our results provide new insight into the role of the mitochondrial compartment in DCA biotransformation.

ABBREVIATIONS

Cl⁻

chloride

DCA

dichloroacetate

FA

fumarylacetone

GSH

glutathione

GSH-IS

¹³C-GSH internal standard

GSSG

oxidized glutathione

GSTZ1

glutathione transferase ζ1

PDC

pyruvate dehydrogenase complex

PDK

pyruvate dehydrogenase complex kinase

MA

maleylacetone

S-D

Sprague–Dawley