Effect of Sulfur Amino Acid Depletion and Acetaminophen on Plasma Glutatione

The objectives of this research are to evaluate toxicological perturbations of metabolism due to complex chemical mixtures and diet-chemical interactions using high-resolution 1H-NMR. These experiments are designed to provide data on the extent of normal variation in metabolic profiles within healthy individuals and, in combination with key GSH redox measurements, will show whether changes due to chemical exposure and diet can be discriminated. The specific aims of the current protocol are: 1. To determine whether SAA-free diet (2 days) and APAP (2 doses at 15 mg/kg) independently alter SAA homeostasis and metabolism and thiol-disulfide redox state (GSH/GSSG and Cys/CySS). 2. To determine whether 2 doses of APAP interact with SAA-free diet (2 days) in effects on plasma GSH concentration and redox state. 3. To use 1H-NMR spectroscopy to determine whether metabolic changes induced by the combined exposure to APAP and SAA-free diet are quantitatively or qualitatively different from that induced by either alone. A. BACKGROUND AND SIGNIFICANCE: Sulfur amino acids are involved in central metabolic processes. The sulfur-containing amino acids, methionine and cysteine, are required for diverse, critical biologic functions, including protein synthesis, methylation, fatty acid metabolism, osmotic regulation and regulation of cell division and growth (1-5). Methionine (Met) is an essential amino acid and is metabolized in individuals by the transulfuration pathway to form cysteine (Cys) (6). In addition to use in the primary sequence of most proteins, both Met and Cys are required for other metabolic functions. Met is converted to S-adenosylmethionine, which is used for methylation reactions (1) for structural and functional modifications of proteins, RNA and DNA, as well as synthesis of phospholipids and signaling molecules. Cysteine is used for biosynthesis of glutathione (GSH) and sulfate (2). GSH functions in redox regulation (7) and detoxification of oxidants and reactive electrophiles (8); sulfate is used as a structural component of oligosaccharides (3), transport of steroid hormones (4) and detoxification of foreign compounds (9). Because the SAA are irreversibly modified in the aforementioned processes, there is a requirement for adequate sulfur amino acid intake that extends beyond the need for adequate amounts to maintain normal protein synthesis and turnover. Sulfur amino acid intake in humans is variable and optimal SAA intake has not been adequately defined. The Recommended Dietary Allowance for SAA is based upon nitrogen and sulfur amino acid balance studies and for an adult male is about 1 g (210 mg Met plus 800 mg Met or Cys) (10,11). The average American diet contains about 100 g/d of protein (2); the mean SAA intake is about 2.4 g, but there is a large variation, ranging from \< 0.3 g to \> 5 g (12). Most of the SAA is derived from animal protein, which constitutes about 2/3 of the protein intake. Except for legumes and some nuts, a gram of plant-derived protein contains only 10-20% of the SAA content of animal protein (13,14). Because plant-derived foods are low in total protein, individuals who do not consume animal protein are at risk of SAA deficiency. The number of Americans with low dietary intake of SAA is relatively small, but most people undergo intermittent periods of SAA deficiency due to food selection, dieting, fasting and illness. The non-protein content of Cys in human liver is about 1 g, largely present in GSH, and approximates the daily SAA requirement (15). Chemical detoxification places a burden upon SAA supply. GSH conjugation and sulfate conjugation represent common and quantitatively important pathways for metabolism of foreign compounds. One example, which is proposed for use in the present study, is acetaminophen (APAP). About 15% of APAP is excreted as the parent compound and about 55% is metabolized by conjugation with glucuronic acid (16-18). The remainder is metabolized by conjugation with sulfate (25%) and GSH (5%). If one considers the metabolism of 4 doses of maximum strength acetaminophen (MW 151), i.e., 4 g daily, will consume 200 mg of Cys during GSH conjugation and 1 g of Cys to supply sulfate for conjugation (15). Without added sulfate or sulfite in the diet, most of the sulfate is derived from dietary SAA (2,19). Thus, the total equivalents of SAA needed for metabolism of the maximum daily dose of APAP is greater than the RDA for SAA. Thus, an interaction between APAP intake and SAA deficiency is expected, and this interaction may be significant even at normal therapeutic dosing and a relatively modest level (e.g., 1 or 2 days) of SAA insufficiency. We propose that this will provide a useful model to investigate chemical interactions in humans and the major metabolites of APAP and Cys are readily measurable so that the converse effects, APAP on SAA metabolism and SAA availability on APAP metabolism, can be determined. Both Cys deficiency and APAP metabolism affect GSH homeostasis, and our clinical assay for GSH redox (7, 20-21) provides a sensitive means to evaluate interactive effects of these 2 experimental manipulations. High resolution 1H-NMR and mass spectrometry provides an approach to investigate complex metabolic effects of diet. Traditional methods to investigate interactive effects of chemical exposure have relied upon detection of specific pathologic and physiologic responses for combinations with specific biologic activities. An alternate approach is to use available high resolution metabolic analyses (22) combined with bioinformatic tools, to examine metabolic effects of chemical interactions in a global manner. In principle, by simultaneously examining effects on hundreds of metabolites, this approach provides the possibility to detect unanticipated interactive effects. A guiding principle for development of high-resolution metabolic analyses is that metabolic profiles in healthy individuals share common features while toxicologic and pathologic consequences of diet, disease, environment or genetics are reflected in variations in that metabolic profile. Studies in rodents show that characteristic patterns of metabolic change are associated with organ-specific toxicity for chemical agents that share no structural similarity (23-24). If this approach is extended to humans to identify healthy and unhealthy metabolic profiles, such patterns could provide a powerful new approach to identify and diagnose toxicities associated with chemical mixtures and diet-drug interactions. This proposal is to conduct a pilot study to determine whether diet-drug interactions can be detected in metabolic patterns of healthy adult subjects. We believe that if the environment is controlled and specific chemical exposures are changed (e.g., +APAP, +dietary SAA), we will observe metabolic changes that can be ascribed to the respective exposures. High resolution 1H-NMR, liquid chromatography, and mass spectrometry (MS) are well suited for investigating changes in body fluid composition because a large number of metabolites are present in body fluids, many metabolites can be detected simultaneously. Examples are available from the studies of Nicholson and coworkers (23, 24).The use of multivariate statistical analysis of NMR data to evaluate metabolic responses of living systems is under development for toxicologic screening of pharmacologic compounds (25). The principles for detection of chemical interactions in free-living humans are the same; in effect, changes in metabolic profile are expected to provide a sensitive means to non-invasively detect conditions of risk for toxicity and pathology. Considerable information is available concerning the assignment of NMR spectra of urine and plasma (e.g., Ref 23-25, 36-37); this information will provide a useful guide to possible metabolites which could vary as a function of SAA intake and APAP exposure. 2-D NMR and other techniques are available for peak identification. However, the goals of this project do not require identification of which chemical species change, only whether changes at specific frequencies can be attributed to variation in SAA intake. Significance. While there is general consensus that chemical interactions pose significant human health risks, identification of such risk remains a largely intractable problem. Detailed metabolic analyses combined with bioinformatic methods provides a potentially powerful approach to identify such risk in terms of perturbed metabolic profile. However, the large variety of foods and drugs consumed by different individuals complicates the already difficult task to distinguish between signal and noise and validate stability and recovery for hundreds of metabolites. The significance of the current proposal is that we will use repetitive measures within individuals under highly controlled environmental and dietary conditions to determine whether exposure to a chemical under non-toxic conditions results in a sufficient perturbation of metabolism to be detected by this promising analytical approach. The results will show whether 1H-NMR has the sensitivity to detect metabolic effects due to variations of chemical exposure and dietary intake that are common among humans. If so, the results will establish that 1H-NMR analysis can be used not only to study consequences of chemical toxicity but also to study risk of toxicity due to chemical interactions that occur under usual occupational and therapeutic conditions. B. PRELIMINARY STUDIES: Use of GSH/GSSG redox to evaluate oxidative stress. This proposal is based upon our long-standing efforts to understand the control and function of the GSH-dependent antioxidant system in humans. The balance between GSH and GSSG is determined by the balance of oxidation due to oxidative stress and the capacity of the NADPH-dependent reductase to reduce GSSG back to GSH. Studies on the interactions of GSH and cysteine pools in human plasma revealed that GSH redox characteristics are suitable for use as an indicator of the balance between oxidative and antioxidant processes (26). In addition, our studies of human cells in culture show that substantial variation in GSH/GSSG redox occurs during the normal life cycle of cells (13, 27-28) and, most central to the present proposal, substantial oxidation can be induced by Cys deficiency (29). These results imply that variation in GSH/GSSG redox may be more than an indicator of oxidative stress, it may be an indicator of the cell survival/cell death balance which is central to tissue homeostasis. With this interpretation, the GSH/GSSG redox reflects underlying tissue health; an oxidized state serves as the platform for toxicity regardless of whether this oxidation is due to a genetic abnormality, nutritional deficiency, inflammation, or chemical exposure. Thus, any chemical exposure that results in GSH oxidation or cysteine utilization should exacerbate effects due to limited cysteine intake. Because a) sulfur amino acid deficiency is associated with oxidation of thiol/disulfide redox, b) sulfur amino acids are required for so many different metabolic processes, and c) so many enzymes are potentially affected by a change in thiol/disulfide redox state, diverse metabolic effects may occur that are not readily predicable from the metabolic profile of cysteine or from the metabolic pathways of chemicals detoxified by cysteine-dependent pathways. Hence, a central question of the current proposal is whether a chemical exposure which increases demand for sulfur SAA interacts with the dietary availability of SAA to produce perturbations of metabolism that can be detected by high resolution 1H-NMR spectroscopy. Clinical studies of GSH/GSSG redox variation. We made the surprising finding that substantial oxidation (approximately 25 mV) was apparent in individuals \>60 compared to \<43 y old (30). We also found that diabetic individuals were about 20 mV more oxidized than similarly aged individuals without known disease. In a follow-up study, we found that GSH/GSSG redox as a function of age was biphasic, with no apparent association with age below 45 y and a linear oxidation with age after 50 y. Preliminary data from our study of diurnal variation in GSH/GSSG redox status show that there was an increase in GSH during the evening and nighttime hours and that the redox state was more reduced during this time period. These results are consistent with rodent studies showing that hepatic GSH undergoes a diurnal variation linked to sulfur amino acid intake (31-32). These results support our central hypothesis that sulfur amino acid intake is a determinant of plasma GSH/GSSG redox state in humans. Effects of chemical exposure on GSH redox in humans. We have performed 2 clinical studies which indicate that chemical exposure alters sulfur amino acid homeostasis as detected by changes in plasma GSH and Cys pools. As indicated above, one of these showed that GSH redox became more oxidized as a consequence of chemotherapy (33). A second study showed that both Cys and GSH pools were oxidized in individuals who smoked cigarettes compared to those who do not (34, 38). Together, these studies indicate that chemical exposures can induce significant changes in Cys and GSH homeostasis. Interactions of GSH and cysteine pools in human plasma. To determine the interactions of the GSH/GSSG and Cys/CySS pools, thiol and disulfide forms were measured in plasma from 24 healthy individuals aged 25-35 (26). In this study, GSH concentration correlated with Cys concentration, but no correlations were observed between GSSG and CySS or between the reduced and oxidized components. The lack of equilibration of these values supports the interpretation that plasma GSH and Cys redox values are dynamic indicators of the systemic balance between oxidative and antioxidant processes (26). Cys deficiency in vitro results in substantial oxidation of GSH/GSSG redox. In a study of Cys deficiency and readdition of Cys in colon carcinoma HT29 cell line, we found that culture in medium with low Cys and CySS resulted in decreased GSH and GSSG with an associated 80 mV oxidation of the GSH/GSSG redox state (29). Upon addition of either Cys or CySS, redox of GSH/GSSG recovered in 1 h while GSH concentration continued to increase over 8 h (33). These results have now been confirmed in Hela cells and normal human retinal pigment epithelial cells (not shown) indicating that variation in Cys availability to cells has a general effect on GSH/GSSG redox state. Oxidation of extracellular thiol/disulfide redox state sensitizes cells to chemical-induced toxicity. Oxidative stress increases expression of death receptor components, Fas and Fas ligand (35). To determine whether changes in thiol/disulfide redox state could affect sensitivity to chemical toxicity, normal human retinal pigment epithelial cells were cultured in media with systematic variations in Cys and CySS concentrations to give a range of redox values found in vivo in human plasma and then treated with the oxidant t-butylhydroperoxide. The results showed that cells cultured at more reduced redox states were more resistant to oxidant-induced apoptosis than were cells cultured at more oxidized redox states. While it is not yet clear how general this effect is with regard to other cell types and other chemicals, the results suggest that the change in redox state, per se, may be an important determinant of susceptibility to toxicity. Together with the above findings, these results provide strong justification for exploring whether an interaction between chemical exposure and diet can induce detectable redox changes and whether diverse metabolic consequences of variation in thiol/disulfide redox state can be detected by high resolution metabolic analysis. High-resolution 1H-NMR of human urine. To examine the sensitivity of our instrumentation to obtain NMR spectra suitable for measurements of metabolic changes in response to varied sulfur amino acid intake, we obtained spectra from urine collected on 2 successive days without and with oral supplementation with 3 g Cys. Samples (300 μl) were mixed with 150 μl of 200 mM sodium phosphate buffer, pH 7.4, and 50 μl 2H2O prior to recording spectra with a Varian Inova 600 MHz spectrometer. Although the data are too preliminary to make conclusions, results demonstrate that spectral quality is similar to that reported by Nicholson and colleagues (23-25, 36-37) and suitable for detection of changes in concentrations of a large number of metabolites. In particular, peaks identified by Nicholson were present and several differences were evident between samples with supplemental Cys and without.