Glutathione Synthesis Is Diminished in Patients with Uncontrolled Diabetes and Restored by Dietary Supplementation with Cysteine and Glycine

Pathophysiology/Complications ORIGINAL ARTICLE

Glutathione Synthesis Is Diminished in Patients With Uncontrolled Diabetes and Restored by Dietary Supplementation With Cysteine and Glycine

1,2 1,2 RAJAGOPAL V. SEKHAR, MD VASUMATHI T. REDDY, PHD polyol pathway, advanced glycation end 3,4 1,2 SIRIPOOM V. MCKAY, MD ASHOK BALASUBRAMANYAM, MD product formation, protein kinase C acti- 1,2 3,4 SANJEET G. PATEL, MD FAROOK JAHOOR, PHD 1,2 vation, and the hexosamine pathway, a ANURADHA P. GUTHIKONDA, MD common feature is increased oxidative stress marked by elevated levels of reactive oxygen species (ROS) (1). The ability OBJECTIVE — Sustained hyperglycemia is associated with low cellular levels of the antiox- of a cell to resist damage caused by oxida- idant glutathione (GSH), which leads to tissue damage attributed to oxidative stress. We tested tive stress is determined by the capacity of the hypothesis that diminished GSH in adult patients with uncontrolled type 2 diabetes is an array of antioxidant defense systems, attributed to decreased synthesis and measured the effect of dietary supplementation with its precursors cysteine and glycine on GSH synthesis rate and oxidative stress. among which reduced glutathione (GSH) is the most ubiquitous and abundantly RESEARCH DESIGN AND METHODS — We infused 12 diabetic patients and 12 available within human cells. GSH is a 2 nondiabetic control subjects with [ H2]-glycine to measure GSH synthesis. We also measured tripeptide synthesized from glutamate, intracellular GSH concentrations, reactive oxygen metabolites, and lipid peroxides. Diabetic cysteine, and glycine in two steps cata- patients were restudied after 2 weeks of dietary supplementation with the GSH precursors lyzed by ␥-L-glutamyl-L-cysteine:glycine cysteine and glycine. ligase and glutathione synthetase. Diabe- tes is associated with decreased cellular RESULTS — Compared with control subjects, diabetic subjects had significantly higher fasting glucose (5.0 Ϯ 0.1 vs. 10.7 Ϯ 0.5 mmol/l; P Ͻ 0.001), lower erythrocyte concentrations of glutathione concentrations (2–5), but the glycine (514.7 Ϯ 33.1 vs. 403.2 Ϯ 18.2 ␮mol/l; P Ͻ 0.01), and cysteine (25.2 Ϯ 1.5 vs. 17.8 Ϯ cause of GSH deficiency currently is 1.5 ␮mol/l; P Ͻ 0.01); lower concentrations of GSH (6.75 Ϯ 0.47 vs. 1.65 Ϯ 0.16 ␮mol/g Hb; unknown. P Ͻ 0.001); diminished fractional (79.21 Ϯ 5.75 vs. 44.86 Ϯ 2.87%/day; P Ͻ 0.001) and Oxidative stress and ROS formation absolute (5.26 Ϯ 0.61 vs. 0.74 Ϯ 0.10 ␮mol/g Hb/day; P Ͻ 0.001) GSH synthesis rates; and are markedly increased by uncontrolled higher reactive oxygen metabolites (286 Ϯ 10 vs. 403 Ϯ 11 Carratelli units [UCarr]; P Ͻ 0.001) hyperglycemia (2,6); conversely, lower- Ϯ Ϯ Ͻ and lipid peroxides (2.6 0.4 vs. 10.8 1.2 pg/ml; P 0.001). Following dietary supplemen- ing blood glucose concentrations lowers tation in diabetic subjects, GSH synthesis and concentrations increased significantly and plasma oxidative stress (7,8). Decreased oxida- oxidative stress and lipid peroxides decreased significantly. tive stress could be an important mecha- CONCLUSIONS — Patients with uncontrolled type 2 diabetes have severely deficient syn- nism whereby glycemic control thesis of glutathione attributed to limited precursor availability. Dietary supplementation with diminishes the incidence of diabetic mi- GSH precursor amino acids can restore GSH synthesis and lower oxidative stress and oxidant crovascular complications (9,10). How- damage in the face of persistent hyperglycemia. ever, there are practical limitations to blunting oxidative stress through glyce- Diabetes Care 34:162–167, 2011 mic control alone, despite strenuous at- tempts to implement evidence-based iabetes is the leading worldwide macrovascular complications including guidelines, a majority of patients are un- cause of blindness, end-stage renal myocardial ischemia and strokes. Al- able to achieve the glycemic goals (e.g., D disease, and amputations. Diabetes though multiple pathways are involved in A1C Ͻ7%) advocated by the American also is associated with an elevated risk of mediating tissue damage, including the Diabetes Association (11). Consequently, despite the clear message of landmark tri- ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● als such as the Diabetes Control and Com- From the 1Translational Metabolism Unit, Baylor College of Medicine, Houston, Texas; the 2Division of plications Trial and the UK Prospective Diabetes, Endocrinology, and Metabolism, Department of Medicine, Baylor College of Medicine, Hous- Diabetes Study regarding the need for ex- ton, Texas; the 3Department of Pediatrics, Baylor College of Medicine, Houston, Texas; and the 4Chil- dren’s Nutrition Research Center, Agriculture Research Service, U.S. Department of Agriculture, Baylor cellent glycemic control, diabetes remains College of Medicine, Houston, Texas. the leading cause of blindness, renal fail- Corresponding author: Rajagopal V. Sekhar, [email protected]. ure, and amputations. There is an urgent Received 26 May 2010 and accepted 30 September 2010. Published ahead of print at http://care. need for novel strategies to reduce the rate diabetesjournals.org on 7 October 2010. DOI: 10.2337/dc10-1006. The contents of this article do not necessarily reflect the views or the policies of the USDA. Mention of trade of diabetes complications in patients un- names, commercial products, and organizations does not imply endorsement by the U.S. Government. able to achieve stable glycemic control. © 2011 by the American Diabetes Association. Readers may use this article as long as the work is properly We therefore investigated whether oxida- cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons. tive stress associated with low levels of org/licenses/by-nc-nd/3.0/ for details. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby GSH could be ameliorated through the marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. alternative strategy of increasing cellular

162 DIABETES CARE, VOLUME 34, NUMBER 1, JANUARY 2011 care.diabetesjournals.org Sekhar and Associates

GSH levels in diabetic patients with un- The GCRC protocol consisted of intrave- and left in the dark at room temperature controlled hyperglycemia. nous infusions of stable isotopes to mea- for 20 min for development of the GSH- Because circulating concentrations of sure GSH synthesis in the fasted state. All MBB derivative. After adding 0.5 ml of a protein depend on the balance between subjects were studied under baseline con- 20% perchloric acid, the sample was cen- its rates of production and consumption, ditions; only the diabetic subjects were trifuged, and the supernatant containing we hypothesized that GSH deficiency in studied again after 14 days of dietary sup- the MBB derivative was sealed and frozen uncontrolled diabetes occurs because of plementation with 0.81 mmol/kg/day of at Ϫ80°C for later analysis of GSH. Con- diminished synthesis. We further hy- cysteine (given as n-acetylcysteine) and centrations of GSSG were measured by pothesized that short-term dietary sup- 1.33 mmol/kg/day of glycine. Subjects first converting oxidized glutathione to plementation of two key amino acid were asked to consume their usual habit- reduced glutathione with the addition of a precursors of GSH, glycine and cysteine, ual diets from 2 weeks before beginning reducing agent (5 mmol/l dithiothreitiol) would increase intracellular GSH synthe- the study to the end of the study period. and measuring this as total GSH; the cal- 2 sis and concentrations and thus lower ox- Sterile solutions of [ H2]glycine culated difference between total GSH and idative stress, despite continuing (Cambridge Isotope Laboratories, reduced GSH is the concentration of hyperglycemia. To test these hypotheses, Woburn, MA) were prepared. After a GSSG. we used stable isotope methods to com- 10-h fast, subjects were admitted to the Erythrocyte GSH was isolated as an pare GSH synthesis rates and concentra- GCRC for the study, where two intrave- red blood cell-free aliquot, and the con- tions within erythrocytes, as well as nous catheters were inserted into superfi- centration measured high-performace plasma markers of oxidant damage, in cial veins for continuous infusion of the liquid chromatography (Waters, Milford, adult patients with poorly controlled type tracer solutions and blood sampling. After MA) using a 717 Plus autosampler com- 2 diabetes matched to nondiabetic con- a basal blood sample was drawn, a primed plexed to a 2475 fluorescent detector and trol subjects. The diabetic patients were constant intravenous infusion of equipped with a reverse-phase ODS Hy- 2 ␮ ␮ ϫ studied before and after 14 days of dietary [ H2]glycine (prime dose 20 mol/kg; in- persil column (5 m, 4.6 200 mm; Wa- supplementation with cysteine and fusion dose 15 ␮mol/kg/h) was main- ters). Elution of GSH was accomplished glycine. tained for 8 h. Additional blood samples with a 3–13.5% acetonitrile linear gradi- were taken at 2, 3, 4, 5, 6, 7, and8hfor ent in 1% acetic acid (pH 4.25) at a flow RESEARCH DESIGN AND measurement of erythrocyte GSH derived rate of 1.1 ml/min. The GSH eluate was METHODS — The study was ap- glycine isotopic enrichments. collected using a fraction collector, dried, proved by the institutional review board The primary outcome variables were and hydrolyzed for4hin4mol/l HCl at for Human Studies at Baylor College of fractional and absolute synthesis rates of 110°C. Medicine. Twelve adults with uncon- GSH within erythrocytes, erythrocyte trolled type 2 diabetes (A1C 8–10%) and GSH, cysteine, glycine and glutamate Erythrocyte free amino acids 12 nondiabetic control subjects matched concentrations, plasma lipid peroxide A 1-ml aliquot of blood was centrifuged, for age, sex, and BMI were recruited. In- levels, and plasma oxidative stress mea- and erythrocytes were then washed thrice formed consent was obtained from all sured as reactive oxygen metabolites. with 3 ml sodium chloride solution (9 subjects. All subjects were free of thyroid g/l). RBCs were then lysed by freeze-thaw disorders, hypercortisolemia, liver or re- Sample analyses action with the use of liquid nitrogen, and nal impairment and malignancy, and had Blood chemistries. Baseline plasma cellular proteins were precipitated by us- no infections or major illnesses during the samples were aliquotted into tubes for the ing 10% perchloric acid solution. After preceding 6 months. All had sedentary various assays, and stored at Ϫ80°C for centrifugation, the supernatant fluid was lifestyles and none consumed unusual di- later analyses. Hemoglobin, reactive oxy- used for erythrocyte free amino acid anal- ets or dietary supplements. All subjects gen metabolites (DROMS), (Diacron In- ysis. Before derivatization for gas chro- were instructed to abstain from alcoholic ternational, Grosetto Italy), plasma matographic–mass spectrometric beverages during the study. chemistries, and lipid peroxides were analysis, erythrocyte-free glycine was iso- All patients received diabetic manage- measured. lated by cation-exchange (Dowex 200ϫ; ment from their primary physicians. To Erythrocyte GSH analyses. Erythro- Bio-Rad Laboratories, Hercules, CA) prevent acute swings of blood glucose, cyte GSH concentration and isotopic en- chromatography. Samples of glycine de- and to achieve comparable glycemic lev- richment of GSH were measured as rived from erythrocyte glutathione and els before and after supplementation with described next (12). Briefly, duplicate ali- plasma and erythrocyte-free glycine sam- cysteine and glycine, we excluded pa- quots of 1 ml whole blood were centri- ples were converted to the n-propyl ester, tients receiving insulin therapy and re- fuged to separate packed cells and heptafluorobutyramide derivative. The cruited only subjects who recently were measure GSH and glutamate, cysteine, tracer-to-tracee ratio for glycine in various diagnosed and were being treated with ei- and glycine, respectively. The first aliquot samples was determined by negative ther lifestyle modification or oral antidia- of packed erythrocytes was washed thrice chemical ionization gas chromatograph- betic agents. with normal saline and 1 ml of monobro- ic–mass spectrometric analysis with se- Subjects were studied in the adult mobimane (MBB) buffer (5 mmol/l MBB, lective monitoring of ions at mass-to- general clinical research center (GCRC) of 17.5 mmol/l Na2EDTA, 50 mmol/l potas- charge ratios of 293–295 on an Agilent the Baylor College of Medicine. After sium phosphate, 50 mmol/l serine, and 6980 gas chromatograph complexed to a measurement of blood counts, glucose 50 mmol/l boric acid) was added. Cells 5973 mass spectrometer (Agilent Tech- concentrations, glycosylated hemoglo- were immediately lysed by rapid freeze- nologies, Wilmington, DE). Erythrocyte bin, and liver and renal profiles, subjects thaw with liquid nitrogen, and the eryth- amino acid concentrations were deter- participated in the first infusion study. rocyte-MBB buffer mixture was shaken mined by high-performace liquid chro- care.diabetesjournals.org DIABETES CARE, VOLUME 34, NUMBER 1, JANUARY 2011 163 Type 2 diabetes, glutathione, oxidative stress matography analysis with a Waters Table 1—Clinical, hematological, and biochemical characteristics of nondiabetic control sub- system (Millipore, Milford, MA). jects and pretreatment and posttreatment data for subjects with type 2 diabetes

Oxidant markers Diabetic Diabetic To assess the level of oxidative stress, the Nondiabetic subjects subjects derivatives of DROMs were determined in Parameters subjects pretreatment posttreatment P serum. Briefly, plasma is reacted with an acidic acetate buffer (pH 4.8), which lib- Age (years) 50.4 Ϯ 3.8 51.0 Ϯ 3.1 51.0 Ϯ 3.1 NS erates transition metal ions that catalyze BMI (kg/m2) 28.0 Ϯ 0.9 30.4 Ϯ 0.7 30.0 Ϯ 0.9 NS the decomposition of the hydroperoxides Hemoglobin (g/l) 14.2 Ϯ 1.8 13.7 Ϯ 1.4 13.7 Ϯ 1.2 NS to alkoxy and peroxyl radicals. These Fasting plasma glucose (mmol/l) 5.0 Ϯ 0.1 10.7 Ϯ 0.5* 10.6 Ϯ 0.4 Ͻ0.001* newly formed radicals in turn oxidize the A1C (%) 5.5 Ϯ 0.1 9.1 Ϯ 0.2* 9.0 Ϯ 0.2 Ͻ0.001* spectrophotometric marker (N,N- Blood urea nitrogen (mmol/l) 5.3 Ϯ 0.3 5.4 Ϯ 0.6 5.4 Ϯ 0.4 NS diethyl-p-phenylenediamine), which is Creatinine (␮mol/l) 88.4 Ϯ 3.4 79.6 Ϯ 4.2 79.6 Ϯ 3.8 NS detectable by absorption at 505 nm as Alanine aminotransferase (U/l) 20.4 Ϯ 1.2 23.8 Ϯ 5.0 24.4 Ϯ 4.4 NS UCarr (where 1 UCarr is equal to 0.8 mg/l Aspartate aminotransferase (U/l) 17.6 Ϯ 1.4 18.2 Ϯ 2.4 17.9 Ϯ 2.6 NS hydrogen peroxide). Glutamate (mmol/l) 530.1 Ϯ 88.3 451.5 Ϯ 120.7 482.4 Ϯ 98.2 NS Cysteine (␮mol/l) 25.2 Ϯ 1.5 17.8 Ϯ 1.5* 25.5 Ϯ 1.9† Ͻ0.01* Lipid peroxides Ͻ0.05† Briefly, this was measured using freshly Glycine (␮mol/l) 514.7 Ϯ 33.1 403.2 Ϯ 18.2* 521.6 Ϯ 19.4† Ͻ0.01* prepared buffers containing ammonium Ͻ0.01† Ϯ Ϯ Ϯ Ͻ ferrous sulfate, xylenol orange, H2SO4, DROMs (UCarr) 286 10 403 11* 359 10†‡ 0.001* BHT in 90% vol/vol methanol, and triph- Ͻ0.05† enylphosphine in methanol. Standard so- Ͻ0.01‡ ␮ Ϯ Ϯ Ϯ Ͻ lutions were made using 30% H2O2. Lipid peroxide ( mol/l) 2.6 0.4 10.8 1.2* 6.2 0.9†‡ 0.001* Heparinized blood was centrifuged, and Ͻ0.01† 10 ␮l of 10 mmol/l TPP solution was Ͻ0.05‡ added to 90 ␮l of plasma (control vials) *Nondiabetic control subjects versus diabetic subjects: pre-treatment. †Diabetic subjects: pretreatment then 10 ␮l of methanol was added (test versus posttreatment. ‡Diabetic subjects: posttreatment versus nondiabtic control subjects. vials), and this solution was incubated at room temperature for 30 min. After adding appropriate buffers, each vial was in- the diabetic group presupplementation Erythrocyte GSH kinetics and cubated at room temperature for 30 min, and the control group and also between concentration of glycine, cysteine, centrifuged, and the absorbance of the su- the diabetic group postsupplementation glutamate, and GSH pernatant was determined by spectropho- and the control group. Differences in out- Compared with nondiabetic control sub- tometry. The hydroperoxide content was come measures in the diabetic patients jects, subjects with poorly controlled dia- determined from test controls and as- studied pre- and postsupplementation betes had 73.8% lower erythrocyte- sayed against the standard curve (13). was determined using a paired t test. Data reduced glutathione concentrations analysis was performed with the Statmate (6.75 Ϯ 0.47 vs. 1.65 Ϯ 0.16 ␮mol/g Hb; Calculations statistical software (GraphPad Software, P Ͻ 0.001) (Fig. 1A) and higher concen- The fractional synthesis rate (FSR) of LA Jolla, CA). Results were considered to trations of erythrocyte-oxidized GSSG erythrocyte GSH (FSRGSH) was calculated be statistically significant at P Ͻ 0.05. (0.10 Ϯ 0.01 vs. 0.33 Ϯ 0.07 ␮mol/g Hb; according to the precursor-product equa- P Ͻ 0.05). Compared with control sub- tion: FSR (%/day) ϭ (IR Ϫ IR )/ GSH t7 t5 jects, total glutathione concentrations (IR ϫ 1,200/t Ϫ t ), where IR Ϫ IR rbc 7 5 t7 t5 RESULTS (6.75 Ϯ 0.47 vs. 1.65 Ϯ 0.16 ␮mol/g Hb; is the increase in the isotope ratio of eryth- P Ͻ 0.001) and the ratio of GSH to GSSG rocyte GSH-bound glycine between the Baseline characteristics fifth and seventh hours of infusion, when were both significantly lower in diabetic The average ages of the control and dia- subjects (59.15 Ϯ 4.12 vs. 6.30 Ϯ 1.30; the isotope ratio of erythrocyte-free gly- Ϯ Ͻ cine, IR , had reached a steady state. The betic subjects were 50.4 3.8 and P 0.001). Diabetic subjects also had rbc 51.0 Ϯ 3.1 years, respectively (Table 1). 43.4% slower GSH FSR (79.21 Ϯ 5.75 vs. units of FSR are percentage per day (%/ Ϯ Ͻ day). The absolute synthesis rate (ASR) of There were no group differences in BMI, 44.86 2.87%/day; P 0.001) and Ϯ erythrocyte GSH per day was calculated hematocrit and hemoglobin concentra- 85.5% slower ASR (5.26 0.61 vs. Ϯ ␮ Ͻ as follows: ASR ϭ erythrocyte GSH con- tions, renal functions, or liver enzymes. 0.74 0.10 mol/g Hb/day; P 0.001) centration ϫ FSR. The units of ASR are The control subjects were euglycemic, (Figs. 1B and C). Compared with control expressed as micromols of GSH per gram whereas the diabetic subjects had signifi- subjects, diabetic subjects also had signif- of hemoglobin per day. cantly higher fasting glucose concentra- icantly lower RBC concentrations of gly- tions and glycosylated hemoglobin. There cine (514.7.7 Ϯ 33.1 vs. 403.2 Ϯ 18.2 Statistics were no differences in hematologic pa- ␮mol/l; P Ͻ 0.01) and cysteine (25.2 Ϯ Data are expressed as means Ϯ SE. An rameters, renal function, or liver enzymes 1.5 vs. 17.8 Ϯ 1.5, ␮mol/l; P Ͻ 0.01) but independent unpaired t test was used to before or after cysteine and glycine sup- not glutamate (530.1 Ϯ 88.3 vs. 451.5 Ϯ compute differences in means between plementation in the diabetic group. 120.7 ␮mol/l; P ϭ 0.61).

164 DIABETES CARE, VOLUME 34, NUMBER 1, JANUARY 2011 care.diabetesjournals.org Sekhar and Associates

ated with signiﬁcantly higher concentrations of markers of oxidative damage (plasma DROMs and lipid peroxides) (Ta- ble 1) than in control subjects. After 14 days of cysteine and glycine supplementation, there was a signiﬁcant fall in these parameters in the diabetic subjects, although not to the levels observed in the nondiabetic control subjects.

CONCLUSIONS — The results of this study demonstrate that intracellular concentrations of GSH, as well as of its precursor amino acids cysteine and glycine, are decreased in adult patients with poorly controlled type 2 diabetes compared with nondiabetic subjects. Whereas previous studies also have found that diabetic patients have deficient cellular levels of GSH (3–5), the present data go further to demonstrate that the GSH deficiency seems to be attributed to a markedly lower synthetic rate of GSH and is associated with increased oxidant stress and elevated plasma markers of oxidant damage. Fourteen days of oral dietary supplementation with cysteine and glycine in diabetic subjects restored the fractional synthesis rates of GSH to those observed in nondiabetic control subjects. This was accompanied by significant de- clines in both oxidative stress and plasma markers of oxidant damage. Suboptimal restoration of the absolute synthesis rate of GSH after cysteine and glycine supplementation, despite full normalization of its fractional synthesis, suggests that there is also persistent, ac- celerated consumption of GSH in diabetic ␮ Figure 1—A: Erythrocyte GSH concentrations ( mol GSH/g Hb). B: GSH fractional synthesis patients with uncontrolled hyperglyce- rate (%/day). C: GSH absolute synthesis rate (␮mol GSH/g Hb /day). RBC ϭ erythrocytes. Ͻ ϱ mia. Furthermore, although the levels of *Control subjects versus diabetes pretreatment, P 0.001; Diabetes pretreatment versus dia- the oxidative stress markers were signifi- betes posttreatment, P Ͻ 0.001). cantly diminished after 2 weeks of precursor supplementation, they did not attain Diabetic subjects received treat- control subjects diabetic after supplemen- the low levels observed in the nondiabetic ment with cysteine and glycine supple- tation, they remained at 59.7% (2.72 Ϯ control subjects. A longer duration of in- mentation for 14 days, and this led to 0.15 vs. 6.75 Ϯ 0.47 ␮mol GSH/g Hb; P Ͻ tervention may be required to normalize 85.1% increase in erythrocyte GSH FSR 0.001) and 58.7% (0.74 Ϯ 0.10 vs. 5.26 Ϯ synthesis and intracellular concentrations (44.86 Ϯ 2.87 vs. 83.03 Ϯ 3.66%/day; 0.61 ␮mol GSH/g Hb/day; P Ͻ 0.001) of GSH. P Ͻ 0.001), resulting in a 64.4% in- lower, respectively, in the diabetic subjects GSH is a tripeptide of glutamate, cys- crease in erythrocyte GSH concentra- postsupplementation (Figs. 1A–C). Com- teine, and glycine, and measurement of tions (1.65 Ϯ 0.16 vs. 2.72 Ϯ 0.15 pared with presupplementation values, the these amino acids within erythrocytes ␮mol/g Hb; P Ͻ 0.001) and a 193.8% post supplementation values of erythrocyte showed low levels of cysteine and glycine increase in GSH ASR (0.74 Ϯ 0.10 vs. concentrations of GSSG (0.33 Ϯ 0.07 vs. but not of glutamate. Deficiency of cys- 2.17 Ϯ 0.17 ␮mol GSH/g Hb/day; P Ͻ 0.28 Ϯ 0.07 mmol/g Hb; P ϭ NS) and the teine and glycine in diabetic humans also 0.001). Precursor supplementation in- ratio of GSH to GSSG (6.30 Ϯ 1.30 vs. has been reported previously in the liter- creased GSH FSR to the level of nondi- 12.63 Ϯ 3.15; P ϭ NS) did not change ature (14). Because catabolic processes of abetic control subjects and led to significantly. virtually all amino acids cycle through significant increases in GSH concentra- glutamate production, it is not surprising tions and GSH ASR. However, when Plasma oxidant parameters that glutamate levels were not different GSH concentrations and ASR were com- The slower rates of GSH synthesis in the from those of nondiabetic subjects. The pared between diabetic subjects and diabetic subjects at baseline were associ- deficiency in cysteine and glycine is in-

care.diabetesjournals.org DIABETES CARE, VOLUME 34, NUMBER 1, JANUARY 2011 165 Type 2 diabetes, glutathione, oxidative stress triguing because both these amino acids controlled type 2 diabetes, there is a true thus combat oxidative stress and prevent traditionally are considered to be “nones- deficiency of glutathione. chronic complications in patients with sential,” meaning that they can be synthe- Sustained hyperglycemia is linked to diabetes. sized endogenously. Why are patients increased oxidative stress, and with an in- with uncontrolled type 2 diabetes unable creased risk of diabetic microvascular and to synthesize cysteine and glycine ade- macrovascular complications. Mecha- Acknowledgments— This work was sup- nisms implicated in hyperglycemia- ported by the Young Investigator Award in Ge- quately? The plasma flux of a nonessential riatric Endocrinology funded by the Atlantic amino acid results from the sum of its rate driven tissue damage in diabetes include Philanthropies, John A. Hartford Foundation; of release from protein breakdown, de abnormal signaling through protein ki- the American Diabetes Association; the Asso- novo synthesis, and dietary absorption. nase C, elevated advanced glycation end ciation of Specialty Professors (to R.V.S.); and Studies evaluating protein turnover in di- products, and the aldose reductase path- the National Institutes of Health (NIH) Train- abetic patients have reported abnormal way (1). ROS is known to stimulate these ing Program in Molecular Endocrinology overall protein balance (15), and disrup- pathways by activation of aldose reduc- (T32-DK07696; to A.G.). Some of the work tion of metabolic pathways resulting from tase, protein kinase C isoforms, and nu- was performed at the Baylor Children’s Nutri- clear factor-␬B and induction of tion Research Center, which is supported by hyperglycemia could impose higher di- the U.S. Department of Agriculture (USDA)/ etary protein requirements (15). Hence, a diacylglycerol and advanced glycation end-product formation (22). Lowering Agricultural Research Service under coopera- combination of impaired protein turn- tive agreement no. 5862-5-01003. This work over and dietary deficiency could under- levels of mitochondrial ROS (and thereby was also supported in part by NIH Grants lie inadequate availability of cysteine and oxidative stress) successfully prevents ac- (M01-RR00188, GCRC; and P30DK079638, glycine for GSH synthesis in type 2 diabe- tivation and induction of these mecha- NIH Diabetes and Endocrinology Research tes. Additional studies are needed to eval- nisms (23). Because ROS production is Center) at Baylor College of Medicine. uate the mechanisms underlying increased by hyperglycemia, optimizing No potential conflicts of interest relevant to decreased availability of cysteine and gly- glycemic control in diabetic subjects this article were reported. R.V.S. was responsible for study design, cine in type 2 diabetes. should decrease ROS, but, in clinical practice, normalizing glycemia remains a conduction, supervision of the study, sample Although the exact mechanisms un- and data analyses, manuscript preparation, derlying cysteine and glycine deficiency major challenge in diabetes management. It is estimated that up to 55% of diabetic and submission. S.V.M. contributed to study in type 2 diabetes are not clear, the net conduction and sample analyses. S.G.P. ana- patients in the U.S. do not attain the result of this deficiency could have a dom- lyzed samples. A.G.P. recruited subjects and American Diabetes Association’s recom- ino effect leading to oxidative stress and conducted studies. V.T.R. analyzed samples. mended glycemic goals, and 67% do not tissue damage. First, GSH synthesis A.B. reviewed and edited the manuscript. F.J. attain the more stringent glycemic targets would be blunted, and this would result contributed to study design and manuscript of the American Association of Clinical review. in a critical imbalance between GSH- Endocrinologists. Therefore, interven- We thank Dina Harleaux, RN; Lynne Scott, driven antioxidant protection and the tions directly aimed at lowering ROS- MS, RD; Varsha Patel, RPh; and the nursing harmful effects of unopposed elevated ox- mediated oxidative stress are needed to staff of the Baylor GCRC for excellent care of idative stress in uncontrolled diabetes. prevent diabetic tissue damage even in the subjects and meticulous attention to protocol. Deficiency of sulfur amino acids (16) or presence of hyperglycemia. The present protein content (17) in the diets of data suggest that increasing GSH levels healthy humans has been previously References with oral precursor supplementation is a 1. Brownlee M. Biochemistry and molecular shown to result in suppression of GSH viable intervention to target diabetic oxi- turnover in vivo. Further, animals fed di- cell biology of diabetic complications. Na- dative stress directly and could constitute ture 2001;414:813–820 ets specifically lacking GSH precursor a novel, safe, and inexpensive form of nu- 2. Whiting PH, Kalansooriya A, Holbrook I, amino acids, especially cysteine, develop tritional treatment. When used as an ad- Haddad F, Jennings PE. The relationship GSH deficiency (18–21). Together with junct to standard glycemic management, between chronic glycaemic control and the present findings, these data indicate this approach could significantly attenu- oxidative stress in type 2 diabetes melli- that an important underlying cause for in- ate tissue damage due to oxidative stress tus. Br J Biomed Sci 2008;65:71–74 tracellular GSH deficiency in diabetes is in patients with diabetes. 3. Sundaram RK, Bhaskar A, Vijayalingam S, Viswanathan M, Mohan R, Shanmu- decreased in vivo GSH synthesis because In conclusion, these data show that of a reduced availability of the precursor gasundaram KR. Antioxidant status and an important reason underlying elevated lipid peroxidation in type II diabetes mel- amino acids cysteine and glycine. oxidative stress in type 2 diabetes is defi- litus with and without complications. Could the lower concentrations of ciency of glutathione, which occurs be- Clin Sci (Lond) 1996;90:255–260 GSH be a result of impaired cycling because of the decreased synthesis caused 4. Vijayalingam S, Parthiban A, Shanmu- tween its oxidized and reduced isoforms, by limited availability of the precursor gasundaram KR, Mohan V. Abnormal anti- with a greater proportion of glutathione amino acids cysteine and glycine. Dietary oxidant status in impaired glucose tolerance trapped as the oxidized form? To answer supplementation of these amino acid pre- and non-insulin-dependent diabetes melli- this question, we also measured concen- cursors restores fractional synthesis of tus. Diabet Med 1996;13:715–719 trations of total GSH by converting oxi- GSH and significantly reduces oxidative 5. Memisogullari R, Taysi S, Bakan E, Capo- glu I. Antioxidant status and lipid peroxi- dized glutathione to its reduced isoform stress and markers of oxidant damage. dation in type II diabetes mellitus. Cell with a reducing agent, and found that Providing supplemental cysteine and gly- Biochem Funct 2003;21:291–296 subjects with diabetes still had signifi- cine in the diet could be a safe, inexpen- 6. Menon V, Ram M, Dorn J, Armstrong D, cantly lower concentrations of total gluta- sive, and innovative approach to increase Muti P, Freudenheim JL, Browne R, Schun- thione. This suggests that in poorly GSH synthesis and concentrations and emann H, Trevisan M. Oxidative stress and

166 DIABETES CARE, VOLUME 34, NUMBER 1, JANUARY 2011 care.diabetesjournals.org Sekhar and Associates

glucose levels in a population-based sam- glutathione concentration and rate of syn- the metabolic response to tumor necrosis ple. Diabet Med 2004;21:1346–1352 thesis. Curr Opin Clin Nutr Metab Care factor alpha in rats fed a low protein diet. 7. Bravi MC, Armiento A, Laurenti O, Cas- 2000;3:385–390 J Nutr 1992;122:2066–2073 sone-Faldetta M, De Luca O, Moretti A, 13. Nourooz-Zadeh J. Ferrous ion oxidation 19. Jahoor F, Wykes LJ, Reeds PJ, Henry JF, De Mattia G. Insulin decreases intracellu- in presence of xylenol orange for detec- del Rosario MP, Frazer ME. Protein-defi- lar oxidative stress in patients with type 2 tion of lipid hydroperoxides in plasma. cient pigs cannot maintain reduced gluta- diabetes mellitus. Metabolism 2006;55: Methods Enzymol 1999;300:58–62 thione homeostasis when subjected to the 691–695 14. De Luca G, Calpona PR, Caponetti A, Ma- stress of inflammation. J Nutr 1995;125: 8. Unger J. Reducing oxidative stress in pa- caione V, Di Benedetto A, Cucinotta D, Di 1462–1472 tients with type 2 diabetes mellitus: a pri- Giorgio RM. Preliminary report: amino 20. Bella DL, Hahn C, Stipanuk MH. Effects mary care call to action. Insulin 2008;3: acid profile in platelets of diabetic pa- of nonsulfur and sulfur amino acids on 176–184 tients. Metabolism 2001;50:739–741 the regulation of hepatic enzymes of cys- 9. The Diabetes Control and Complications 15. Gougeon R, Morais JA, Chevalier S, teine metabolism. Am J Physiol 1999; Trial Research Group. The effect of inten- Pereira S, Lamarche M, Marliss EB. Deter- 277:E144–E153 sive treatment of diabetes on the develop- minants of whole-body protein metabo- 21. Cresenzi CL, Lee JI, Stipanuk MH. Cys- ment and progression of long-term lism in subjects with and without type 2 teine is the metabolic signal responsible complications in insulin-dependent dia- diabetes. Diabetes Care 2008;31:128– for dietary regulation of hepatic cysteine betes mellitus. N Engl J Med 1993;329: 133 977–986 16. Lyons J, Rauh-Pfeiffer A, Yu YM, Lu XM, dioxygenase and glutamate cysteine ligase 10. UK Prospective Diabetes Study (UKPDS) Zurakowski D, Tompkins RG, Ajami AM, in intact rats. J Nutr 2003;133: Group. Intensive blood-glucose control Young VR, Castillo L. Blood glutathione 2697–2702 with sulphonylureas or insulin compared synthesis rates in healthy adults receiving 22. Nishikawa T, Edelstein D, Brownlee M. with conventional treatment and risk of a sulfur amino acid-free diet. Proc Natl The missing link: a single unifying mech- complications in patients with type 2 di- Acad SciUSA2000;97:5071–5076 anism for diabetic complications. Kidney abetes (UKPDS 33). Lancet 1998; 17. Jackson AA, Gibson NR, Lu Y, Jahoor F. Int Suppl 2000;77:S26–S30 352:837–853, Synthesis of erythrocyte glutathione in 23. Nishikawa T, Edelstein D, Du XL, Yamag- 11. Fan T, Koro CE, Fedder DO, Bowlin SJ. healthy adults consuming the safe ishi S, Matsumura T, Kaneda Y, Yorek Ethnic disparities and trends in glycemic amount of dietary protein. Am J Clin Nutr MA, Beebe D, Oates PJ, Hammes HP, Gi- control among adults with type 2 diabetes 2004;80:101–107 ardino I, Brownlee M. Normalizing mito- in the U.S. from 1988 to 2002. Diabetes 18. Grimble RF, Jackson AA, Persaud C, chondrial superoxide production blocks Care 2006;29:1924–1925 Wride MJ, Delers F, Engler R. Cysteine three pathways of hyperglycaemic dam- 12. Reid M, Jahoor F. Methods for measuring and glycine supplementation modulate age. Nature 2000;404:787–790

care.diabetesjournals.org DIABETES CARE, VOLUME 34, NUMBER 1, JANUARY 2011 167