Pathophysiology/Complications ORIGINAL ARTICLE

Transmembrane Electron Transfer in Diabetic Nephropathy

ELENA MATTEUCCI, MD definite interrelationship of the various OTTAVIO GIAMPIETRO, MD sulfhydryl populations on the RBC mem- brane in the plasma and on the plasma pro- teins (albumin) has been suggested to be operating so that the cellular redox balance affects the extracellular environment (4). OBJECTIVE — Erythrocytes (red blood cells [RBCs]) reduce extracellular ferricyanide by causes both enhanced oxida- transmembrane transfer of reducing equivalents involving ascorbate recycling. tive stress and impaired regeneration of protective antioxidants (5). Abnormalities RESEARCH DESIGN AND METHODS — Because ascorbate regeneration is glu- in ascorbic acid metabolism (6), impaired tathione (GSH) dependent and cells may be depleted of GSH in diabetes, we measured RBC GSH, plasma sulfhydryl (SH) groups, and RBC-mediated ferricyanide reduction in 30 type 1 GSH synthesis and thiol transport (7), and diabetic patients (age 34 ± 10 years, disease duration 20 ± 8 years; no complications, n = 10; significant defects of chain-breaking antiox- retinopathy, n = 10; nephropathy, n = 10), their 36 siblings (age 39 ± 13 years), and matched idant blood protection (8) have all been healthy volunteers. described in patients with both type 1 and . Increased cytosolic reduc- RESULTS — Fasting plasma glucose was 15 ± 7 mmol/l (vs. 5 ± 1 in control subjects, P tive stress associated with an increased ratio 0.001), HbA1c 8.4 ± 1.5% (vs. 5.4 ± 0.3, P 0.001), GSH 0.76 ± 0.12 mg/ml packed RBCs (vs. of NADH to NAD is a candidate mecha- 0.88 ± 0.18, P 0.01), SH groups 401 ± 72 µmol/l (vs. 444 ± 56, P 0.05), and ferrocyanide nism of early diabetes-induced glomerular generation 15 ± 5 µmol/ml RBC per h (vs. 13 ± 5, NS). In comparison with 10 normoalbu- dysfunction, manifested as increases in minuric diabetic subjects with retinopathy, 10 patients with diabetic nephropathy had similar blood flow, glomerular filtration rate, and fasting plasma glucose, HbA1c, and SH groups; lower RBC GSH (0.73 ± 0.08 vs. 0.85 ± 0.11, P 0.05); and higher ferrocyanide generation (18 ± 4 vs. 14 ± 5, P 0.05). The 10 patients (9). without complications differed from the 10 healthy volunteers in glycemic control and RBC In the present study, we have mea- GSH. RBC electron transfer correlated with plasma lactate (r = 0.8, P = 0.01) only in the uncom- sured the erythrocyte-mediated ferricyanide plicated group. No difference was detected between siblings and healthy control subjects or reduction system with intact erythrocytes between siblings of subjects in the nephropathy and retinopathy groups. Among diabetic from type 1 diabetic patients with or with- patients, the rate of ferrocyanide generation was associated with urinary albumin excretion, out complications from , plasma , and SH groups (multiple r = 0.6, P 0.01). from their nondiabetic siblings, and from healthy control subjects. We wished to CONCLUSIONS — Transmembrane electron transfer is selectively increased in diabetic explore the status of transplasma mem- nephropathy, where RBC GSH is also depleted. The abnormality is peculiar to the nephropa- brane redox activity in the patients with thy group and not contributed by familial or hereditary components because the electron flow was normal in siblings. The close relationship between cytosolic NADH and RBC electron diabetes, its relationship with diabetic com- transfer observed in diabetic patients without complications seems to be lost in the microan- plications, and the contribution of familial giopathic patients. Whereas patients with retinopathy alone still had normal activity of the factors to the electron transfer rate. RBC-reducing system, patients with nephropathy showed significantly increased activity, unrelated to metabolic parameters or plasma lactate concentration and correlated with renal RESEARCH DESIGN AND function parameters and plasma thiols. METHODS Diabetes Care 23:994–999, 2000 Study subjects Patients with (n = 30) and control age-matched nondiabetic subjects uman erythrocytes reduce extracellu- the export of electrons (2). Ascorbate regen- (n = 30) were studied (Table 1). Diabetic 3 lar ferricyanide, [Fe(CN)6] , to fer- eration from dehydroascorbate is strongly patients had been treated with at least 2 rocyanide, [Fe(CN) ]4, in a reaction dependent on glycolysis in red blood cells H 6 daily injections from the time of that involves the transmembrane transfer of (RBCs) and occurs via intracellular reduced diagnosis (disease duration 20 ± 8 years). reducing equivalents (1). Ascorbate has glutathione (GSH) in a process involving None received medical treatment except been implicated as the preferred substrate in both NADH and NADPH (3). Moreover, a insulin (0.65 ± 0.13 U/kg) or possibly anti- hypertensive drugs. The group consisted of 10 patients without diabetic complications, From the Department of Internal Medicine, University of Pisa, Pisa, Italy. 10 patients with retinopathy (background Address correspondence and reprint requests to Ottavio Giampietro, Dipartimento di Medicina Interna or proliferative, determined by fluorescein (Clinica Medica II), via Roma 67, 56100 Pisa, Italy. E-mail: [email protected]. angiography), and 10 patients with Received for publication 20 September 1999 and accepted in revised form 5 April 2000. Abbreviations: GSH, glutathione; RBC, red blood cell; SH, sulfhydryl; UAER, urinary albumin excretion rate. nephropathy. Among the patients with A table elsewhere in this issue shows conventional and Système International (SI) units and conversion nephropathy, 1 was on dialysis (serum cre- factors for many substances. atinine levels 133 µmol/l), 4 had persis-

994 DIABETES CARE, VOLUME 23, NUMBER 7, JULY 2000 Matteucci and Giampietro

Table 1—Clinical and metabolic characteristics of the study subjects absorbance at 535 nm with 4,7-diphenyl- 1,10-phenanthroline and sulfonated as the color-developing agent against a 0 time Characteristic Type 1 diabetic subjects Control subjects P blank. The initial rate of ferrocyanide gen- Sex (M/F) 11/19 14/16 — eration was expressed relative to the real Age (years) 34 ± 10 36 ± 10 — packed cell volume (determined by an BMI (kg/m2) 24 ± 2 24 ± 3 — automated cell counter) as micromoles per Mean (mmHg) 92 ± 13 90 ± 11 — milliliter packed RBCs per h. The within- Plasma glucose (mmol/l) 15 ± 7 5 ± 1 0.001 and between-run coefficients of variation (µmol/l) 371 ± 60 226 ± 17 0.001 were 6 and 9%, respectively. HbA1c (%) 8.4 ± 1.5 5.4 ± 0.3 0.001 The level of erythrocyte GSH was mea- Plasma lactate (mmol/l) 1.7 (1.2–2.4) 1.7 (1.3–2.9) — sured by the method of Beutler et al. (12) Plasma pyruvate (mmol/l) 0.06 (0.02–0.2) 0.04 (0.01–0.1) — and expressed as milligrams per milliliter GSH (mg/ml RBC) 0.76 ± 0.12 0.88 ± 0.18 0.01 RBCs. Plasma sulfhydryl (SH) groups were SH groups (µmol/l) 401 ± 72 444 ± 56 0.05 measured using fresh plasma and the col- Plasma albumin (mg/dl) 4,394 ± 478 4,576 ± 457 — orimetric assay described by Ellman (13). Ferrocyanide (µmol ml1 h1) 15 ± 5 13 ± 5 — The cytosolic ratio of NADH to NAD was Data are n, means ± SD, or medians (range). inferred from the ratio of plasma lactate to pyruvate. Plasma glucose, fructosamine, lac- tate, and pyruvate were measured by stan- tent macroalbuminuria (urinary albumin sisted of 120 mmol/l NaCl, 5 mmol/l KCl, dard enzymatic methods (reagents from excretion rate [UAER] 200 µg/min and 1 mmol/l MgSO , 16 mmol/l Na HPO , Boehringer Mannheim, Mannheim, Ger- 4 2 4 serum creatinine 133 µmol/l), and 5 had and 5 mmol/l glucose (pH 7.4), and cen- many). HbA1c was evaluated by the Bio-Rad persistent microalbuminuria (defined as a trifuged at 1,320g for 5 min at 4°C. Diamat Fully Automated Glycosylated UAER 20 µg/min in 2 of 3 consecutive To measure ferricyanide reduction, 0.5 Hemoglobin Analyzer System (Bio-Rad). 24-h urine collections within 6 months in ml of the packed cells were diluted to 5 ml Albumin was quantified by the kinetic the absence of a or with phosphate-buffered saline containing immunonephelometric method with an heart failure). 1 mmol/l ferricyanide and incubated under automated instrument (Behring Institute Two of the patients with retinopathy magnetic stirring in a water bath at 37°C. nephelometer and reagents; Scoppito, and 6 of the patients with nephropathy were At 5, 10, 20, and 40 min, 0.3-ml aliquots of L’Aquila, Italy). taking drugs, including ACE inhibitors, cal- the cell suspension were removed and cen- cium antagonists, and/or vasodilators. trifuged at 3,110g for 4 min at 4°C (10). An Statistical analysis We compared 36 nondiabetic nor- aliquot of 0.05 ml of the supernatant was All data are expressed as means ± SD, motensive siblings of the diabetic patients sampled, and the ferrocyanide content was except plasma lactate and pyruvate, which (age 39 ± 13 years, BMI 25 ± 4 kg/m2, and measured by the assay of Avron and Shavit are given as median and range. Results were mean blood pressure 92 ± 14 mmHg) with (11). Ferrocyanide was determined from its analyzed by a commercial software package 36 matched control subjects (39 ± 10, 25 ± 4, and 93 ± 10). Twelve were siblings of probands with nephropathy and 16 of Table 2—Clinical and metabolic characteristics of 10 type 1 diabetic patients with nephropathy probands with retinopathy and normoal- versus 10 normoalbuminuric patients with retinopathy buminuria. No study subject was taking antioxidant vitamins at the time of blood Type 1 diabetic subjects Type 1 diabetic subjects sampling. All subjects gave their informed Characteristic with normoalbuminuria with nephropathy P consent to participate in the study. Participants were interviewed with Sex (M/F) 9/1 4/6 — regard to their own and familial medical his- Age (years) 39 ± 9 36 ± 10 — tory. They underwent a routine clinical Duration of diabetes (years) 24 ± 7 23 ± 7 — assessment. Height and weight were mea- Insulin dose (U/kg) 0.7 ± 0.2 0.7 ± 0.1 — sured (in indoor clothing without shoes). BMI (kg/m2) 24 ± 2 24 ± 2 — The mean blood pressure was calculated as Mean blood pressure (mmHg) 90 ± 10 98 ± 18 — the diastolic pressure plus one-third of the Plasma glucose (mmol/l) 15 ± 8 18 ± 8 — pulse pressure. Fructosamine (µmol/l) 381 ± 55 384 ± 75 — HbA1c (%) 8.9 ± 1.2 8.9 ± 1.7 — Biochemical analyses Plasma lactate (mmol/l) 1.8 (1.2–2.2) 1.7 (1.3–2.7) — After an overnight fast, blood was collected Plasma pyruvate (mmol/l) 0.06 (0.02–0.2) 0.05 (0.03–0.2) — in heparin-rinsed syringes and immedi- GSH (mg/ml RBC) 0.85 ± 0.12 0.73 ± 0.08 0.05 ately centrifuged at 54g for 9 min at 20°C SH groups (µmol/l) 382 ± 70 373 ± 74 — to remove platelet-rich plasma. Packed Plasma albumin (mg/dl) 4,259 ± 170 4,298 ± 716 — RBCs were washed 3 times in 10 volumes Ferrocyanide (µmol ml 1 h 1) 14 ± 5 18 ± 4 0.05 of phosphate-buffered saline, which con- Data are n, means ± SD, or medians (range).

DIABETES CARE, VOLUME 23, NUMBER 7, JULY 2000 995 Ferricyanide reduction in diabetic nephropathy

No significant difference emerged when comparing the rate of ferricyanide reduction of patients with complications who were taking drugs (n = 8, 13 ± 6 µmol ml1 h1) with those who were not (n = 12, 15 ± 5 µmol ml1 h1). The remaining 10 patients without dia- betic complications were younger than the groups with complications and showed both shorter duration of diabetes (14 ± 9 years, P 0.05) and lower glycated hemo- globin levels (P 0.05). These 10 patients were compared with 10 matched healthy control subjects (Table 3). They differed from the control group in glycemic control and erythrocyte GSH. No difference was detected when comparing siblings of dia- betic patients with healthy control subjects (ferrocyanide 13 ± 6 vs. 14 ± 5 µmol ml1 h1, respectively) as well as siblings of patients with nephropathy with siblings of patients with retinopathy and normoalbu- Figure 1—Individual distribution of erythrocyte rate of ferricyanide reduction in diabetic subgroups minuria (ferrocyanide 14 ± 5 vs. 13 ± 5 1 1 and control subjects. Bars identify mean values. µmol ml h , respectively; other data not shown). Erythrocyte GSH showed sig- nificant negative correlation with fasting

(StatView, Abacus Concepts, Berkeley, CA) to plasma glucose, fructosamine, HbA1c, lac- plasma glucose (r = 0.3, P = 0.05) and using a Student’s unpaired t test or for not tate-to-pyruvate ratio, and plasma thiols. In fructosamine (r = 0.3, P = 0.01). In the dia- normally distributed data, a Mann-Whitney the patients, the only significant differences betic group without complications, a sig- U test. A 2 test was used to compare preva- were the lower erythrocyte concentrations of nificant relation was observed between lence among groups. A P value 0.05 was GSH and the higher ferricyanide-reducing ferricyanide reduction capacity of erythro- considered significant. Correlations were capacity in the nephropathy group compared cytes and plasma lactate concentration (r = sought by using stepwise multiple linear with the retinopathy group (Table 2). Indi- 0.8, P = 0.01; Fig. 2), whereas no relation regression analysis. Before regression analy- vidual distribution of erythrocyte ferricyanide was present in the diabetic groups with sis, nonparametric values were log-trans- reduction capacity in diabetic subgroups and microangiopathic complications. Among formed. Retrospectively, the patient sample control subjects is shown in Fig. 1. all patients with type 1 diabetes, multiple size had a 65% power to identify differences (P 0.05) in velocity of ferricyanide reduc- tion of a 30% magnitude (14). Table 3—Clinical and metabolic characteristics of 10 type 1 diabetic patients without complications versus 10 matched healthy volunteers RESULTS — Type 1 diabetic patients and control subjects were well matched for age, Type 1 diabetic subjects sex distribution, BMI, and arterial blood Characteristic without complications Control subjects P pressure (Table 1). The diabetic patients had significantly elevated levels of fasting plasma Sex (M/F) 4/6 3/7 — glucose, fructosamine, and glycated hemo- Age (years) 29 ± 10 30 ± 8 — Duration of diabetes (years) 14 ± 9 — — globin (HbA1c) compared with the control subjects. In type 1 diabetic patients, both ery- Insulin dose (U/kg) 0.6 ± 0.1 — — throcyte GSH levels and plasma SH groups BMI (kg/m2) 24 ± 2 22 ± 2 — were significantly lower than those in control Mean blood pressure (mmHg) 87 ± 10 85 ± 11 — subjects (Table 1). Changes in erythrocyte- Plasma glucose (mmol/l) 13 ± 4 5 ± 0.4 0.001 mediated ferricyanide reduction were not Fructosamine (µmol/l) 348 ± 47 229 ± 11 0.001 observed in the diabetic group compared HbA1c (%) 7.4 ± 1.3 5.2 ± 0.4 0.001 with the control group. The patients with a Plasma lactate (mmol/l) 1.5 (1.14–2.3) 1.7 (1.3–2.8) — known history of diabetic nephropathy (n = Plasma pyruvate (mmol/l) 0.05 (0.01–0.1) 0.04 (0.01–0.1) — 10) were compared with 10 type 1 diabetic GSH (mg/ml RBC) 0.70 ± 0.11 0.88 ± 0.16 0.05 patients of similar age, sex distribution, BMI, SH groups (µmol/l) 447 ± 51 438 ± 53 — and long duration of diabetes, who had Plasma albumin (mg/dl) 4,626 ± 323 4,782 ± 404 — retinopathy but normal UAERs (Table 2). Ferrocyanide (µmol ml 1 h 1) 13 ± 4 13 ± 5 — The 2 groups were comparable with regard Data are n, means ± SD, or medians (range).

996 DIABETES CARE, VOLUME 23, NUMBER 7, JULY 2000 Matteucci and Giampietro

inhibitors of glycolysis, p-chloromercu- riphenyl sulfonate (suggesting the involve- ment of SH groups on the outside of the membrane in the electron transfer process). In diabetes (as in other diseases involv- ing oxidative stress) the plasma and mem- brane may be observed to have lower thiol levels than in normal healthy volunteers. Indeed, our type 1 diabetic patients had lower levels of both erythrocyte GSH and plasma sulfhydryl groups. Reduced GSH concentration in erythrocytes showed negative correlations with short- and mean- term metabolic control parameters, such as fasting plasma glucose and fructosamine,

rather than with long-term HbA1c. Dis- crepancies among reports with regard to erythrocyte GSH in diabetes and its rela- tionship with metabolic parameters and diabetic complications (16–19) usually have been ascribed to differences in glycemic control, which was on average Figure 2—Linear association between plasma lactate concentration and ferricyanide reduction worse in our complicated groups. How- capacity of erythrocytes in 10 type 1 diabetic patients without diabetic complications. ever, this finding does not explain the dif- ference in erythrocyte GSH content between the nephropathy and retinopathy logistic regression of the electron transport tron flow was normal in siblings. The activ- groups who showed similar metabolic con- capacity gave positive correlations with uri- ity of the erythrocyte transmembrane- trol. With regard to the low GSH levels of nary albumin excretion and plasma creati- reducing system has been measured in a the patients without complications, abnor- nine and a negative correlation with plasma clinical setting now for the first time. Pre- malities of chain-breaking antioxidant sta- thiol concentration (multiple r = 0.6, P vious studies examined the mechanism by tus and erythrocyte GSH depletion have 0.01; Table 4). Plasma SH groups corre- which human erythrocytes transfer elec- been observed in patients with either type 1 lated negatively with disease duration (r = trons to extracellular ferricyanide and the and type 2 diabetes, both at disease onset 0.4, P = 0.01) and circulating albumin (r = ultimate source of the reducing equivalents and in later stages (8,16). No relationship 0.4, P = 0.01). Plasma albumin concentra- (1–3,10,15). The addition of an oxidating has been found between the level of GSH in tion decreased with increasing duration of agent such as ferricyanide or Ellman’s erythrocytes and the duration of diabetes diabetes (r = 0.4, P 0.05). reagent to the medium of erythrocyte sus- (17). Only a follow-up will tell us whether pensions has been suspected of inducing the patients with low levels of GSH in early CONCLUSIONS — The major novel the thiol-disulfide exchange reaction that stages of the disease are more prone to dia- observation in our study of type 1 diabetic occurs on the esofacial surface of the mem- betic nephropathy. patients with or without microangiopathic brane (4). As a result of the oxidative stim- The ratio of plasma lactate to pyru- complications, their siblings, and healthy ulus at the membrane surface, the cytosolic vate, which is a reliable parameter of the control subjects is the finding that the rate GSH pool is both oxidized and depleted to cytosolic ratio of free NADH to NAD (20), of ferrocyanide generation by erythrocytes regenerate the disulfide formed. This reac- has been considered a marker of hypergly- is significantly elevated in diabetic patients tion requires electrons and energy from the cemic pseudohypoxia (21). with nephropathy. The abnormality is cell. Few compounds have been found to In our patients without diabetic com- peculiar to the nephropathy group and act on transplasma membrane ferricyanide plications, cytosolic NADH was the major seemingly not contributed by familial or reduction by intact human erythrocytes determinant of erythrocyte electron transfer hereditary components because the elec- (15): dehydroascorbate supplementation, capacity (Fig. 2). In contrast, in patients

Table 4—Details of simple and multiple regression analysis between the electron transport capacity and plasma creatinine, log-transformed UAER, and plasma SH groups among all patients with type 1 diabetes

Simple regression Multiple regression Intercept Slope SEM rPIntercept 18.1 (slope) r = 0.6 (SEM) P 0.01 (t value) Ferrocyanide vs. creatinine 12.8 0.01 0.007 0.4 0.04 0.01 0.007 1.5 Ferrocyanide vs. UAER 10.0 1.51 0.6 0.4 0.01 1.0 0.6 1.6 Ferrocyanide vs. SH groups 24.1 0.02 0.013 0.3 0.05 0.02 0.013 1.5

DIABETES CARE, VOLUME 23, NUMBER 7, JULY 2000 997 Ferricyanide reduction in diabetic nephropathy with diabetic complications, the close rela- water molecules is the accelerating factor in Leguen C, Baxter MA, Thorpe GHG, Jones tionship between NADH generation and diabetic nephropathy). Which dynamics of AF, Barnett AH: Antioxidant status in transmembrane electron passage seemed to the intervening medium really accelerated patients with uncomplicated insulin-depen- be lost. Whereas patients with retinopathy the electron transfer along functional path- dent and non-insulin-dependent diabetes alone still had normal activity of the ery- ways is still to be examined. mellitus. Eur J Clin Invest 27:484–490, 1997 9. Ido Y, Kilo C, Williamson JR: Cytosolic throcyte-reducing system, patients with We observed that if too many electrons NADH/NAD , free radicals, and vascular nephropathy showed significantly increased are passed on at a time, harmful free radi- dysfunction in early diabetes mellitus. Dia- activity related only to renal function cals result that are candidate mechanisms betologia 40:5115–5117, 1997 parameters (plasma creatinine and urinary involved in the process of renal damage. 10. May JM, Qu Z-C, Whitesell RR: Ascorbic albumin) and plasma thiols. The data presented support the hypothesis acid recycling enhances the antioxidant The possibility that chronic renal fail- that the transmembrane-reducing system reserve of human erythrocytes. Biochem- ure and/or dialysis treatment could have of the erythrocytes is selectively involved in istry 34:12721–12728, 1995 contributed to our findings was excluded diabetic nephropathy, where erythrocyte 11. Avron M, Shavit N: A sensitive and simple by the composition of the nephropathy GSH is also depleted. Future studies are method for determination of ferrocyanide. group: no patients had renal failure but needed to investigate the underlying mech- Anal Biochem 6:549–554, 1963 one. To exclude any measurable contribu- anisms and the chronological order in 12. Beutler E, Duron O, Kelly BM: Improved method for the determination of blood glu- tion of the antihypertensive treatment to which these abnormalities develop. Ery- tathione. J Lab Clin Med 61:882–888, 1963 the increased rate of ferricyanide reduction throcyte transmembrane electron transfer 13. Ellman GL: Tissue sulfhydryl groups. Arch of the nephropathy group, we compared may be a useful tool to directly monitor cel- Biochem Biophys 82:70–77, 1959 the patients with complications who were lular redox state and, moreover, to evaluate 14. Altman DG: Statistics and ethics in medical taking drugs and those who were not: no efficacy of preventive interventions. research: how large a sample? Br Med J 281: significant difference emerged. On the 1336–1338, 1980 other hand, the indicated drugs have been 15. Himmelreich U, Kuchel PW: 13C-NMP reported to have no adverse effect on glu- Acknowledgments — We acknowledge the studies of transmembrane electron transfer cose metabolism (22); moreover, only one superb technical assistance of Vincenzo Cinapri to extracellular ferricyanide in human ery- and Stefano Quilici. throcytes. Eur J Biochem 246:638–645, 1997 patient was receiving the SH-containing 16. Dominguez C, Ruiz E, Gussinye M, Carras- , which seemed to preserve cosa A: Oxidative stress at onset and in early myocardium Q10 and cytochrome oxidase References stages of type 1 diabetes in children and activity (23). 1. Orringer EP, Roer MES: An ascorbate-medi- adolescents. Diabetes Care 21:1736–1742, In living organisms, thiols, thiolates, ated transmembrane-reducing system of the 1998 and disulfides are interdependent and con- human erythrocyte. J Clin Invest 63:53–58, 17. Jain SK, McVie R: Effect of glycemic con- stitute elements of an electron transfer sys- 1979 trol, race (white versus black), and duration tem along with metalloproteins (24). 2. May JM, Qu Z-C, Whitesell RR: Ascorbate is of diabetes on reduced glutathione content Parameters controlling the rates of electron the major electron donor for a transmem- in erythrocytes of diabetic patients. Metab- transfer reactions have been investigated by brane oxidoreductase of human erythro- olism 43:306–309, 1994 18. Di Simplicio P, De Giorgio LA, Cardaioli E, using a range of model systems (25). Earlier cytes. Biochim Biophys Acta 1238:127–136, 1995 Lecis R, Miceli R, Rossi R, Anichini R, Mian work established the rate dependence on 3. Himmelreich U, Drew KN, Serianni AS, M, Seghieri G, Franconi F: Glutathione, glu- driving force and distance. In addition, the Kuchel PW: 13C NMR studies of vitamin C tathione utilizing enzymes and thioltrans- medium separating the donor and acceptor transport and its redox cycling in human ferase in platelets of insulin-dependent also has to play an important role. The reor- erythrocytes. Biochemistry 37:7578–7588, diabetic patients: relation with platelet aggre- ganization energy () may be another 1998 gation and with microangiopathic compli- important factor for many systems, owing 4. Reglinski J, Hoey S, Smith WE, Sturrock cations. Eur J Clin Invest 25:665–669, 1995 to reorientation of water molecules. In the RD: Cellular response to oxidative stress at 19. Thornalley PJ, McLellan AC, Lo TWC, absence of diabetic complications, the rate sulphydryl group receptor sites on the ery- Benn J, Sönksen PH: Negative association of electron transfer was, as expected, pro- throcyte membrane. J Biol Chem 263: between erythrocyte glutathione concen- tration and diabetic complications. Clin Sci portional to the driving force (NADH-to- 12360–12366, 1988 5. Parker L: The role of anti-oxidative treat- 91:575–582, 1996 NAD ratio). At both equal oxidative ment in diabetes mellitus. Diabetologia 36: 20. Williamson DH, Lund P, Krebs HA: The stimulus and driving force, changes in the 1212–1213, 1993 redox state of free nicotinamide-adenine transfer matrix element have to account for 6. Sinclair AJ, Girling AJ, Gray L, Leguen C, dinucleotide in the cytoplasm and mito- the loss of the proportionality observed in Lunec J, Barnett AH: Disturbed handling of chondria of rat liver. Biochem J 103:514– the diabetic groups with microangiopathy. ascorbic acid in diabetic patients with and 527, 1967 The electron flow among the redox without microangiopathy during high dose 21. Williamson JR, Chang K, Frangos M, Hasan centers is in part controlled by protein ascorbate supplementation. Diabetologia KS, Ido Y, Kawamura T, Nyengaard JR, Van components (covalent switching). The 34:171–175, 1991 Den Enden M, Kilo C, Tilton RG: Hyper- development of diabetic complications may 7. Yoshida K, Hirokawa J, Tagami S, glycemic pseudohypoxia and diabetic com- plications. Diabetes 42:801–813, 1993 be significant for electron transfer rate, Kawakami Y, Urata Y, Kondo T: Weakened cellular scavenging activity against oxidative 22. Bonora E, Targher G, Alberiche M, owing to irreversible chemical modifica- stress in diabetes mellitus: regulation of glu- Bonadonna RC, Saggiani F, Zenere MB, tions and cross-links in proteins (26) or tathione synthesis and efflux. Diabetologia Uleri S, Muggeo M: Effect of chronic treat- reorientation of solvent water molecules 38:201–210, 1995 ment with lacidipine or on intra- (to determine if the reorientation of solvent 8. Maxwell SRJ, Thomason H, Sandler D, cellular partitioning of glucose metabolism

998 DIABETES CARE, VOLUME 23, NUMBER 7, JULY 2000 Matteucci and Giampietro

in type 2 diabetes mellitus. J Clin Endocrinol function and energy production in the left 25. Daizadeh I, Medvedev ES, Stuchebrukhov Metab 84:1544–1550, 1999 ventricle in rabbits with smoke mitochon- AA: Effect of protein dynamics on biologi- 23. Gvozdjakova A, Simko F, Kurcharska J, drial cardiomyopathy. Biofactors 10:61–65, cal electron transfer. Proc Natl Acad Sci U S A Braunova Z, Psenek P, Kyselovic J: Capto- 1999 94:3703–3708, 1997 pril increased mitochondrial coenzyme 24. Halliwell B: Switches in enzymes. Nature 26. Brownlee M: Glycation and diabetic com- Q10 level, improved respiratory chain 354:191, 1991 plications. Diabetes 43:836–841, 1993

DIABETES CARE, VOLUME 23, NUMBER 7, JULY 2000 999