Ebselen does not improve oxidative stress and vascular function in patients with diabetes: a randomized, crossover trial

This study tested the glutathione peroxidase mimetic ebselen in clinically relevant doses and found no significant impact on either oxidative stress or conduit artery vascular function in subjects with diabetes.

Abstract

Oxidative stress is a key driver of vascular dysfunction in diabetes mellitus. Ebselen is a glutathione peroxidase mimetic. A single-site, randomized, double-masked, placebo-controlled, crossover trial was carried out in 26 patients with type 1 or type 2 diabetes to evaluate effects of high-dose ebselen (150 mg po twice daily) administration on oxidative stress and endothelium-dependent vasodilation. Treatment periods were in random order of 4 wk duration, with a 4-wk washout between treatments. Measures of oxidative stress included nitrotyrosine, plasma 8-isoprostanes, and the ratio of reduced to oxidized glutathione. Vascular ultrasound of the brachial artery and plethysmographic measurement of blood flow were used to assess flow-mediated and methacholine-induced endothelium-dependent vasodilation of conduit and resistance vessels, respectively. Ebselen administration did not affect parameters of oxidative stress or conduit artery or forearm arteriolar vascular function compared with placebo treatment. There was no difference in outcome by diabetes type. Ebselen, at the dose and duration evaluated, does not improve the oxidative stress profile, nor does it affect endothelium-dependent vasodilation in patients with diabetes mellitus.

NEW & NOTEWORTHY

This study tested the glutathione peroxidase mimetic ebselen in clinically relevant doses and found no significant impact on either oxidative stress or conduit artery vascular function in subjects with diabetes.

cardiovascular diseases are the primary cause of death and disability in patients with diabetes mellitus (3). The metabolic abnormalities accompanying diabetes, including hyperglycemia, liberation of excess nonesterified fatty acids, and insulin resistance, create an environment permissive for the development of endothelial dysfunction and atherosclerosis. One mechanism by which these metabolic disturbances may cause endothelial dysfunction is generation of oxidative stress.

Oxidative stress induced by diabetes mellitus is chronic and progressive. Patients with diabetes have increased oxidative stress (27, 28), and reduced glutathione concentrations in human erythrocytes inversely correlate with severity of chronic hyperglycemia (33). Duration of diabetes is inversely correlated with concentrations of endogenous antioxidants (32). Mechanisms that impact oxidative stress may be important to vascular dysfunction and atherosclerosis.

The relative importance of the different oxygen-derived free radicals, including, but not limited to, superoxide anion (O₂⁻), hydroxyl radical, peroxynitrite, and H₂O₂, remains incompletely understood in humans. In animal models, excess O₂⁻ is associated with atherogenesis, whereas induction of superoxide dismutase retards development and progression of atherosclerosis (6, 39). We previously demonstrated that infusion of ascorbate at doses sufficient to compete with superoxide dismutase and scavenge superoxide restores the bioavailability of endothelium-derived nitric oxide in patients with type 1 and type 2 diabetes mellitus (T1DM and T2DM) (34, 35). The contribution of H₂O₂ to vascular function in patients with diabetes, however, is not established.

The glutathione peroxidase (GPx) family of antioxidant enzymes converts H₂O₂ to water via a series of coupled reactions that oxidize reduced glutathione to glutathione disulfide in a NADPH-dependent manner, playing an important role in preservation of endothelial function. GPx reduces H₂O₂ to two water molecules by oxidizing reduced glutathione. Glutathione reductase, by oxidizing NADPH, restores reduced glutathione. In rodents, absence of GPx is associated with increases in oxidative stress, uncoupling of endothelial nitric oxide synthase (eNOS), and reduced endothelium-dependent vasodilation (25). In humans with diabetes, plasma GPx, but not catalase or superoxide dismutase, is upregulated with increasing states of oxidative stress (32). Despite increased expression, glutathione-dependent H₂O₂ degradation is decreased by 50% in the endothelium, suggesting that deficiencies in cofactors contribute to lower enzyme activity (11). Decreased intracellular NADPH has been demonstrated in endothelial cells exposed to high glucose in vitro (11), which limits the ability of the endothelium to regenerate reduced glutathione. Inadequate GPx activity and peroxide detoxification capacity may create an environment that promotes increased oxidative stress and adversely affects vascular function.

Ebselen, a seleno-organic compound, is a GPx mimetic that catalyzes the reduction of H₂O₂ at the expense of a thiol (5). Ebselen primarily binds albumin covalently in plasma but can move and act intracellularly via interaction with thiols (8). It has been demonstrated to act as an effective GPx mimetic, specifically in the albumin-bound form in whole blood. Thus the aim of this study is to determine whether ebselen, a GPx mimetic, can decrease oxidative stress and improve endothelium-dependent vasodilation in patients with diabetes.

METHODS

Subjects.

Twenty-six volunteers with diabetes (16 T1DM and 10 T2DM, 12 women and 14 men) were recruited by newspaper advertisement and gave written informed consent. All subjects underwent screening history, physical examination, and laboratory analysis, including complete blood count, serum electrolytes, fasting glucose, blood urea nitrogen, creatinine, transaminases, alkaline phosphatase, and a lipid profile. Individuals with treated hypertension were eligible. Exclusion criteria included tobacco use within 1 yr, LDL cholesterol or total cholesterol greater than the 75th percentile for age and sex, use of statins, cardiovascular disease, and diabetes-related microvascular complication or other active diseases, including chronic kidney disease, liver disease, human immunodeficiency virus infection, and malignancy. Cyclooxygenase inhibitors, alcohol, and caffeine were prohibited for 12 h before study initiation. Four healthy subjects [2 men and 2 women, mean age 36 yr, mean body mass index (BMI) 26] who met the above criteria, but without diabetes, were recruited for fasting measures of oxidative stress alone to establish the levels of oxidants to be expected in healthy subjects tested with these techniques. None of the four healthy subjects had known medical conditions or used medications. The protocol was conducted at a single academic site, approved by the Human Research Committee of Brigham and Women's Hospital, and listed at clinicaltrials.gov (NCT00762671).

Intervention.

The effect of ebselen on conduit artery and forearm arteriolar endothelial function was studied using a randomized, double-masked, placebo-controlled, crossover design.

Clinical investigators and participants were masked to treatment assignment. Treatment arms were ebselen (150 mg po twice daily; Rhone Poulenc Rorer, Strasbourg, France) and matching placebo. The dose of ebselen was chosen on the basis of dose-ranging and kinetic studies (16, 18) and a clinical trial of ebselen in stroke patients in whom functional recovery improved compared with placebo (38). Treatment periods were 4 wk each, with a 4-wk washout interval between treatments; 4-wk periods were chosen to ensure repeat evaluation at the same phase of the female reproductive cycle. The Investigational Drug Service at Brigham and Women's Hospital performed randomization using a sequential number algorithm at the time of first study drug allocation and maintained the study concealment until trial lock. Treatment adherence was assessed by pill count. Study assessment visits occurred at the end of each treatment period, conducted in the morning after overnight fast.

Forearm arteriolar endothelial function.

Bilateral forearm blood flow (FBF) was measured by venous occlusion, mercury-in-Silastic, strain-gauge plethysmography by established methods (1). During data acquisition, wrist cuffs were inflated to 200 mmHg to exclude the hand circulation. A venous occlusion pressure of 40 mmHg was generated by cuffs placed on each arm above the elbow for each measurement of blood flow, which is reported as milliliters per 100 ml of tissue per minute. Arterial blood pressure was measured via the brachial artery cannula, which was attached to a pressure transducer contiguous with an amplifier on a Gould physiological recorder. Heart rate was determined by the R-R interval of a continuous ECG monitor. The vascular research laboratory was quiet, dimly lit, and temperature-controlled at 23°C. Participants rested for ≥30 min after insertion of the catheters before baseline hemodynamic data were acquired.

Effects of ebselen and placebo treatment on endothelial function were investigated in all participants with diabetes. First, in the postprandial condition, basal FBF and the blood flow response to 4-min intra-arterial infusions of incremental doses of methacholine chloride (0.3, 1.0, 3.0, and 10.0 μg/min) were assessed to determine endothelium-dependent vasodilation. To ascertain whether ebselen affected vascular smooth muscle function, a subset of 18 subjects (10 T1DM and 8 T2DM) were studied on a separate occasion with the calcium channel blocker verapamil at doses of 10, 30, 100, and 300 μg/min. FBF was measured during ebselen and placebo treatment conditions.

Conduit artery function.

Brachial artery ultrasonography was performed in a temperature-controlled, quiet environment to evaluate vascular function after each treatment period. An ultrasound scanner (Vivid 7, General Electric) equipped with a high-resolution broad-band linear-array transducer (7.5–12.5 MHz) was used to image the brachial artery. Flow-mediated endothelium-dependent (FMD) and nitroglycerin-mediated endothelium-independent dilation of the brachial artery were determined as we reported previously and according to established guidelines (7, 24). In our laboratory the sphygmomanometric cuff is placed above the olecranon. We previously demonstrated that this cuff location induces endothelium-dependent, nitric oxide-mediated, flow-mediated vasodilation (FMD) (14, 26).

Two investigators used software from Medical Imaging Applications to acquire and analyze the digitized images in a blinded manner. The vessel wall-lumen interface was determined using a derivative-based edge-detection algorithm after identification of the region of interest by the investigator. The maximum diameter of the vessel was then determined. In our laboratory, this method is associated with an interobserver variability of 0.05 ± 0.16% and intraobserver variability of 0.01 ± 0.15%. The same arm and site were used for all measurements.

Laboratory analysis.

The routine laboratory tests were performed in the Clinical Laboratory Improvement Amendments-certified clinical laboratory at Brigham and Women's Hospital. Blood for oxidative stress assessments was collected after each treatment period, centrifuged, and stored in aliquots in a −80°C freezer for future analyses.

We assessed oxidative stress by determining the ratio of reduced to oxidized glutathione (GSH/GSSG) and 8-isoprostane levels, as described previously (20, 30). Plasma GSH levels were measured using the Bioxytech GSH-400 enzymatic method (Oxis Research) according to the manufacturer's instructions. Results were read at 412 nm on a microplate reader (Spectramax Gemini XS, Molecular Devices). Plasma nitrotyrosine levels were measured by enzyme immunoassay using the Bioxytech nitrotyrosine enzyme immunoassay kit (Oxis Research) according to the manufacturer's instructions. Results were read at 450 nm on a microplate reader (Spectramax Gemini XS). Plasma was stored at −80°C until the time of assay, and samples were measured in duplicate. Total 8-isoprostane levels were measured in plasma utilizing the 8-isoprostane ELISA kit (Cayman Chemical) according to the manufacturer's instructions. For measurement of total 8-isoprostanes, samples were treated with 15% KOH and neutralized with 1 mol/l potassium phosphate buffer, pH 7.4. Results were read at 415 nm on a microplate reader (Spectramax Gemini XS). Samples were standardized for comparison by protein levels measured using a Bradford assay (13, 19).

Statistical analysis.

The sample size was modeled on our previous work using ascorbate as an antioxidant in subjects with diabetes. In our previous work, successful antioxidant therapy required 10 subjects to improve endothelium-dependent vasodilation in subjects with T1DM (34) and T2DM (35). Baseline characteristics are described as mean and SD or median and interquartile range. Baseline characteristics between groups were compared using ANOVA, Kruskal-Wallis equality-of-populations test, or χ² test, as appropriate. The primary outcome assessment was brachial artery FMD response during ebselen therapy compared with placebo treatment. The distribution of FMD at both time points was normally distributed [P = not significant (NS) by Shapiro-Wilk test]. Consequently, a paired t-test was performed to assess the difference in mean FMD between ebselen and placebo (the primary end point). We then used repeated-measures ANOVA to assess carryover effect with terms for treatment (ebselen vs. placebo), period (ebselen given first or second), and their interaction. For arteriolar function, basal FBF and laboratory measures were compared by paired 2-tailed t-tests. Statistical analyses of the dose-response curves for each drug (methacholine and verapamil) were conducted by the absolute increase in blood flow from the resting flow rate. Prespecified subgroup analysis evaluated T1DM and T2DM separately. Other subgroup descriptive analyses were performed by paired Student's t-test or Wilcoxon's signed-rank test, as needed. P < 0.05 was considered statistically significant. SPSS version 23 for Mac (IBM, Armonk, NY) was used for all analyses.

RESULTS

Twenty-six patients with diabetes were enrolled, randomly assigned, received both ebselen and placebo interventions, and were analyzed for the primary outcome. The trial was completed when the full intended cohort was enrolled. Compliance was confirmed by pill count. The participants had a mean age of 41 yr, included 12 women, and had normal lipid levels and renal function (Table 1). The mean BMI was 30 kg/m². No subject had proteinuria, retinopathy, or neuropathy. The only difference in demographics between the T1DM and T2DM patients was BMI: 28 and 33 kg/m² (P < 0.01), respectively. We measured oxidative stress markers in four healthy subjects and found evidence of oxidative stress in T1DM and T2DM patients compared with the healthy subjects (Table 2). There was no difference in these markers between the T1DM and T2DM patients.

Table 1.

Demographics

	All (n = 26)	T1DM (n = 16)	T2DM (n = 10)
Age, yr	41 ± 11	40 ± 12	43 ± 8
Female/male	12/14	8/8	4/6
Duration of diabetes, yr	17 ± 9	24 ± 7	6 ± 4
Body mass index, kg/m2	30 ± 3	28 ± 2	33 ± 3*
Hemoglobin A1c, %	8.2 ± 1.2	8.2 ± 0.9	8.2 ± 1.7
Mean arterial pressure, mmHg	99 ± 8	100 ± 8	98 ± 7
Cholesterol, mg/dl
Total	174 ± 32	175 ± 31	171 ± 34
HDL	52 ± 12	58 ± 5	46 ± 15
LDL	100 ± 33	96 ± 36	106 ± 29
Creatinine, mg/dl	0.9 ± 0.2	0.9 ± 0.1	0.8 ± 0.2

Values are means ± SD. T1DM and T2DM, types 1 and 2 diabetes mellitus.

^*P < 0.01.

Table 2.

Baseline oxidative stress

	Subjects			P Value
	T1DM	T2DM	Healthy
GSSG, mmol·l−1·mg protein−1	17.5 ± 3.4	15.9 ± 2.5	5.0 ± 1.7	<0.001
GSH, mmol·l−1·mg protein−1	1269 ± 146	1219 ± 92	1297 ± 180	0.59
GSH/GSSG	146 ± 25	151 ± 22	653 ± 43	<0.001
Nitrotyrosine, μmol·l−1·mg protein−1	552 ± 252	611 ± 362	69 ± 8	0.01
8-Isoprostane, pg·ml−1·mg protein−1	36.8 ± 10.0	42.3 ± 4.7	34.3 ± 4.6	0.2

Values are means ± SD. GSSG, glutathione disulfide; GSH, glutathione.

Effect of ebselen on oxidative stress and other laboratory measures.

Compared with placebo, ebselen had no effect on mean arterial pressure, BMI, glucose, hemoglobin A_1c, creatinine, or lipid levels (Table 3). There was no difference in the markers of oxidative stress, plasma 8-isoprostane, reduced and oxidized glutathione, and nitrotyrosine levels, between ebselen and placebo treatment in T1DM and T2DM patients (Table 3) or when T1DM and T2DM subgroups were considered separately (data not shown).

Table 3.

Effect of ebselen on laboratory parameters

	Treatment
	Placebo	Ebselen	P Value
Glucose, mg/dl	177 ± 80	195 ± 87	NS
Creatinine, mg/dl	0.71 ± 0.15	0.78 ± 0.14	0.09
ALT, mg/dl	19.4 ± 11.6	26.0 ± 27.4	NS
Hemoglobin A1c, %	7.6 ± 0.9	7.7 ± 0.9	NS
Cholesterol, mg/dl
Total	171 ± 43	165 ± 40	0.15
HDL	46 ± 14	44 ± 16	NS
LDL	104 ± 23	104 ± 21	NS
8-Isoprostane, pg·ml−1·mg protein−1	38.3 ± 8.4	37.5 ± 8.9	NS
GSSG, mmol·l−1·mg protein−1	17.0 ± 3.1	17.5 ± 4.0	NS
GSH, mmol·l−1·mg protein−1	1,252 ± 131	1,192 ± 142	NS
GSH/GSSG	146 ± 23	150 ± 22	NS
Nitrotyrosine, μmol·l−1·mg protein−1	564 ± 294	603 ± 418	NS

Values are means ± SD. ALT, alanine aminotransferase; NS, not significant.

Effect of ebselen on vascular function.

Basal FBF did not differ between ebselen and placebo treatment (1.8 ± 0.2 and 2.2 ± 0.4 ml·100 ml⁻¹·min⁻¹, respectively, P = NS). Incremental doses of intra-arterial methacholine increased FBF after both treatments, and the response did not vary by treatment condition (P = NS; Table 4). Similarly, there was no significant difference in the FBF response to verapamil between ebselen and placebo treatment (P = NS). Adjustment for diabetes type and hemoglobin A_1c did not change the effect of diabetes on the response to intra-arterial methacholine or verapamil (data not shown).

Table 4.

Forearm arteriolar vascular function

	Methacholine Dose, μg/min				P Value
	0.3	1	3	10
Placebo	2.30 ± 0.27	3.99 ± 0.53	7.04 ± 0.91	13.00 ± 1.8
Ebselen	1.71 ± 0.22	4.44 ± 2.46	8.16 ± 0.93	14.28 ± 1.43	NS

	Verapamil Dose, μg/min
	10	30	100	300	P Value
Placebo	1.00 ± 0.27	1.99 ± 0.48	4.32 ± 1.16	6.87 ± 2.05
Ebselen	1.22 ± 0.29	2.40 ± 0.59	5.22 ± 1.24	7.38 ± 1.89	NS

Values are means ± SD.

Baseline brachial artery diameter and the postocclusion reactive hyperemia stimulus used to increase flow and induce endothelium-dependent vasodilation were not different after ebselen and placebo therapy (Table 5). Mean FMD was not different after ebselen treatment (10.6 ± 1.1% and 9.8 ± 1.1% after ebselen and placebo, respectively, P = NS). Nitroglycerin-mediated, endothelium-independent vasodilation was not different between ebselen and placebo treatment (18.1 ± 1.9% and 16 ± 1.1%, respectively, P = NS). In separate models, repeated-measures ANOVA showed no significant carryover effect (treatment-period interaction, P = NS) and no significant difference in treatment effect between the diabetic groups (treatment-group interaction, P = NS).

Table 5.

Conduit artery vascular function

	Treatment		P Value
	Placebo	Ebselen
	Flow-mediated vasodilation
Baseline diameter, mm	3.66 ± 0.13	3.55 ± 0.14	0.09
RH diameter, mm	4.00 ± 0.12	3.92 ± 0.13	0.2
FMD, %	9.8 ± 1.1	10.6 ± 1.1	NS
RH stimulus, fold increase	6.1 ± 1.0	5.8 ± 0.9	NS
	Nitroglycerin-mediated vasodilation
Baseline diameter, mm	3.71 ± 0.12	3.62 ± 0.14	0.14
NTG diameter, mm	4.28 ± 0.12	4.23 ± 0.12	NS
NMD, %	16.0 ± 1.1	18.1 ± 1.9	0.2

Values are means ± SE. RH diameter, brachial artery diameter after reactive hyperemia (RH); FMD, flow-mediated vasodilation; NTG diameter, brachial artery diameter after nitroglycerin; NMD, nitroglycerin-mediated vasodilation.

There was no correlation between baseline diameter, reactive hyperemic stimulus, and FMD and any measure of oxidative stress. The medication was well tolerated by all subjects, save one who suffered a minor allergic reaction that responded to diphenhydramine. Compliance was assured by pill count. Three subjects missed one dose of a medication (2 placebo, 1 ebselen), and one subject missed two doses (ebselen). No subject missed a dose on the day of the evaluation.

DISCUSSION

Flow-mediated conduit artery and methacholine-induced forearm arteriolar endothelium-dependent vasodilation are impaired in T1DM and T2DM patients (2, 10, 37). We previously demonstrated that endothelial function in diabetes was attenuated, in part, because of an increase in oxidative stress (34, 35). In our previous work, FMD was ∼12% in healthy subjects (21), and the response to the peak dose of methacholine was ∼20 ml·100 ml⁻¹·min⁻¹ (35). The vascular function results in the diabetic patients of this study were attenuated compared with our previous experience with healthy subjects. In this investigation we examined the effect of ebselen, a GPx mimetic, on oxidative stress and vascular function in patients with diabetes. Using a dosing strategy informed by dose-ranging, kinetic, and stroke outcome studies (16, 18, 38), we report that ebselen failed to alter oxidative stress and vascular function.

Oxidative stress and endothelial function in diabetes.

An increase in the production of oxygen-derived free radicals has been demonstrated in diabetes. We report the same findings in this study, confirming previous findings and assay conditions. This increase in reactive oxygen species has been linked to decreased endothelium-derived nitric oxide bioavailability in humans. On the basis of a series of investigations, including our own, intra-arterial administration of ascorbic acid at doses needed to compete effectively with superoxide dismutase for O₂⁻, 3–10 mmol/l, improved endothelium-dependent vasodilation (1, 31). Oral antioxidant therapy, which is unable to match this concentration, fails to restore vascular function in most (23), but not all (2), studies, implicating O₂⁻ as an important determinant of endothelium-dependent vasodilation.

The primary sources of O₂⁻ in endothelial cells are enzymatic systems, including NADPH oxidases, uncoupled eNOS, mitochondrial respiration, and xanthine oxidoreductases (4). These enzymes reduce molecular oxygen to create O₂⁻, which either self-dismutates or is oxidized by superoxide dismutase to H₂O₂. H₂O₂ may exert its effects through oxidative modification of a variety of proteins, including protein kinases, phosphatases, transcription factors, ion channels, and metabolic enzymes. For example, H₂O₂ may activate protein kinase C or mitogen-activated protein kinases. As a second messenger that regulates the activity of signaling proteins and enzymes, H₂O₂ modulates a variety of endothelial cell effects.

The effect of H₂O₂ on vascular function differs from that of O₂⁻. O₂⁻ typically promotes vasoconstriction by decreasing the bioavailability of nitric oxide and generating peroxynitrite. H₂O₂, however, has vasodilator and vasoconstrictor properties. It causes vasodilation in mesenteric, pulmonary, and coronary arteries (4). It has been shown to be an endothelium-dependent hyperpolarizing factor, as well as an activator of eNOS through activation of phosphoinositide 3-kinase/Akt and Erk1/2. However, H₂O₂ also increases the degradation of nitric oxide by fostering the reaction between O₂⁻ and nitric oxide, and it can stimulate NAD(P)H oxidase, reduce levels of tetrahydrobiopterin, and promote eNOS uncoupling, a state associated with production of O₂⁻ and, thereby, less vasodilation (22). Moreover, conversion of H₂O₂ to hydroxyl radical may also impair endothelium-dependent vasodilation.

GPx serves as the primary endogenous mechanism for reducing plasma-based H₂O₂ (17). GPx is a selenium-dependent protein, and increased expression and activity of GPx in rats, through increased dietary selenium, augments acetylcholine-induced aortic ring dilation. This endothelium-dependent vasodilation is blocked by nitro-l-arginine methyl ester, indicating the role of GPx in increasing the bioavailability of nitric oxide (15). These observations were extended to humans with coronary artery disease by Vita and colleagues, who improved endothelium-dependent vasodilation by augmenting intracellular glutathione with administration of l-2-oxothiazolidine-4-carboxylic acid (36). Diabetes is associated with impaired glutathione synthesis, glutathione deficiency, and oxidative stress, all of which are reversed with cysteine and glycine administration to increase glutathione levels (29). Similarly, GPx activity is reduced in diabetic patients compared with healthy subjects (9). GPx activity in atherosclerotic lesions is markedly reduced, implying a role for impaired peroxide detoxification in the development of atherosclerosis (12).

Ebselen, oxidative stress, and vascular function in humans.

To examine the role of H₂O₂ in diabetes, we administered ebselen, a GPx mimetic. Ebselen has demonstrated safety in human use. In a double-blind, placebo-controlled trial, ebselen improved the functional recovery of patients with stroke as measured by the Glasgow Outcome Score without significant difference in adverse reactions compared with placebo (38).

In our hands, ebselen did not affect parameters of oxidative stress, conduit artery vascular function, or forearm arteriolar vascular function. There are several reasons for this. 1) The use of ebselen in humans has been rare and limited to neurological outcome studies. There has been no direct study in humans evaluating its effect on oxidative stress, so the dose may have been inadequate. 2) H₂O₂ plays both a salutary and an adverse role in the endothelium, such that nonspecific scavenging of H₂O₂ may not have significantly altered the oxidative balance. 3) Similarly, as a mild vasodilator, activator of eNOS, activator of NAD(P)H oxidase, and oxidizer of tetrahydrobiopterin, the effect of ebselen may just have been balanced. More thorough pharmacology would be necessary to understand proper dosing with ebselen for future studies.

Safety.

One subject had a mild allergic reaction to ebselen that responded to diphenhydramine. The medication was otherwise well tolerated.

Limitations.

This investigation has three limitations. 1) Administration of ebselen was limited to clinically successful doses tried previously in humans. Whether another dose of ebselen may have affected parameters of oxidative stress significantly is not known and cannot be determined by this investigation. 2) It is possible that a larger subject group may have demonstrated an effect of ebselen. However, based on the size of the changes noted in biochemical results and vascular function, we would need to increase the population by >10-fold to see a significant effect. 3) These diabetic patients were relatively healthy, and a sicker group may have had a response. We believe that this is not the case, as there was evidence of a significant increase in oxidative stress in both types of diabetes, but it is possible that patients with a more severe vascular pathology may have responded to the medication. Also, it is possible that novel techniques, such as cubital vein endothelial cell harvest and ex vivo analysis, may provide additional information about endothelial cell-specific oxidative conditions in response to ebselen treatment.

Conclusion.

In this single-site, randomized, double-masked, placebo-controlled, crossover trial of 4 wk of ebselen therapy in adults with diabetes and increased oxidative stress, we observed that ebselen affected neither markers of oxidative stress nor conduit and arteriolar vascular function. These data raise the following question: Is ebselen an effective antioxidative agent, at least at the dose and for the duration given? Further work is needed to understand the role of GPx and H₂O₂ in oxidative stress and vascular function in patients with diabetes.

GRANTS

This study was supported by National Institutes of Health Grants K23 HL-04169 and P30 DK-036836.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.A.B., A.B.G., and M.A.C. developed the concept and designed the research; J.A.B. and J.A.L. performed the experiments; J.A.B. analyzed data; J.A.B., A.B.G., J.A.L., and M.A.C. interpreted the results of the experiments; J.A.B. drafted the manuscript; J.A.B., A.B.G., J.A.L., and M.A.C. approved the final version of the manuscript; A.B.G., J.A.L., and M.A.C. edited and revised the manuscript.