Dehydroepiandrosterone Administration Counteracts Oxidative Imbalance and Advanced Glycation End Product Formation in Type 2 Diabetic Patients

  1. Enrico Brignardello, MD1,
  2. Cristina Runzo, MD2,
  3. Manuela Aragno, PHD3,
  4. Maria Graziella Catalano, MD4,
  5. Maurizio Cassader, PHD2,
  6. Paolo Cavallo Perin, MD2 and
  7. Giuseppe Boccuzzi, MD14
  1. 1Oncological Endocrinology Unit, San Giovanni Battista Hospital, Turin, Italy
  2. 2Department of Internal Medicine, University of Turin, Turin, Italy
  3. 3Department of Experimental Medicine and Oncology, University of Turin, Turin, Italy
  4. 4Department of Clinical Pathophysiology, University of Turin, Turin, Italy
  1. Address correspondence and reprint requests to Prof. Giuseppe Boccuzzi, Dipartimento di Fisiopatologia Clinica, Via Genova, 3-10126 Torino, Italy. E-mail: giuseppe.boccuzzi{at}


OBJECTIVE—Dehydroepiandrosterone (DHEA) has been shown to prevent oxidative stress in several in vivo and in vitro models. This study aimed to evaluate the effects of DHEA administration on oxidative stress, pentosidine concentration, and tumor necrosis factor (TNF)-α/TNF-α receptor system activity in patients with type 2 diabetes.

RESEARCH DESIGN AND METHODS—Twenty patients were enrolled in the study and randomly assigned to the DHEA (n = 10) or placebo (n = 10) group. Twenty healthy sex- and age-matched subjects with normal glucose levels served as control subjects. DHEA was given as a single daily dose of 50 mg for 12 weeks.

RESULTS—Oxidative stress parameters were significantly higher in diabetic patients versus control subjects. Pentosidine levels, as well as soluble TNF receptor (sTNF-R)I and sTNF-RII, were also higher in diabetic patients. After DHEA, plasma levels of reactive oxygen species and hydroxynonenal dropped by 53 and 47%, respectively, whereas the nonenzymatic antioxidants glutathione and vitamin E increased (+38 and +76%, respectively). The same changes in oxidative parameters were detected in peripheral blood mononuclear cells (PBMCs). DHEA treatment also induced a marked decrease of pentosidine plasma concentration in diabetic patients (−50%). Moreover, the TNF-α/TNF-α receptor system was shown to be less activated after DHEA treatment, in both plasma and PBMCs.

CONCLUSIONS—Data indicate that DHEA treatment ameliorates the oxidative imbalance induced by hyperglycemia, downregulates the TNF-α/TNF-α receptor system, and prevents advanced glycation end product formation, suggesting a beneficial effect on the onset and/or progression of chronic complications in type 2 diabetic patients.

The onset and progression of diabetes complications involves a complex interplay between ranges of pathogenic mechanisms. However, emerging evidence suggests that a single early phenomenon, i.e., the overproduction of superoxide by the respiratory chain, plays a key role in the pathogenesis of both microvascular and macrovascular chronic complications (13).

The production of advanced glycation end products (AGEs) is among the main mechanisms recruited by oxidative stress and is involved in the pathogenesis of tissue injury (3,4). AGEs progressively accumulate with time at the sites of diabetic microvascular disease and mediate tissue damage by activation of specific receptors at distant sites, via circulation (3,5). Moreover, the AGE/AGE receptor interaction, along with hyperglycemia-induced oxidative stress, possibly serves as a key activator of upstream kinases, leading to increased production of inflammatory cytokines thought to be involved in the progression of chronic diabetes complications (6).

Interruption of free radical overproduction by antioxidants counteracts AGE formation (2). Nevertheless, despite convincing experimental results, clinical trials with traditional antioxidants have been disappointing (7)—the activity of the antioxidants used in those trials is limited to scavenging already-formed oxidants and is stoichiometric. A compound of physiological origin that possesses multi-targeted antioxidant properties is dehydroepiandrosterone (DHEA), a multifunctional steroid that has been shown to prevent tissue damage induced by hyperglycemia in several in vivo and in vitro models (8,9). It also prevents the upregulation of AGE receptors observed in the hippocampus of streptozotocin-induced diabetic rats (10).

This study aimed to examine the effects of DHEA administration on oxidative stress, pentosidine (a marker of AGE-biogenesis) concentration, and tumor necrosis factor (TNF)-α/TNF-α receptors in patients with recently diagnosed type 2 diabetes, controlled with diet alone, and without any evidence of chronic complications, i.e., with the disease at a very early stage.


The study was approved by the ethics committee at our institution, and written informed consent was obtained from all recruited subjects. It was a randomized, double-blind, placebo-controlled, small-scale study of 12-week duration.

The study group comprised 20 patients with recently diagnosed type 2 diabetes treated with diet alone, having good glycemic control, and taking no drugs potentially interfering with redox status. All patients were nonsmokers and showed no evidence of chronic diabetes complications. The control group comprised 20 healthy subjects, matched by sex and age, with normal glucose levels.

After consent, patients were randomly assigned to the DHEA (n = 10) or the placebo (n = 10) group. DHEA was purchased from DHEAPharma (Miami, FL) and given as a single daily oral dose of 50 mg at 0800 h for 12 weeks. The 10 patients enrolled in the placebo group received pills that were identical in appearance to the DHEA formulation. Compliance was checked by pill counts. At baseline and at the end of the treatment, all patients were subjected to complete physical examination and fasting blood samples collected to evaluate oxidative stress parameters, pentosidine levels, and TNF-α/TNF receptors.

Serum DHEA, DHEA sulfate, glucose, serum insulin, homeostasis model assessment, and A1C

Serum DHEA and DHEA sulfate (DHEAS) were determined by specific radioimmunoassay (Diagnostic System Laboratories, Oxford, U.K.). Fasting serum glucose was measured by the glucose oxidase method (HITACHI 911 Analyzer; Sentinel Ch., Milan, Italy), and fasting serum insulin by immunoradiometric assay (Radim S.p.A., Pomezia, Italy). A1C was determined by standardized affinity high-performance liquid chromatography (BioRad). The homeostasis model assessment (HOMA) of insulin resistance index was calculated as the product of basal glucose and insulin levels divided by 22.5 (11).

Cytosol extracts from PBMCs

To isolate peripheral blood mononuclear cells (PBMCs), heparin-anticoagulated blood was subjected to density-gradient centrifugation on Lymphoprep (Fresenius Kabi Norge, Oslo, Norway). The cytosolic extracts were prepared as reported elsewhere (12). PBMCs were resuspended in buffer containing 20 mmol/l HEPES, pH 7.9, 1 mmol/l MgCl2, 0.5 mmol/l EDTA, 1% NP-40, 1 mmol/l EGTA, 1 mmol/l DTT, 0.5 mmol/l PMSF, 5 μg/ml aprotinin, and 2.5 μg/ml leupeptin, then centrifuged at 1,000g for 5 min at 4°C. Supernatants were centrifuged at 105,000g at 4°C for 40 min to obtain cytosolic fraction, and protein content was determined.


Plasma samples (200 μl) were treated with 6 mol/l hydrochloric acid for 2 h at 40°C and centrifuged (4,000 rpm), and 20 μl supernatant was injected.

A Thermo-Finnigan Surveyor instrument (Thermo Electron, Rodano, Italy), equipped with autosampler and PDA-UV 6,000 LP detector, was used. Mass spectrometry analyses were performed using a LCQ Deca XP Plus spectrometer, with electrospray interface and ion trap as mass analyzer.

The chromatographic separations were run on a Varian Polaris C18-A column (150 × 2 mm, particle size 3 μm) (Varian, Leinì, Italy), with a flow rate of 200 μl/min. Gradient mobile phase composition was adopted: 95/5 to 0/100 vol/vol 5 mmol/l heptafluorobutanoic acid in water/methanol for 13 min.

The liquid chromatography column effluent was delivered to a UV detector (200–400 nm) and then to the ion source using nitrogen as sheath and auxiliary gas (Claind Nitrogen Generator apparatus; Claind, Lenno, Italy). The tuning parameters adopted for the electrospray ionization source were as follows: source current 80.00 μA, capillary voltage 3.00 V, and tube lens offset 15 V; and for ions optics, multipole one offset −5.25 V, intermultipole lens voltage −16.00 V, and multipole two offset −9.00 V. Mass spectra were collected in tandem mass spectrometry (MS) mode: MS2 of (+) 379 m/z with 33% capillary electrophoresis in the range 100–400 m/z.

Oxidative biochemical parameters

Reduced glutathione (GSH), hydroxynonenal (HNE), and reactive oxygen species (ROS) levels were evaluated as reported elsewhere (12). For GSH levels, a mixture was directly prepared in a cuvette: 2.25 ml of 0.1 mol/l K-phosphate buffer, pH 8.0; 0.2 ml of the sample (plasma or PBMCs cytosol fractions); and 25 μl of Ellman’s reagent (10 mmol/l dithionitrobenzoic acid in methanol). After 1 min, the assay absorbance was measured at 412 nm and the GSH concentration calculated by comparison with a standard curve.

HNE was determined on plasma or PBMC cytosol fractions. An aliquot of cytosol (0.5 ml) was added to an equal volume of acetonitrile:acetic acid (96:4 vol:vol). After centrifugation, the supernatant was injected into a high-performance liquid chromatography (HPLC; Waters Associated, Milford, MA) Symmetry C18 column (5 mm, 3.9 × 150 mm). The mobile phase was acetonitrile:bidistilled water (42% vol:vol). The HNE concentration was calculated by comparison with a standard solution of HNE of known concentration.

ROS was measured in plasma or PBMC cytosol fractions using probe 2′,7′-dichlorofluorescin diacetate (DCFH-DA). DCFH-DA is a stable, nonfluorescent molecule that readily crosses the cell membrane and is hydrolyzed by intracellular esterases to the nonfluorescent 2′,7′-dichlorofluorescin (DCFH), which is rapidly oxidized, in the presence of peroxides, to highly fluorescent 2′,7′-dichlorofluorescein (DCF); DCF is measured fluorimetrically.

α-Tocopherol (vitamin E) was assayed using the method described by Burton et al. (13)—after extraction of 0.5 ml plasma with 1 ml n-heptane and centrifugation, the heptane phase was collected for HPLC analysis. A Supercosil-LC-Si column (25 cm × 4.6 mm; Supelco, Bellefonte, PA) was used, the mobile phase being n-hexane-isopropanol (99:1, vol:vol) and the flow rate 2.0 ml/min; the fluorescence detector was set to 298 nm excitation and 325 nm emission.

TNF-α and TNF receptors I and II

Serum levels

Serum levels of TNF-α, its soluble receptors TNF-RI (sTNF-RI) and TNF-RII (sTNF-RII) were measured using enzyme-linked immonosorbent assay tests (for TNF-α: Immunotech, Marseille, France; for sTNF-RI and sTNF-RII: Quantikine HS, R&D Systems, Minneapolis, MN) following the manufacturers’ instructions.

Real-time RT-PCR

Total RNA was extracted from PBMCs using TRIzol Reagent (Invitrogen, Groningen, the Netherlands), as reported elsewhere (12). DNase was added to remove remaining genomic DNA. RNA was reverse transcribed using the iScript cDNA Synthesis Kit (BioRad Laboratories).

Primers were designed using the Beacon 5 program. Primers used were as follows: for TNF-α, forward 5′-CGC CAC CAC GCT CTT CTG C, reverse 5′-GGG CTA CAG GCT TGT CAC TCG; for TNF-RI, forward 5′-CTG CCA GGA GAA ACA GAA CAC C, reverse 5′-GCG TCC TCA GTG CCC TTA ACA TTC; for TNF-RII: forward 5′-CGG TGT GGG CTG TGT CGT AG, reverse 5′- GAG GCT GCG GCT GTG GAG; and for β-actin, forward 5′-GCG AGA AGA TGA CCC AGA TC, reverse 5′-GGA TAG CAC AGC CTG GAT AG.

Real-time PCR was performed using a BioRad iQ iCycler Detection System with SYBR green fluorophore. Reactions were run in a total volume of 25 μl including 12.5 μl IQ SYBR Green Supermix (BioRad Laboratories), 1 μl of each primer at 10 μmol/l concentration, and 5 μl of the reverse-transcribed cDNA template. The protocol is as follows: denaturation (95°C for 5 min) and amplification repeated 40 times (95°C for 15 s, 60°C for 1 min). A melt curve analysis was performed to ensure a single amplified product for every reaction. All reactions were carried out in at least triplicate. Analysis of relative gene expression was performed using Gene Expression Macro software (BioRad Laboratories).

Statistical analysis

Results are presented as means ± SE. Differences between means were analyzed for significance using the Student’s t test. A P value <0.05 was considered significant.


Oxidative stress was significantly higher in diabetic patients than control subjects, as demonstrated by higher levels of ROS and HNE and reduced levels of GSH and vitamin E (Table 1). Pentosidine levels, as well as sTNF-RI and sTNF-RII, were also higher in diabetic patients than in control subjects, whereas serum TNF-α levels were within the normal range (Table 1).

Effects of DHEA on BMI and glycemic control

DHEA treatment had no effect on patients’ BMI. Basal glucose concentration, A1C level, and HOMA index were also unaffected (Table 2).

Effects of DHEA on oxidative state

After DHEA treatment, DHEA and DHEAS levels were significantly increased (P < 0.05, Table 3). Oxidative stress parameters were significantly modified by DHEA treatment, both in plasma and in PBMCs. Plasma levels of ROS dropped by 53% (Table 3 and Fig. 1A). Likewise, a 47% reduction of plasma HNE was observed after DHEA treatment (Table 3 and Fig. 1A), whereas plasma levels of the nonenzymatic antioxidants GSH and vitamin E increased by 38 and 76%, respectively (Table 3 and Fig. 1A). The same trends were found in PBMCs (Table 3 and Fig. 1B). No correlation was found between DHEA or DHEAS levels and both oxidative and antioxidant parameters either before or after DHEA treatment.

Effects of DHEA on pentosidine levels

DHEA treatment markedly decreased pentosidine plasma concentration in diabetic patients (−50%; Table 3 and Fig. 1A).

Effects of DHEA on TNF-α and TNF-α receptors

After DHEA treatment, no modification of serum TNF-α and sTNF-RI was observed (Table 3), whereas in PBMCs mRNA expression was reduced by 33 and 29%, respectively (Fig. 1C). On the contrary, serum sTNF-RII levels were reduced after DHEA treatment (Table 3), without any modification in mRNA expression (Fig. 1C).

All the above parameters were unchanged in diabetic patients given placebo (data not shown).


Type 2 diabetic patients with good glycemic control and no evidence of chronic diabetes complications, such as those enrolled in this study, show a redox imbalance characterized by increased production of highly reactive oxygen species and lower-than-normal antioxidant potential. We show here that DHEA treatment counteracts this oxidative imbalance—after 12 weeks, the concentrations of ROS and HNE were greatly reduced in both plasma and cytosol of PBMCs, whereas levels of the nonenzymatic antioxidants GSH and vitamin E were increased. These results are in agreement with the multi-targeted antioxidant effect of DHEA previously reported by our group (9,10,12). We could not find any correlation between serum DHEA or DHEAS levels and both the oxidative and the antioxidant parameters either before or after DHEA treatment. This result is not surprising, since DHEA is not a scavenger compound acting in a stoichiometric manner and exerts its antioxidants effects in a complex and non–completely defined way (10,12).

Moreover, DHEA treatment significantly reduced the plasma concentration of pentosidine in these patients, in line with its effects on AGEs and AGE receptors, which have been reported in experimental diabetes (12). AGEs, whose production is triggered by oxidative stress, are clearly implicated in the development and progression of chronic diabetes complications (3,4). Among AGEs, pentosidine is a well-characterized compound and is used as a marker of AGE biogenesis (14); it is considered a good predictor for the development of microvascular complications in diabetic patients (15). Interestingly, DHEA reduces pentosidine concentration in type 2 diabetic patients without any influence on glycemic control, strongly suggesting that this effect can involve its ability to improve redox balance.

Hyperglycemia-induced oxidative stress may, either directly or through the AGE/AGE receptor interaction, serve as a key activator of upstream kinases, leading to an increase in the plasma inflammatory cytokine concentrations that is thought to be involved in the progression of chronic diabetes complications (6), as well as in the development of insulin resistance (16).

Compared with control subjects, diabetic patients showed higher levels of sTNF-RI and sTNF-RII, whereas TNF-α plasma concentrations were similar. After binding of TNF-α to its receptors, cleavage of the extracellular parts elicits the soluble forms, known as sTNF-RI and sTNF-RII, which are thought to reflect the degree of TNF system activation (17). TNF-α is not a very stable protein, and its serum level may not adequately reflect its activity. On the contrary, serum concentration of soluble TNF-α receptors have been shown to be increased in diabetic patients (18) and the level of sTNF-RII has been considered the best predictor of TNF-α system activation, as well as a marker of insulin resistance (19). DHEA treatment reduced the serum concentration of sTNF-RII and mRNA expression of TNF-α and TNF-RI in PBMCs. A reduction in mRNA expression of TNF-α indicates a downregulation of this system, which is reflected by a concomitant reduction in TNF-RI expression. As far as TNF-RII is concerned, our observation of a reduction in its serum levels without any change in its mRNA expression implies reduced shedding of the protein, suggesting improved insulin sensitivity after DHEA treatment. A relationship between TNF-α system and insulin resistance has indeed been reported (20), as has a negative correlation between TNF-RII shedding and insulin sensitivity (21). Moreover, ROS, which are markedly reduced by DHEA treatment, have recently been reported to play a causal role in insulin resistance (16). In accordance with previous observations (22), the present data suggest that DHEA treatment might influence insulin sensitivity in type 2 diabetic patients, through its effects on the TNF-α system, despite the absence of HOMA index modifications that we observed.

In conclusion, these data, together with the experimental data from rodents, suggest that DHEA treatment might prevent many of the events that lead to cellular damage induced by hyperglycemia, thus counteracting the onset and/or progression of chronic complications in type 2 diabetic patients. Of interest, this result was obtained in the absence of any improvement of glycemic control. Present management of hyperglycemia is based upon the assumption that the best way to reduce the risk of diabetes complications is to achieve optimal glycemic control. However, it should be pointed out that patients receiving intensive therapy designed to achieve glycemic control still develop diabetes complications, even though their prevalence is reduced (23). If hyperglycemia cannot be effectively prevented, the only way to impede diabetes complications will be to interrupt the pathways that lead from hyperglycemia to target organ damage. AGEs clearly represent one such pathway.

A similar preventive activity against hyperglycemia-induced oxidative stress has been postulated for drugs such as statins, ACE inhibitors, angiotensin II type 1 receptor blockers, calcium channel blockers, and thiazolidinediones, and their clinical use to prevent chronic complications in diabetic patients has been suggested (24). Compared with these drugs, DHEA has the advantage of being a physiologic steroid without side effects at the dosage used in this study. The usefulness of this novel approach to protect diabetic patients against tissue damage appears to be worth further exploration through multicenter clinical trials.

Figure 1—

Effects of DHEA on variations of oxidative stress parameters, TNF-α, and TNF-α receptors. A: Changes in levels of ROS, GSH, vitamin E, HNE, and pentosidine in plasma of diabetic patients treated with DHEA for 12 weeks. B: Changes in levels of ROS, GSH, and HNE in PBMCs of diabetic patients treated with DHEA for 12 weeks. C: Changes in mRNA expression of TNF-α, TNF-RI, and TNF-RII in PBMCs of diabetic patients treated with DHEA for 12 weeks. Variation is expressed as percentage with respect to levels at baseline. Significance vs. baseline: *P < 0.05, **P < 0.01, ***P < 0.001.

Table 1—

BMI and serum concentrations of glucose, ROS, GSH, vitamin E, HNE, pentosidine, and cytokines in the study group at baseline

Table 2—

Characteristics of patients at baseline and after DHEA treatment (n = 10) or placebo administration (n = 10)

Table 3—

Effects of DHEA treatment on DHEA, DHEAS, oxidative and antioxidant parameters, and TNF-α/TNF-α receptor system in plasma and/or PBMCs


This study was supported by the Special Project “Oncology,” Compagnia San Paolo, Turin, by MIUR (Ministero Università e Ricerca), and by Regione Piemonte.


  • Published ahead of print at on 17 August 2007. DOI: 10.2337/dc07-1110.

    A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.

    The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C Section 1734 solely to indicate this fact.

    • Accepted August 9, 2007.
    • Received June 22, 2007.


| Table of Contents

This Article

  1. Diabetes Care vol. 30 no. 11 2922-2927
  1. All Versions of this Article:
    1. dc07-1110v1
    2. 30/11/2922 most recent