The role of oxidative stress in the onset and progression of diabetes and its complications: asummary of a Congress Series sponsored byUNESCO-MCBN, the American Diabetes Association and the German Diabetes Society



This review summarises the results and discussions of an UNESCO-MCBN supported symposium on oxidative stress and its role in the onset and progression of diabetes. There is convincing experimental and clinical evidence that the generation of reactive oxygen species (ROI) is increased in both types of diabetes and that the onset of diabetes is closely associated with oxidative stress. Nevertheless there is controversy about which markers of oxidative stress are most reliable and suitable for clinical practice. There are various mechanisms that contribute to the formation of ROI. It is generally accepted that vascular cells and especially the endothelium become one major source of ROI. An important role of oxidative stress for the development of vascular and neurological complications is suggested by experimental and clinical studies. The precise mechanisms by which oxidative stress may accelerate the development of complications in diabetes are only partly known. There is however evidence for a role of protein kinase C, advanced glycation end products (AGE) and activation of transcription factors such as NFκB, but the exact signalling pathways and the interactions with ROI remain a matter of discussion. Additionally, results of very recent studies suggest a role for ROI in the development of insulin resistance. ROI interfere with insulin signalling at various levels and are able to inhibit the translocation of GLUT4 in the plasma membrane. Evidence for a protective effect of antioxidants has been presented in experimental studies, but conclusive evidence from patient studies is missing. Large-scale clinical trials such as the DCCT Study or the UKPDS Study are needed to evaluate the long-term effects of antioxidants in diabetic patients and their potential to reduce the medical and socio-economic burden of diabetes and its complications. Copyright © 2001 John Wiley & Sons, Ltd.


Diabetes (and its complications) is one of the major disorders worldwide. The prevalence of diabetes will rise from 6% to over 10% in the next decade. The major contributor to the increasing number of diabetic patients will be type 2 diabetes which is characterised by an inability to utilise glucose in the presence of insulin. The future increase of type 2 diabetes will not only be a problem faced by the so-called developed countries but also by the developing world. Therefore UNESCO, the World Organisation for Education, decided to organise a scientific meeting on diabetes and its complications in order to highlight the medical and socio-economic burden of this disorder. Since there is no global therapeutic solution to treating diabetes and its complications, it was felt that scientific discussions involving representatives from political organisations about possible therapeutic solutions should be included, based on the current state of research knowledge. The organising committee, comprising Prof. A. Azzi (Bern, Switzerland), Prof. L. Packer (Berkeley, CA, USA), Prof. P. Rösen (Düsseldorf, Germany), Dr. Hans-J. Tritschler (Frankfurt, Germany) and Prof. G. King (Boston, MS, USA), decided to focus this first UNESCO-sponsored scientific conference on diabetes research on the pathogenic role of oxidative stress in the onset and progression of diabetes and its complications.

There is emerging evidence that diabetes leads to depletion of the cellular antioxidant defence system and increased levels of reactive oxygen species (ROI). This new concept of oxidation stress, being an important trigger in the onset and progression of diabetes and its complications, may offer a unique therapeutic option for the treatment of diabetes and its complications by using antioxidants or nutrients with high antioxidant capacity. Antioxidants have been shown to reduce indices of oxidative stress measures in experimental disease models and – most importantly – also in humans. So far their possible therapeutic role has been underestimated in diabetes research. Since they are safe and inexpensive in comparison to hypoglycaemic therapy (such as intensified insulin therapy), they may offer a potential treatment regime applicable worldwide.

Until now, only long-term intensified insulin treatment and long-term oral hypoglycaemic therapy have been shown to be efficacious in the treatment of diabetes and its complications. This type of diabetes therapy is not applicable to the majority of diabetic patients because they fail to reach near normal glycaemic levels for several years, which is necessary to retard the onset and progression of diabetic late complications. Furthermore, some forms of hypoglycaemic therapy are very expensive. Intensified insulin treatment costs US$1600 per year per patient and is therefore not affordable for all citizens, especially in the developing world. Therefore, further innovations in diabetes therapy are needed. Consequently the organisers of this UNESCO conference with the title ‘The Role of Oxidative Stress in the Onset and Progression of Diabetes and its Complications’ hoped to stimulate both scientific and political discussion about future alternatives in the therapy of diabetes and its complications. Epidemiological studies suggest a prevalence of 300 million diabetic patients in the year 2010. This number of potential patients should lead to intensified government-sponsored research activities in the quest for effective methods of diabetes management.

Oxidative stress and oxidation damage in diabetes and its complications

The biomedical literature claims that ‘free radicals’ and other ‘reactive species’ are involved in many human diseases and that the increased formation of ‘free radicals’ accompanies tissue injury1–7. There is emerging evidence that free radicals make a significant contribution to the progression of diabetes and its complications8–13. The term ‘oxidative stress’ is widely used in the literature, but is rarely defined. In essence, it refers to the situation of a serious imbalance between the production of free radicals and antioxidant defence, leading to potential tissue damage14. The term ‘free radical species’ summarises a variety of highly reactive molecules that can be divided into different categories, e.g. reactive oxygen intermediates (ROI), reactive nitrogen species (RNS) and reactive chlorine species (RCS). The most prominent members of such categories include superoxide O2·−, hydroxyl radical OH·, peroxy radical ROO· in the ROS group, and nitric oxide NO· in the RNS group, and are summarised in Figure 114.

Figure 1.

Overview of different types of free reactive species

Free radical reactions are essential for host defence mechanisms as with neutrophils, macrophages and other cells of the immune system, however if free radicals are overproduced they cause tissue injury and cell death2, 3, 7. In order to avoid free radical overproduction, antioxidants are present in tissues to neutralise these free radicals3, 5–7. What is an antioxidant? B. Halliwell defined an antioxidant as ‘any substance that, when present at low concentrations compared to those of an oxidisable substrate, significantly delays or inhibits oxidation of this substrate’14. L. Packer favours the definition of the antioxidant as a metabolic intermediate, i.e. ‘an antioxidant is a substrate which protects biological tissues from free radical damage, which is able to be recycled or regenerated by biological reductants’15.

A manifold of compounds [flavonoids, uric acid, bilirubin, albumin, vitamin E (RRR-α-tocopherol), vitamin C (ascorbate), α-lipoic acid (thioctic acid) and glutathione] and various enzymes (catalase, superoxide dismutase, glutathione peroxidase) have been described as antioxidants, however only a few of them have been studied in detail. Therefore this review focuses on the most prominent of the antioxidants such as vitamins E and C, α-lipoic acid and glutathione which comprise an antioxidant network which enables these antioxidants to be recycled and regenerated in a co-ordinated way20–22. This does not exclude the fact that other antioxidative compounds are also active or even more effective and that improved, more specific compounds have to be developed. It is the intention of this review to present evidence that oxidative stress plays an important pathophysiological role in the onset of diabetes and in the development of diabetic complications and that antioxidants might principally be helpful in the treatment of diabetic patients. The antioxidant network is usually activated by RRR-α-tocopherol. After RRR-α-tocopherol is oxidised by free radicals, the RRR-α-tocopherol free radical is formed, which in turn reacts with ascorbate to non-enzymatically regenerate RRR-α-tocopherol20. Glutathione, together with the assistance of specific enzymes, can reduce the ascorbate radical20, 21. Recent research has shown that α-lipoic acid can play a unique role in this antioxidant network. α-Lipoic acid has a redox potential of −320 mV, which is even lower than that of the glutathione system (−280 mV)22. Thus, in its reduced form lipoic acid can regenerate glutathione22. Furthermore, α-lipoic acid has been shown to induce the synthesis of the endogenous antioxidant, glutathione, by reducing the glutathione precursor molecule cysteine to cystine, which is also important for the antioxidant network function23. An example of how the antioxidant network works with respect to the antioxidants RRR-α-tocopherol, ascorbate, glutathione and α-lipoic acid is shown in Figure 2. Besides antioxidant molecules, there are antioxidant defence enzymes (such as CuZnSOD, MnSOD) and glutathione peroxidase24. Depletion of antioxidants and antioxidant enzymes are important factors in the delicate pro-oxidant and antioxidant balance, which is essential for the function of the cell25. An imbalance between free radicals and antioxidant status is thought to lead to tissue damage and disease.

Figure 2.

Proposed mechanism of antioxidant regeneration by a co-ordinated intrinsic recycling process. RRR-α-Tocopherol itself will be converted to a tocopheroxyl radical when it scavenges a free radical species. The tocopheroxyl radical is reduced back to its active antioxidant form as RRR-α-tocopherol by a number of compounds including ubiquinol, cytochrome c and ascorbate. Ascorbate can be regenerated through reaction of thiols such as glutathione or lipoic acid. These can be returned to their reduced ‘active’ form by various mechanisms, drawing on the reducing power of NADH or NADPH

To assess the importance of oxidative damage in human disease, accurate methods for measuring ‘oxidative stress’ are essential. One has to establish markers of oxidative damage in experimental disease models and then test them in human disease conditions.

Which markers are available?

That several ROI are able to damage DNA has been well demonstrated in vitro and in humans, and several DNAdamage products have been identified in human urine including 8-hydroxydeoxyguanosin (8-OHdG), 8-hydroxyadenine and 7-methyl-8-hydroxyguanine, but the most exploited is 8-OHdG26–29.

Lipid peroxidation is important in vivo for several reasons, in particular because it strongly contributes to the development of atherosclerosis30–32. Many assays are available to measure lipid peroxidation, such as malonic dialdehyde (MDA) by the thiobarbituric acid (TBA) test and diene conjugation33–39. Recently, ferrous oxidation with Xylenol Orange (FOX) assay coupled with triphenylphosphine has shown to be a reliable marker for determining levels of hydroperoxides (ROOH) in health and disease33. Another reliable marker of lipid peroxidation is 8-epi-prostaglandin F (8-epi-PGF), an oxidative stress marker derived from the free radical oxidation of phospholipids containing arachidonic acid40–44.

Damage to proteins may also be important in vivo because it affects the function of receptors, enzymes, transport proteins, etc, and it contributes to secondary damage of other biomolecules, e.g. by inactivating antioxidant defence enzymes or repair enzymes45. The attack of various RNS (ONOO·−, NO·, NO2Cl) upon tyrosine leads to production of 3-nitrotyrosine, which can be measured immunologically or by HPLC46, 47. Reactive chlorine species can produce chlorinated products, e.g. 3-chlorotyrosine, and these have been detected in human atherosclerotic lesions31. Several assays to measure damage to specific amino acid residues in proteins by ROI/RNS/RCS have been developed. They include valine-hydroxides, tryptophan hydroxylation, and the ring-opener products 8-oxo-histidine, dithyrosine, and ortho- and meta-thyrosines48. A ‘general’ assay for oxidative protein damage is the carbonyl assay, which is based on the ability of several ROI to attack amino acid residues in proteins to produce carbonyl functions (Figure 3)48–50.

Figure 3.

Overview of biomarkers of free radicals, which have been applied to study free radical damage in human disease

There seems to be general agreement that the production of free radicals is increased in diabetic patients. Several clinical studies show increases in levels of oxidative stress markers, for instance ROOH, 8-epi-PGF, 8-OHdG and oxLDL in type 1 and type 2 diabetes when compared to healthy age-matched subjects51–64. Dandona showed an approximately four-fold higher median concentration of 8-OHdG in mononuclear cells of diabetic patients compared to corresponding controls. This difference was statistically significant and demonstrated for the first time greater oxidative damage to DNA in diabetic patients when compared to controls51.

Clinical studies by Halliwell and Nourooz-Zadeh demonstrated an approximately two-fold increase in the plasma oxidative stress measures 8-epi PGF and ROOH in diabetic patients when compared to healthy subjects61–63. The relationship of hydroperoxides (as an index of oxidative stress) and the level of the antioxidant RRR-α-tocopherol (standardised to cholesterol) was assessed. A 3–6-fold imbalance between increased levels of oxidative stress and the depletion of RRR-α-tocopherol was observed in the plasma of diabetic patients (Table 1)63. In addition, Davi et al. reported very recently that the increased urinary release of 8-epi-PGF observed in type 1 and type 2 diabetic patients can be normalised by improvement of the metabolic control64.

Table 1. Parameters of oxidative stress in healthy and type 2 diabetic subjectsa
VariablesHealthyType 2 diabetesp
  • a

    There is a more than two-fold increase in the lipid peroxidation product ROOH, a 35% depletion of cholesterol-standardised α-tocopherol and a 3.5-fold imbalance between elevated levels of oxidative stress and antioxidant depletion (ROOH/α-tocopherol/cholesterol) in type 2 diabetes patients compared to corresponding healthy controls.

  • b

    Measured by FOX assay.

  • c

    Range given in parentheses.

  • Data are mean±SD.

Total cholesterol (mmol/l)5.0±1.16.0±1.3<0.002
Fasting glucose (mmol/l)4.9±0.412.1±5.1>0.0005
HBA1c (%)11.0±2.4
ROOHb (µmol/l)4.1±2.29.4±3.3<0.0005
α-Tocopherol23.8±8.3 10.647.0c19.6±7.5 (8.6−44.3)c<0.05
α-Tocopherol (µmol/l) /cholesterol (mmol/l)5.1±2.3 (1.9−13.0)c3.3±1.0 (1.5−6.2)c<0.0005
ROOH/(α-tocopherol/cholesterol)0.9±0.6 (0.1−2.7)c3.2±1.6 (0.7−8.3)c<0.0005

One major hypothesis is that oxidised lipoprotein (oxLDL) contributes to the cardiovascular complications of diabetes. Several studies have demonstrated increased LDL oxidation in diabetic patients when compared to their corresponding controls60, 65–73. The concept of imbalance between oxidative stress measures and antioxidant depletion has been applied to the pro-atherogenic LDL oxidation process in diabetic patients. Leonhardt reported an imbalance between elevated oxLDL levels anddecreased RRR-α-tocopherol levels (cholesterol-standardised) in the LDL particles of diabetic patients when compared to healthy controls74, 75. The increase in LDL modification by oxidation may therefore not only correlate with increased levels of ROS but also with the diminished antioxidant protection.

A decrease in antioxidant capacity has been observed in the plasma of diabetic patients by several research groups54, 63, 75–85. In a prospective study, Salonen et al. found a decline of plasma RRR-α-tocopherol levels after the onset of type 2 diabetes86. Thornally et al. reported a decrease in the endogenous antioxidant, glutathione, in the erythrocytes of diabetic patients87. The decline in the plasma levels of the antioxidants RRR-α-tocopherol or glutathione may be explained as a result of the increased production of free radicals in diabetic patients. However, other explanations also need to be taken into considerations. Rösen et al. presented data from two nutritional studies showing a dietary-induced RRR-α-tocopherol deficiency in older subjects and especially in diabetic patients88, 89. Dietary recommendations are for diabetic patients aiming to reduce their fat intake in order to decrease weight. The lipophilic antioxidant RRR-α-tocopherol is present in the fat-soluble phase of food ingredients and may therefore be under-represented in a fat-reduced diabetic diet. In addition, the RRR-α-tocopherol content of dietary recommendations of diabetic patients was analysed in the food records of 100 diabetic patients and those values compared to the recommended dietary allowance (RDA) of RRR-α-tocopherol. In about 43% of the eating plans and dietary recommendations the RRR-α-tocopherol intake was lower than the RDA would have achieved. A correlation was found between energy intake and RRR-α-tocopherol levels. Furthermore, there was a negative relationship in the energy content of the diabetic diet and the dietary RRR-α-tocopherol intake, resulting in a potentially severe α-tocopherol deficiency in those diabetic patients who were obese and who received a slimming diet. Long-term dietary restriction of RRR-α-tocopherol leads to reduced RRR-α-tocopherol plasma levels in humans88, 89. The contribution of impaired nutritional antioxidant intake to the observed imbalance between increased plasma oxidative stress measures and decreased antioxidant plasma levels in diabetic patients has to be further evaluated.

The effects of antioxidant supplementation on oxidative stress measures is an indirect way to further prove the pathogenic concept of elevated free radicals and antioxidant depletion leading to diabetic complications. It has been shown in several studies that whole plasma and isolated LDL from type 2 diabetes patients are more prone to peroxidation when compared to the corresponding controls, and that supplementation with RRR-α-tocopherol decreases these peroxidation processes in diabetic patients68, 90. Supplementation with α-lipoic acid or α-tocopherol decreases the oxidative stress measured as ROOH (Table 2)4, 91, LDL-oxidation90, 92 as well as the imbalance between ROOH and cholesterol-standardised RRR-α-tocopherol levels91. Despite the arguments that can be raised about the validity of some individual biomarkers, the sum of evidence from biomarkers reporting oxidative damage to DNA, lipids and proteins supports the concept of increased oxidative stress in diabetes. Nevertheless there are some unanswered questions:

Table 2. Description of the patients (n=107) allocated to two groups according to the use of α-lipoic acid
 Without α-lipoic acidWith α-lipoic acid
  1. a

    The antioxidant α-lipoic acid (600 mg) reduces the oxidative stress measure ROOH to levels of healthy individuals. These studies suggest that elevated levels of oxidative stress in diabetic patients can be reduced by the antioxidant, α-lipoic acid. Nominal data are presented as number of patients. Continuous data are given as mean±SD. The two samples two-tailed Student's t-test was used to compare the continuous data of the two groups. Significant differences between the groups are expressed as *p<0.05 and ***p<0.005.

Patients (n)7433
Type 14519
Type 22914
Age (years)48.77±13.4954.09±12.56
Diabetes duration (years)15.31±10.4521.79±11.17*
HbA1c (%) 8.12±1.75 8.86±1.83
ROOH (µmol/l) 7.16±3.22 4.76±2.49***
  • When does oxidative stress as a pathogenic event occur in the disease process? Is it in the beginning of the disease or in the later disease stages, when metabolic and vascular abnormalities of hyperglycaemia have already occurred?

There is much evidence from experimental studies that the formation of ROI is a direct consequence of hyperglycaemia. Incubation of endothelial and smooth muscle cells with increasing concentrations of glucose initiates the formation of ROI93–99. A significant increase was already observed at glucose concentrations as high as 10 mM93. In addition, advanced glycation end products (AGE) have been shown to stimulate the formation of ROI by a receptor-mediated process100–103. Very recently Giardino et al.96 have shown that the intracellular formation of AGE and lipid peroxidation are closely dependent processes. Inhibition of lipid peroxidation also prevented the formation AGE products. That the autoxidation of glucose leads to the formation of ROI has already been shown by Wolf et al.104, 105. Furthermore, the regeneration of glutathione is delayed in the presence of high glucose causing an impairment of the antioxidant defence106. That hyperglycaemia and oxidative stress are linked very closely is also supported by in vivo evidence107. Thus, there are various mechanisms known by which ROI may be generated in diabetes, and much experimental evidence has been accumulated to show that various types of vascular cells are able to produce ROI under hyperglycaemic conditions. Observations that oxidative stress impairs endothelium-dependent vasodilatation8–12, 98, 99, 108, triggers the expansion of mesangial cells109 and is related to the incidence of type 2 diabetes86 suggest that oxidative stress may play an important role in the initiation of the pathophysiological cascade of events leading to vascular and other diabetic complications110.

Conversely, Baynes111 presented evidence that oxidative stress may not occur early in the disease process of diabetes, but may rather be an underlying pathogenic factor in the progression of the disease. One of the major problems in assessing the questions concerning when oxidative damage occurs in the disease process and whether there is an accumulation of free radical-derived tissue damage during the duration of the disease is the stability of the oxidation products111. In the case of DNA, it is well known that DNA alterations can be repaired by a sophisticated repair process utilising repair enzymes that act by excision and replacement of the modified base or nucleotide111. For proteins, lipids and RNA, the kinetics of the turnover of the affected molecules appear to be the critical factor limiting the accumulation of oxygen radical damage. Long-lived unrepaired protein molecules, such as collagen, products of oxygen radical reactions may accumulate with time111–114 and therefore act as a unique sensor for exposure to oxidative stress over time. Studies on glycation of proteins and Maillard reactions of glycated proteins have yielded indirect evidence for increased oxidative modifications of collagen in diabetes109.

  • What kind of free radical damage occurs at what stage of the disease process of diabetes and its complications? This question is of great importance for the evaluation and design of antioxidant therapy in diabetes and should be addressed in the near future. It might be that the form of oxidative stress and its localisation is dependent on the disease process itself.

  • Which compartment is mainly affected by oxidative stress? Which tissues and cells are the main sources for the generation of ROI in vivo? Is it possible to extrapolate the experimental findings on the disease process occurring in diabetic patients?

Diabetic polyneuropathy

Diabetic polyneuropathy is one of the leading complications of diabetes mellitus. Up to 50% of diabetic patients develop diabetic polyneuropathy115, 116. The DCCT Study and the UKPDS Study have shown that intensified insulin treatment can reduce the neurological deficits of diabetic polyneuropathy117–119. These results provide clinical proof that hyperglycaemia per se is the pathogenic factor for the onset and progression of diabetic neuropathy. Patients with intensified insulin treatment for 4 years (HbA1c levels of 7%) showed a clinically relevant and statistically significant effect on nerve function when compared to the conventionally treated group with HbA1c levels of approximately 9%117–119.

In recent years, several experimental studies suggest a multifactorial pathogenesis of diabetic polyneuropathy with a complex interaction between metabolic and vascular abnormalities. The ‘sorbitol pathway’ by which glucose is converted to sorbitol and fructose under hyperglycaemic conditions has been shown to lead to reduction of myoinositol, which in turn may reduce the activity of Na+,K+-ATPase, an enzyme known to be important for nerve conduction velocity120–123. Alternatively or in addition, free radicals may impair the endothelium-dependent vasodilatation either by changes in the generation and bioactivity of nitric oxide or by a reduced synthesis of vasodilating prostaglandins124. The diminished generation of vasodilating mediators may lead to a reduced endoneurial blood flow, which in turn causes endoneurial hypoxia and/or ischemia, which are responsible for the destruction of neuronal and Schwann cells and finally nerve degeneration (Figure 4)124–130. The hypothesis that hypoxia may play a major pathogenic role in diabetic neuropathy is supported by experimental and clinical observations131, 132. In animal models of diabetic neuropathy, which show similar neurophysiological and morphometric alterations to those observed in humans, free radical scavengers such as RRR-α-tocopherol and probucol can protect against neurovascular dysfunction133, 134. The antioxidant α-lipoic-acid has even shown a ten-fold stronger effect than RRR-α-tocopherol in improving endoneurial microcirculation and nerve conduction velocity135 and in improving the impaired endoneurial microcirculation in a dose-dependent fashion136. This neurovascular improvement is associated with reduction of the endoneurial oxidative stress markers, malondialdehyde and TBARS (thiobarbituric acid reactive substances), and a dose-dependent increase in the endoneurial glutathione content136, 137. These neurovascular effects of α-lipoic acid are associated with an increase in motor and sensory nerve conduction velocity136. This protective effect of α-lipoic acid is further strengthened by application of γ-linolenic acid (GLA)135.

Figure 4.

Hypothetical schema of the causal relationship between metabolic changes, oxidative stress and nerve degeneration in diabetes. The increased auto-oxidation of glucose and the increased formation of advanced glycation end products (AGE) produce reactive oxygen species (ROI). The elevated flux of glucose through the first half of the polyol pathway consumes NADPH. This impairs the glutathione redox cycle so that antioxidant protection is reduced. These changes lead to endothelial dysfunction which is a key element in the impairment of endoneurial blood flow and nerve dysfunction. All, Angiotensin II; EFA, essential fatty acid; ET, endothelin-1; PGI2, prostacyclin

Another important aspect in the progression of diabetic polyneuropathy is the impaired ability of damaged nerve fibres to regenerate138. Indeed, the levels of several neurotrophic factors, including nerve growth factor (NGF), ciliary neurotrophic factor, brain-derived neurotrophic factor and insulin growth factors which are assumed to be essential for nerve regeneration, have been shown to be reduced in diabetic polyneuropathy139–141. Although the mechanisms and biochemical effects of these neurotrophic factors are not yet understood in detail, it is interesting to note that oxidative stress and impaired microcirculation contribute to the diminution of neurotrophins in the process of neurodegeneration in diabetic polyneuropathy. In line with this hypothesis, administration of the antioxidant α-lipoic acid, which has been shown to reduce endoneurial oxidative stress markers and to increase nerve blood flow135, 136, caused a small increase in the sciatic NGF content and normalised both stimulus-invoked release of substance P and neuropeptide Y in experimental diabetic polyneuropathy142.

In clinical studies, the antioxidant α-lipoic acid has been approved for treatment of diabetic polyneuropathy in several countries. So far, six pilot studies have been completed143–148. Four out of six studies showed beneficial effects on nerve function, such as distal motor latency, heart rate variation and on symptoms of diabetic polyneuropathy143–148. Five multicenter, randomised, double-blind studies using α-lipoic acid in diabetic and cardiac autonomic neuropathy have been conducted (Table 3)149–153. In the Alpha-Lipoic Acid in Diabetic Neuropathy Study (ALADIN II) patients with peripheral polyneuropathy were randomly assigned using two doses of 600 mg, 1200 mg or placebo for 24 months150. Both doses of α-lipoic acid were safe and resulted in improvement of sural and tibial nerve conduction velocity as well as improvement of the distal motor latency and the sural amplitude when compared to placebo. In the ALADIN Study, 328 patients with type 2 diabetes and symptomatic peripheral polyneuropathy were randomly assigned to treatment with intravenous infusion of α-lipoic acid using a daily dose of 1200, 600, or 100 mg or placebo over 3 weeks149. The major symptoms of diabetic polyneuropathy such as pain, burning, paraesthesia and numbness improved significantly in the 1200 and 600 mg α-lipoic acid groups when compared to placebo. These effects of intravenous α-lipoic acid treatment on neuropathic symptoms in diabetic patients could be confirmed in the ‘Orpil Study’ using a 3-week high-dose oral treatment protocol of 1800 mg α-lipoic acid per day in 24 individuals151. In the ALADIN III Study, a total of 509 patients were eligible for intention to treat analysis and received either 600 mg or placebo in a 3-week intravenous treatment interval followed by a 6-month oral maintenance therapy with 600 mg or placebo152. The neuropathy impairment score (NIS) which summarises the clinical relevant nerve function of the whole body in addition to the NIS subscores of the lower limb and great toe showed a clinical meaningful improvement of 1 point compared to placebo which was statistically significant. The 6-month oral maintenance therapy could stabilise these observed effects on the NIS and their subscores. In the DEKAN Study, patients with type 2 diabetes and cardiac autonomic neuropathy were treated with oral doses of 800 mg α-lipoic acid daily (n=39) or placebo (n=34) for 4 months153. Two out of four parameters of heart variability were significantly improved with α-lipoic acid, such as the root mean squared successive differences (RDSSM) and the low frequency band of the power spectrum.

Table 3. Multi-center, randomised, double-blind, placebo-controlled studies using α-lipoic acid in diabetic peripheral and cardiac autonomic neuropathy
 Patients (n)Dose (mg)DurationEffectsSafety
  1. DML, Diatal motora latency; NIS, neuropathy impairment score; NDS, neuropathy disability score; HRV, heart rate variation.

  2. +, Improvement; −, no change; good, no elevated side effect when compared to placebo.

ALADIN328100/600/1200/placebo3 weeksSymptoms +Good
    NDS + 
    HPAL + 
ALADIN II65600/1200/placebo2 years oralSural NCV +Good
    Tibial NCV + 
    SNAP + 
    DML − 
    NDS − 
ALADIN III508600 iv/1800 oral/Placebo3 weeks iv 6 months oralSymptoms +/− 
    HPAL + 
    NIS total +Good
    NIS lower 
    Limb + 
DEKAN73800/placebo4 months oralHRV +Good
    QTc − 
ORPIL241800/placebo3 weeks oralSymptoms +Good
    HPAL + 
    NDS + 

In summary, treatment with antioxidants such as α-lipoic acid is safe and appears to be effective against diabetic polyneuropathy, reducing the symptoms and improving selective parameters of autonomic and peripheral nerve function. These clinical effects, together with the experimental findings, support the concept that oxidative stress plays an important pathogenic role in the development of diabetic polyneuropathy. The synergistic effect of the prostanoid precursor molecule, GLA, and α-lipoic acid on microcirculation and nerve conduction velocity in experimental diabetic polyneuropathy suggests a combined role of oxidative stress and impaired prostaglandin metabolism in the pathogenesis of diabetic polyneuropathy135, 142.

Vascular complications

Microvascular complications are one of the most serious aspects of diabetes that determine the life quality and expectancy of diabetic patients. Diabetic retinopathy and nephropathy are typical complications, but it has become clear in recent years that microvascular complications can also affect the heart and contribute to the development of diabetic neuropathy by limiting the endoneuronal blood flow. Our knowledge about the mechanisms leading to microvascular complications in diabetes is still limited. The following discussion focuses on the specific contribution of ROI to the development of vascular complications. It is not intended here to give a complete overview of the various mechanisms that may play a role [for detailed reviews see Refs11, 12, 97, 154.

Generation of ROI and activation of PKC

It has already been mentioned that both AGE and high glucose are able to induce the generation of ROI in various types of vascular cells93–95, 101–103. That in vivo hyperglycaemia and oxidative stress are closely linked is supported by determination of the total radical trapping antioxidant parameter (TRAP) in control subjects and diabetic patients107. However the underlying mechanisms are not yet known in detail. In addition to the autoxidation of glucose104, 105 three different mechanisms are discussed here.

  • Activation of a membrane-bound, macrophage-like NADH oxidase. Activation of NADH oxidase in endothelial cells and smooth muscle cells has been reported in hypertension and hypercholesterolaemia. Angiotensin II (AII) is a strong activator of this enzyme in vascular cells155–157.

  • Alternatively it has been suggested that the electron flux in endothelial NO synthase (NOS III) becomes uncoupled in hyperglycaemia. In such an uncoupled state the electrons flowing from the reductase domain to the oxygenase domain in the NOS complex are diverted to molecular oxygen rather than to L-arginine. In line with this assumption not only the production of ROI was prevented in human and rat endothelial cells in the presence of inhibitors of NOS, but also the activation of NFκB and the induction of apoptosis93, 157, 158.

  • Very recently Nishigawa et al.159 demonstrated in bovine aortic endothelial cells that the mitochondrial electron flux becomes uncoupled from ATP synthesis in hyperglycaemic conditions. The production of ROI could be prevented by various uncouplers of the mitochondrial electron chain and the overexpression of the uncoupling protein (UCP-2). Furthermore, activation of PKC, the polyol pathway, the transcription factor NFκB and the increased formation of AGE and glucosamine were clearly dependent on the formation of ROI, suggesting that at least in these cultured endothelial cells the formation of ROI is the central and initiating step for the transformation of endothelial cells into an active, pro-thrombotic state. According to these observations, an accelerated substrate flow from either glucose or fatty acids seems to be final cause for the generation of ROI and the oxidative stress159, 160.

These hypotheses could explain why an accelerated conversion of glucose to fructose by the ‘sorbitol pathway’ has a large impact on the redox state of a cell. The conversion of glucose by this pathway consumes NADPH and leads to an increased NADH flow to mitochondria161. In addition, an important co-factor for glutathione peroxidase is diminished, and the regeneration of glutathione is impaired, which may limit the antioxidative capacity of the cells101, 106 and contribute to the occurrence of oxidative stress in diabetes.

That hyperglycaemia leads to an activation of protein kinase C in all cells and tissues, which take up glucose independently from insulin, might therefore not only be mediated of an increase in diacylglycerol (DAG) which is a strong activator of protein kinase C (PKC)162, 163, but also by the enhanced formation of ROI. It is therefore not surprising that diabetes is associated with the activation of various isoforms of PKC154, 164–178. Amongst the 11 PKC isoforms found in vascular cells, PKCβ and α appear to be activated in the aorta and heart of diabetic rats as shown by immunoblotting studies166, 168. Increases in the other isoforms such as PKCα, β1 and ε in the retina165, 167, 175–178 and PKCα, β1 and δ in the glomerular cells from diabetic rats have been described154, 164, 169–171. Various experimental observations and the use of the selective PKC inhibitors such as LY 333531 and RRR-α-tocopherol174 suggest that the PKCβ and PKCα isoforms are predominantly activated in vascular tissue and may be responsible for many of the vascular dysfunctions that occur in diabetes154, 164, 166, 168, 172, 173.

Thus, several lines of evidence suggest AGE and ROI are able to activate PKC and that activation of PKC might be a common downstream mechanism to which multiple cellular and functional abnormalities in the diabetic vascular tissue can be attributed, including changes in vascular blood flow, vascular permeability, extracellular matrix components and cell growth. Certainly ROI play a prominent role in the activation biochemical pathways involved in the pathogenesis of vascular complications. Taken together, we have to state that activation of PKC and the generation of ROI in diabetes present important pathogenic mechanisms, but it is not completely clear which roles the activation of these pathways play in thevarious signal transduction pathways altered in thediabetic vasculature178–181. Taking these observations into consideration, it becomes evident that hyperglycaemia itself is a causative factor in the development of endothelial and vascular dysfunction. Not onlylong-term, but also short-term (postprandial), hyperglycaemia has to be considered as a risk factor182–185.

Activation of redox-sensitive transcription factors by AGE and hyperglycaemia

AGE accumulate at an accelerating rate during the course of diabetes185–188. They are believed to induce endothelial dysfunction characterised by an impaired regulation of blood flow by a highly thrombogenic and pro-coagulant state of vessel wall191–195. These vascular alterations are thought to be of relevance in the onset and progression of diabetic complications188–190. AGE formation has so far mostly been discussed as a process of protein modification111. From recent studies it follows, however, that interactions of AGE-modified proteins with specific AGE receptors do not only serve to degrade AGE proteins, but induce signal transaction pathways102, 103, 189–196 which lead to an impairment of endothelium-dependent vasodilatation, an increased pro-coagulant activity, e.g. tissue factor expression, impairment of anti-coagulant activity, induction of cell adhesion molecules, increased vasoconstriction by inducing endothelin I and other changes in the function of the vessel wall.

As consequence of the activation of one of these signal transduction pathways – binding of AGE to RAGE (receptor for AGE) – vascular cells (smooth cells, macrophages, endothelial cells) generate reactive oxygen species [102,103,196 which results in depletion of cellular antioxidant defence mechanisms (e.g. glutathione, ascorbate) and the activation of redox-sensitive transcription factors such as NFκB186, 189, 195. The activation of NFκB and presumably also other redox-sensitive transcription factors such as AP-1 promotes the expression of a variety of NFκB- and AP-1 regulated genes, such as the pro-coagulant tissue factor, endothelin I, or the adhesion molecule, VCAM-1, all of which have been found to be increased in the diabetic state102, 189, 197. The concept of an AGE-induced oxidative stress which activates the transcription factor NFκB and presumably AP-1 can explain the concomitant occurrence of oxidative stress and changes in the dynamic endothelial balance from an anticoagulant to a pro-coagulant state, from vasodilatation to vasoconstriction and impaired microcirculation (Figure 5)102, 103, 186–197.

Figure 5.

Possible linkage between hyperglycaemia-induced oxidative stress, activation of NFκB, and the development of vascular complications in patients with diabetes mellitus

The AGE-induced activation of NFκB has been studied in vitro in endothelial cells, mesangial cells, neurons and smooth muscle cells. In addition, animal studies demonstrated significantly higher renal expression of NFκB indiabetic animals compared to healthy controls102, 103, 186–197. Nawroth et al. reported increased activation of NFκB in monocytes of type 1 diabetic patients when compared to healthy controls198, 199. A positive correlation (r=0.7) between metabolic control and NFκB was seen187. Intensified insulin treatment, which is known to prevent the onset of diabetic complications, was associated with NFκB inactivation in diabetic patients (Nawroth, unpublished data). The effect of different degrees of endothelial dysfunction and different stages of nephropathy on NFκB activation in peripheral blood mononuclear cells has been evaluated in type 1 diabetic patients. A positive correlation between the endothelial dysfunction marker, thrombomodulin, and NFκB activation has been observed. A similar correlation was found between urinary albuminuria and NFκB activation198. The correlation of thrombomodulin plasma levels, urinary albumin and NFκB binding activity suggests a link between NFκB activation, endothelial dysfunction, as well as diabetic nephropathy.

It is interesting to note that in experiments using vascular cells the generation of ROI and the activation of NFκB is not only caused by AGE, but also by high concentrations of glucose, which might indicate a link between elevated postprandial glucose levels (‘glucose spikes’) and the development of vascular complications93. Taken together, these experimental and clinical data suggest that activation of the oxidative stress-sensitive transcription factors such as NFκB and AP-1 represents an important pathogenic link between hyperglycaemia and the transformation of the vessel wall in diabetes (Figure 5).

In line with this concept, activation of NFκB can be inhibited in experimental studies by antioxidants such as RRR-α-tocopherol, acetylcysteine and α-lipoic acid93, 108, 195, 198–200.

In a clinical trial a decrease in NFκB binding activity in mononuclear blood cells was observed in diabetic patients receiving 600 mg α-lipoic acid per day198, 199. This effect was independent of the level of glycaemic control, the degree of endothelial dysfunction as well as different stages of diabetic nephropathy. In an 18-month pilot study the effect of 600 mg α-lipoic acid supplementation on the progression of diabetic nephropathy was studied201. Plasma thrombomodulin levels (as a marker for endothelial dysfunction) and urinary albuminuria (as a marker for nephropathy) were compared for a group treated with α-lipoic acid and a control group which received no treatment. α-Lipoic acid significantly reduced the time-dependent decrease of plasma thrombomodulin and urinary albuminuria seen in control patients. Treatment with α-lipoic acid was the only significant factor predicting the decrease of plasma thrombomodulin and urinary albuminuria in multiple regression analysis.

In summary, experimental and clinical results provide increasing evidence that redox-sensitive transcription factors are activated by AGE or hyperglycaemia that may play a central role in the development of hyperglycaemia-induced ‘late’ complications.

Vascular blood flow

Abnormalities in vascular blood flow and contractility have been found in many organs of diabetic animals and patients including kidney, retina, heart, peripheral arteries and microvessels of peripheral nerves. One of the first symptoms of changes in the regulation of blood flow seems to be an impairment of endothelium-dependent vasodilatation8–12, 108 which can already be observed in school children between 9 and 11 years of age202. There is considerable evidence that this defect precedes the onset of type 2 diabetes203.

The underlying mechanisms are very complex and may differ in the affected organs and tissues (retina, kidney, peripheral arteries and heart). This becomes very clear if the generation of nitric oxide is considered, one of the main mediators of endothelium-dependent relaxation204. An increase, a diminution or no effect of diabetes on NOS have been reported205. In the kidney, for example, the enhanced release of nitric oxide206–208 and vasodilating prostaglandins (PGE2, PGI2)209, 210 corresponds to the accelerated renal blood flow which is a characteristic finding of type 1 diabetes211, 212 and in experimental diabetes213. Diabetic glomerular hyperfiltration is assumed to be the result of hyperglycaemia-induced decreases in arteriolar resistance leading to increases in glomerular filtration pressure. In the isolated perfused heart of diabetic rats we observed an impaired endothelium-dependent relaxation, but an elevated activity of NOS108, 214. Thus, the balance in the synthesis and release of vasodilating and constricting factors has to be regarded which may be influenced differently by diabetes in the various organs.

In which ways may ROI affect the regulation of blood flow?

There is considerable evidence that nitric oxide reacts in a diffusion-controlled way with superoxide anions to peroxynitrite215. Thus, a simultaneous formation of ROI would consequently reduce the amount of biologically active nitric oxide and lead to an impairment of endothelium-dependent vasodilatation. That such a mechanism may be operative has been demonstrated in hypertension and hypercholesterolemia216, 217. We have suggested a similar mechanism to act in the heart, since the impaired endothelium-dependent relaxation was restored by the addition of superoxide dismutase (SOD), an enzyme which inactivates superoxide anions very effectively, and by pre-treatment of diabetic animals with α-tocopherol108. Furthermore, in endothelial cells exposed to high glucose the formation of peroxynitrite has directly been shown by immunoblotting218. Experimental studies using various types of vessels are in concordance with this hypothesis8–12, 108. In addition to superoxide anions, AGE have been shown by Bucala et al.219 to quench nitric oxide and may thereby important modulators of endothelium-dependent relaxation.

Recent clinical studies have shown that antioxidants are able to reverse or to improve the disturbed endothelium-dependent vasodilatation in diabetic patients, in patients with coronary artery disease, hypercholesterolemia or high remnant lipoprotein levels220–224.

Quenching of nitric oxide does, however, not only affect the vasotonus, but additionally the anti-thrombotic properties of endothelium, since nitric oxide inhibits the expression of adhesion molecules (VCAM-1, ICAM-1) and the proliferation of smooth muscle cells225. Thus, the anti-thrombotic and anti-atherosclerotic defence becomes defective by oxidative stress. Another mechanism by which a ROI-mediated activation of PKC may contribute to vasoconstriction is by increasing the expression of endothelin-1 (ET-1)226, since the expression of ET is dependent on the binding of nuclear proteins to the GATA motif and on activation of the redox-sensitive transcription factor, AP-1227. Furthermore, it is well known that hypoxia, which is associated with the increased formation of ROI, induces the expression of ET-1 and of the vascular endothelial growth factor (VEGF)228. That the expression of ET-1 is increased in the retina of diabetic rats and that the decrease in retinal blood flow is prevented by an endothelin receptor antagonist228 supports this hypothesis. ROI may therefore directly and indirectly contribute to reduction in retinal blood flow, to local hypoxia, and induce VEGF, causing increases in vascular permeability and microaneurysms229, 230. In a clinical study the effect of 1800 mg RRR-α-tocopherol on retinal blood flow and renal function has been evaluated in type 1 diabetes patients in a placebo-controlled, double-shielded trial231. After 6 months there was an increase in retinal blood flow and improvement of renal function in the RRR-α-tocopherol treated group that was statistically significant when compared to placebo. Additional clinical trials are needed to determine whether antioxidant treatment will be effective in the prevention of diabetic late complications.

Taken together, there is much experimental and clinical evidence that the formation of ROI has a major impact on the function and structure of the vessel wall in diabetes. These observations, however, do not exclude the possibility that other mechanisms may also be operative such as the activation of specific isoforms of PKC, the enhanced formation of the endothelium constricting factor or vasoconstricting prostaglandin peroxides. For a complete understanding of the situation we have to know the network of mediators determining the fine-tuning of blood flow and the influence of diabetes on this network.

Vascular permeability and proliferation of vascular cells

Increased vascular permeability and vascular cell proliferation are two important events in the pathogenic cascade of endothelial dysfunction leading to microvascular and macrovascular complications232–237. ROI and PKC activation regulate vascular permeability andvascular cell proliferation (neo-vascularisation)237–239. As in hypertension240, oxidative stress directly induced by hyperglycaemia or indirectly by ET-1 may lead to activation of the redox-sensitive transcription factors, AP-1 and NFκB. As a consequence the expression of various chemokines and cytokines such as vascular endothelial growth factor (VEGF), transforming growth factor-β (TGF-β), monocyte chemo-attracting protein (MCP-1) and others165, 230, 240–244 are up-regulated inducing cell proliferation and changes in permeability. VEGF has been shown to be increased by high glucose and in the ocular fluid from diabetic patients242–244. It has been implicated in the neovascular process of proliferative retinopathy231, 232. Antioxidants as well as inhibition of PKC activity by selective inhibitors decrease cell proliferation and permeability in diabetic retinopathy, in aortic smooth muscle and mesangial cells and supports the close relationship between oxidative stress and PKC activation165, 173, 230, 231, 242, 243.

Extracellular matrix components

Thickening of capillary basement membrane is one of the early structural abnormalities observed in tissues including the vascular system in diabetes244. Histologically, increases in type IV and VI collagen, fibronectin and laminin, in addition to decreases in proteoglycans, are not only observed in mesangium245–253 but also in the heart107 and other vascular beds of diabetic patients253. Experimental studies show that these alterations in the vascular bed can be prevented by high doses of antioxidants such as tocopherol and inhibition of PKC108, 154, 164.

Macrovascular complications: the role of modified lipoproteins, oxidative stress measures and antioxidant depletion in the onset and progression of cardiovascular disease

The imbalance between oxidative stress measures and antioxidant levels in diabetes is present because of the generation of ROI during glycation, glucose and lipid oxidation1, 51, 55, 61–63, 97, 104. These elevated levels of free radicals may be at least in part responsible for the observed depletion of antioxidants such as glutathione, RRR-α-tocopherol and ascorbate in the plasma of diabetic patients63, 74, 75, 87. In addition, there is sufficient evidence for an RRR-α-tocopherol deficiency in the diet of diabetic patients88, 89 which may account for the decrease in RRR-α-tocopherol plasma levels in diabetic patients when compared to controls. RRR-α-tocopherol isthe major antioxidant in the LDL particle, accounting for over 90% of LDL protection against lipid peroxidation (Figure 6)254. Leonhardt et al. demonstrated astatistically significant decrease in cholesterol-standardised RRR-α-tocopherol levels in the LDL fraction from diabetic patients compared to healthy controls74, 75. These preliminary clinical data suggest thatenhanced LDL oxidation in diabetic patients derives not only from glucose-induced free radical production butalso from insufficient antioxidant protection (RRR-α-tocopherol depletion) in LDL. Hyperglycaemia leads not only to increased susceptibility of LDL oxidation but also to increased non-enzymatic glycation of LDL32, 65, 255. These modified LDL particles have been found to be increased in the plasma of diabetic patients compared to healthy controls65–74, 254, 255. Elevated levels of autoantibodies against these modified LDL particles have also been detected in diabetic patients65, 66. Modified LDL exhibits biological properties that promote atherogenesis. Unlike other lipoproteins, oxidised or glycated LDL is also able to transform macrophages and perhaps also smooth muscle cells into foam cells in addition to modulating growth factor and cytokine expression, thus leading to the formation of atherogenic plaques254.

Figure 6.

Schematic overview of the proposed causal pathogenic cascade of hyperglycaemia, oxidative stress, LDL oxidation, arteriogenic plaque formation and myocardial infarction rate

That RRR-α-tocopherol reduces the pro-atherogenic risk factors as oxidised LDL, which are thought to be responsible for the onset and progression of coronary heart disease in diabetic patients, is demonstrated by several observations. Clinical studies have shown an inhibitory effect of RRR-α-tocopherol supplementation on the hyperglycaemia-induced LDL modifications in type 1 and 2 diabetic patients32, 68, 90, 254. Reaven evaluated the effect of 1600 IU of RRR-α-tocopherol supplementation daily for 10 weeks on the hyperglycaemia-induced LDL modifications in a placebo-controlled, randomised trial68. There was a reduction of plasma LDL oxidation of approximately 60% in diabetic patients which was statistically significant when compared to healthy controls. These results could be confirmed by other groups. Jialal et al.256 found that doses equal to or higher than 450 IU are sufficient to significantly ameliorate the susceptibility of LDL to oxidation, indicating that relatively high doses of RRR-α-tocopherol for supplementation are needed. Furthermore, Bellomo and others have shown that the effects of RRR-α-tocopherol supplementation on LDL oxidation are accompanied by a concomitant reduction in autoantibody levels against hyperglycaemia-induced LDL modifications65, 66. It is well established that hyperglycaemia-induced LDL modifications are antigens and induce autoantibodies against these epitopes. Salonen et al.257 could demonstrate that the titre of autoantibodies to MDA-LDL (as an epitope of oxLDL) in asymptomatic middle-aged men predicted the rate of carotid arteriosclerosis. Since this initial report there have been several studies demonstrating a relationship between such autoantibody titres and clinical manifestations of atherosclerosis65, 258, 259.

There is epidemiological and clinical evidence in non-diabetic individuals that RRR-α-tocopherol supplementation is associated with a reduction of coronary heart disease (Table 4). The evidence for other antioxidants is less convincing. The WHO sponsored ‘Monica Study’ found an inverse correlation between plasma RRR-α-tocopherol levels and coronary heart disease (CHD) mortality in different European populations260. The Physician Health Study investigating the beneficial effects of RRR-α-tocopherol supplementation in 80 000 women and 40 000 men found an approximate 40% reduction in deaths from CHD in those patients who ingested 200–400 IU α-tocopherol daily when compared to those who received less than 10 IU per day261, 262. Thehypothesis of cardioprotective effects of RRR-α-tocopherol is further supported by the results of the CHAOS study, showing a 77% reduction of non-fatal myocardial infarction rate in those receiving α-tocopherol compared to the placebo group263. The subgroup analysis of the approximately 170 diabetic patients also enrolled in this study showed also a marked 50% reduction in the non-fatal heart infarction rate which was not statistically significant (p=0.14) due to the small sample size of the diabetic patient group (n=264). The cardioprotective effects of RRR-α-tocopherol seen in epidemiological and clinical studies may relate to its inhibitory effects on LDL oxidation265. However, it should be pointed out very clearly that all the described observations are suggestive of, but do not prove, the clinical efficiency of antioxidants, especially since there are also reports in which the authors did not find evidence for cardiovascular protection by antioxidants266. Also the database is too small to answer various questions (recommended dosage, time scale, interactions between several antioxidants in the antioxidative network, which patients may profit from treatment, etc).

Table 4. Overview of important pharmacological, epidemiological and clinical studies showing the cardio-protective effects of RRR-α-tocopherol supplementation
Monica Study, Circulation 1994; 90: 583–612Subgroup analysis in 38 populations, 21 countriesReverse correlation of vitamin E plasma level and CHD mortality
Salonen et al., Br Med J 1995; 311: 1, 124–1, 127944 type 2 diabetes in Finnish men, 4-year follow-upLow plasma level of vitamin E increases risk of CHD four-fold
Kushi et al., N Engl J Med 1996; 234: 1, 156–1, 16234 486 post-menopausal women without history of CVD, 7-year follow-upInverse correlation of vitamin E intake with CHD mortality
Gey et al., Am J Clin Nutr 1991; 53: 3, 265–3, 34511 000 Europeans, 9-year follow-upVitamin E supplementation decreases ischemic heart disease mortality rates by ∼60%
Rimm et al., N Engl J Med 1993; 328: 1, 450–1, 466/Health professionals follow-up study51 529 menVitamin E supplementation decreases CHD by ∼40%
Stampfer et al., N Engl J Med 1993; 328: 1, 444–1, 449/Nurses Health Study Cohort121 700 women, 8-year follow-upVitamin E supplementation decreases CHD by ∼34%
CHAOS Study, Lancet 1996; 347: 781–7862002 patients with CHDVitamin E supplementation decreases non-fatal MI by ∼77%
Fuller et al., Am J Clin Nutr 1996; 63: 753–759Vitamin E supplementation in patients with diabetes mellitusVitamin E supplementation decreases oxidation of LDL
Reaven et al., Diabetes Care 1995; 18: 753–7591600 IU/day over 10 weeksVitamin E supplementation decreases oxidation of LDL
Ceriello et al., Diabetes Care 1991; 14: 68–72600/1200 IU/day over 2 monthsDecreases of protein glycosylation
Jialal et al., Arterioscler Thromb Vasc Biol 1995; 15: 190–198Dose–response study in ex vivo system in control subjectsDoses >450 IU decreases oxidation of LDL

In summary, there is increasing evidence that hyperglycaemia-induced LDL modifications are one major factor in the pathogenesis of atherosclerosis. Elevated levels of ROI and insufficient antioxidant protection of the LDL particle lead to the enhanced LDL oxidation in diabetes. Supplementation with the antioxidant, RRR-α-tocopherol, which accounts for 90% protection of LDL against oxidation, has been shown to reduce LDL oxidation in diabetic patients. The idea of a cardioprotective effect of RRR-α-tocopherol supplementation in diabetic patients is further supported by epidemiological and clinical studies showing a decrease in the myocardial infarction rate with RRR-α-tocopherol supplementation in the non-diabetic patient population. More placebo-controlled clinical trials are needed to assess the cardioprotective effects of RRR-α-tocopherol supplementation in non-diabetic and diabetic patients.

The role of oxidative stress in metabolism

Type 1 diabetes is characterised by the endogenous loss of insulin production, whereas type 2 diabetes which afflicts 90% of all diabetic patients is characterised by the loss of the ability of insulin-sensitive tissue to respond to insulin. As a consequence the gluconeogenesis in liver is accelerated, whereas the uptake and conversion of glucose by insulin-sensitive tissues such as muscle and fat are severely impaired. To overcome these defects the release of insulin by the β-cells is increased, resulting in the typical features of type 2 diabetes, namely high glucose plasma levels in the presence of hyperinsulinaemia. Whereas no data are available at present about the effects of antioxidants on regulation of gluconeogenesis, several lines of evidence suggest that antioxidants influence insulin signalling and glucose uptake.

Recent research has demonstrated a direct link between the imbalance of oxidative stress and antioxidants leading to impaired glucose uptake. In 3T3-L1 adipocytes glucose uptake is rapidly decreased when they are incubated with glucose oxidase which results in steady production ofH2O2267. The reduction of insulin-dependent 2-deoxyglucose uptake was accompanied by decreased PI3 kinase activity and GLUT4 translocation. This suggests that reactive oxygen species and antioxidant depletion could impair the insulin-mediated PI3 kinase, which results in impaired GLUT4 translocation and defective insulin-mediated glucose uptake. Increased oxidative stress measures in addition to antioxidant depletion leading to decreased glucose uptake were also observed in L6 muscle cells267, 268. Depletion of antioxidants accompanied by decreased glucose uptake has also been detected in type 2 diabetes patients. Thus, Salonen et al. found in 911 Finnish men that a decrease in RRR-α- tocopherol levels is a strong predictor for the onset of type 2 diabetes86. The findings of this 4-year prospective study led to the hypothesis that the imbalance of reactive oxygen species and antioxidants is one important pathogenic factor leading to insulin resistance through impaired stimulation of the insulin signalling pathway. This hypothesis is further supported by additional clinical observations showing a close association between oxidative stress and insulin sensitivity269–272. Epidemiological studies demonstrated a close correlation between low ascorbate plasma levels and the risk of developing type 2 diabetes273. Several groups report on higher levels of oxidative stress measures in subjects who had an impaired glucose tolerance test274–276. Thus, there is increasing evidence that the imbalance between ROI and antioxidant capacity may play an important role in the development and progression of insulin resistance in type 2 diabetic patients.

If this hypothesis is true, then reversal of the imbalance between ROI and antioxidant capacity should improve insulin resistance. This has been demonstrated with the antioxidant α-lipoic acid. α-Lipoic acid was shown in several experimental in vitro and in vivo as well as in diabetic patients to improve glucose utilisation277–288. α-Lipoic acid also improves insulin resistance in cultured L6 muscle cells, 3T3-L1 adipocytes, in rodent animal models of diabetes as well as in diabetic patients276–283. Recently, Bashan and co-workers have found a reversal of the glucose oxidase-induced PI3 kinase-mediated defect in insulin-mediated 2-deoxyglucose uptake in 3T3-L1 adipocytes treated with α-lipoic acid267, 268. Klip and co-workers have demonstrated an activation of IRS-I and PI3 kinase in L6 muscle cells and 3T3-L1 adipocytes in the presence of α-lipoic acid, thus leading to a translocation of the GLUT4 transporter from the intracellular compartment to the cellular surface (Figure 7)279. These results establish the ability of α-lipoic acid to stimulate glucose uptake through activation of the insulin signalling pathway. These antidiabetic effects of α-lipoic acid on glucose uptake and glucose oxidation have been further studied in several models of diabetes, such as in the muscle epitrochelaris of the obese Zucker rat, the sand rat, in the heart of the streptozotocin diabetic rat as well as in isolated rat diaphragm and rat brain277–284. Administration of α-lipoic acid resulted in an increase in cellular glucose uptake in all diabetes models. A decline in blood glucose levels was determined in sand rats and the streptozotocin diabetic rats283–289. The effects of an α-lipoic acid-induced increase in glucose transport and in non-oxidative (glucose incorporation into glycogen) and oxidative (glucose oxidation) metabolism was investigated in the epitrochelaris muscle of the obese Zucker rat. α-Lipoic acid treatment improved insulin-mediated 2-deoxyglucose uptake by 62% and increased both insulin-stimulated glucose oxidation and glycogen synthesis by 33% and 38%, respectively278. These responses to α-lipoic acid treatment were associated with significantly lower (15–17%) plasma levels of insulin and free fatty acids278. Administration of α-lipoic acid further resulted in an increase in glucose uptake and glucose oxidation in the perfused hearts of streptozotocin diabetic rats282. These alterations of glucose metabolism led to increased ATP synthesis and oxygen consumption as a precondition for an improvement of myocardial function depressed in diabetes. In type 2 diabetes patients, acute and chronic treatment with α-lipoic acid in various doses ranging from 500 to1000 mg iv improved insulin-stimulated glucose disposal by 30–50% in type 2 diabetes patients284–287. A placebo-controlled, randomised, double-blind study found a statistically significant 27% improvement in glucose utilisation after oral administration of α-lipoic acid287. Konrad et al. demonstrated a decline in glucose, pyruvate and lactate levels during oral glucose tolerance tests in the presence of α-lipoic acid, supporting the experimental findings of an α-lipoic acid-mediated increase in glucose uptake and glucose oxidation (Table 5)288. α-Lipoic acid is not the only antioxidant that has been shown to improve glucose utilisation in diabetic patients. Antioxidants such as RRR-α-tocopherol, ascorbate and glutathione have also been found to increase glucose utilisation in diabetic patients289–296.

Figure 7.

Schematic diagram of the glucose transporter translocation hypothesis. Insulin-responsive tissue – specifically adipose tissue and skeletal muscle – contain intracellular stores of glucose transporter proteins (GLUT). The binding of insulin and the subsequent increase in tyrosine phosphorylation of the insulin receptor substrates (IRS 1-n), their binding to the enzyme phosphatidylinositol 3-kinase (PI3 kinase), and the resultant activation of PI3 kinase to produce phosphorylated phosphoinositides. This signalling cascade leads to the mobilisation and insertion of stored glucose transporters into the plasma membrane allowing for increased glucose flux into the cell in response to insulin. α-Lipoic acid utilises components of the insulin transduction cascade in its ability to increase glucose uptake

Table 5. Influence of α-lipoic acid (LA) on the insulin-sensitivity in type 2 diabetic patients
Jacob et al., Drug Res 1995; 45: 872–87413 type 2 diabetic patients; acute parenteral administration; 1000 mg/LA vs placebo. Method: glucose clampSignificant increase in MCR of glucose by ∼50% by acute administration of LA
Jacob et al., Exp Clin Endocrinol Diabetes 1996; 104: 284–28820 type 2 diabetic patients; 500 mg/LA iv over 10 days. Method: glucose clampSignificant increase in MCR of glucose by ∼30% by acute administration of LA
Rett et al., Diabetes und Stoffwechsel 1996; 5 (3): 59–6312 type 2 diabetic patients; acute parenteral administration; 600 mg/LA vs placebo, crossover. Method: glucose clampSignificant increases in MCR in 58.3% of patients vs placebo
Jacob et al., Free Radic Biol Med 1999; 27: 309–31474 type 2 diabetic patients; oral administration of 600/1, 200/1, 800 mg/LA or placebo over 4 weeks. Method: glucose clampNo dose effect, significant increase in MCR of ∼27% vs placebo
Konrad et al., Diabetes Care 1999; 22: 280–28720 type 2 diabetic patients; oral administration of 1200 mg/LA over 4 weeks. Method: intravenous glucose tolerance testIncrease in glucose effectiveness

In conclusion, experimental and clinical data suggest an inverse association between insulin sensitivity and levels of ROI. Whether antioxidants represent an attractive therapeutic strategy for the treatment of insulin resistance in type 2 diabetes patients has now to be investigated in large clinical trials.

Recent studies suggest that oxidative stress may also participate in the development of type 1 diabetes297, 298. Non-obese diabetic (NOD) mice are known to develop autoimmune diabetes following the infiltration of inflammatory cells into pancreatic islets299, 300. Infiltrating cells are composed of T and B lymphocytes, macrophages and natural killer cells, which may exhibit their β-cell cytotoxicity through ROI297, 298. Studies have found an increase in ROI such as superoxide radical and hydrogen peroxide in β-cells infiltrated by these inflammatory cells297, 298. Pancreatic β-cells are especially vulnerable to oxidative stress, probably on account of their low levels of ROI scavenging enzymes such as superoxide dismutase (SOD), catalase and glutathione (GSH) peroxidase300–307. Antioxidants such as RRR-α-tocopherol and nicotinamide have been shown to have protective effects against diabetes in NOD mice308–310. In order to prove the concept of oxidative stress-mediated pancreatic β-cell destruction Miyazaki and co-workers have generated a NOD transgenic mouse overexpressing thioredoxin, a small (12 kDa) protein with antioxidant function exclusively in pancreatic β-cells. Spontaneous diabetes was prevented or delayed in the NOD transgenic mice310. The results support the hypothesis that ROI play an essential role in the development of type 1 diabetes and that antioxidant treatment (such as thioredoxin overexpression) has a protective effect against the development of type 1 diabetes. Large-scale clinical trials have been started to investigate the protective role of the antioxidant, nicotinamide, in the onset and development of type 1 diabetes. These clinical trials will prove or disprove the hypothesis of oxidative stress-induced pancreatic β-cell destruction in type 1 diabetes.


There is emerging evidence that the generation of ROI is one major factor in the onset and the development of diabetes and its complications. Research over the last 10 years suggests that oxidative stress plays a role in this pathway as a link between hyperglycaemia and the typical observed pathophysiological features of diabetes, associated with the onset and progression of diabetic late complications. Recently, research has also revealed an oxidative stress-induced inactivation of the signal pathway between the insulin receptor and the glucose transporter system that is known to be defective in the skeletal muscle of type 2 diabetes patients. Therefore, oxidative stress may also be an important underlying factor leading to the onset and progression of insulin resistance in type 2 diabetes patients.

Antioxidants might therefore be helpful for treating diabetic patients and their complications. Among the variety of antioxidants such as ascorbate, glutathione, acetylcysteine and others, α-lipoic acid and RRR-α-tocopherol are the two substances which have been shown in several experimental and clinical studies to have beneficial effects on diabetes and its complications. RRR-α-Tocopherol reduces LDL oxidation and autoantibody titres against LDL modification proteins that are prominent and established risk factors for the manifestation of atherosclerosis. Clinical and epidemiological studies have shown an inverse relation between RRR-α-tocopherol and the myocardial infarction rate. RRR-α-Tocopherol was furthermore shown to exert beneficial effects on retinal flow of blood and renal function in diabetic patients. α-Lipoic acid has been shown in several studies to interfere with the development of diabetic polyneuropathy. Clinically, it reduces neuropathic symptoms and improves autonomic and peripheral functions in diabetic patients. α-Lipoic acid has been further shown to reduce the imbalance between oxidative stress measures and antioxidant capacity in diabetic patients. This antioxidant effect can explain the inactivation of the oxidative stress-sensitive transcriptional factor, NFκB, in plasma mononuclear cells, which in turn may be associated with the improvement of renal and endothelial function in diabetic patients. Improvement of insulin-stimulated glucose disposal in type 2 diabetes patients in the presence of α-lipoic-acid strengthens its therapeutic value. These experimental and clinical results suggest the importance of studies to further evaluate the potential value of antioxidants for the treatment of diabetes and its complications. Large-scale clinical trials, such as the DCCT Study or the UKPDS Study are needed to evaluate the long-term effects of these antioxidants in diabetic patients. The medical and socio-economic burden of diabetes and its complications in the near future demand further innovation in diabetes therapy. Therefore potential therapeutic agents such as antioxidants should be tested without delay by means of long-term prospective studies.