Pioglitazone Treatment Improves Nitrosative Stress in Type 2 Diabetes

This was a randomized, double-blinded trial with a placebo run-in period and two treatment arms. The primary efficacy measures were the vascular and NO responses of the foot dorsum to ischemic block and local warming.

Patients were divided into two groups. Group 1 received placebo for 12 weeks followed by 12 weeks of treatment with pioglitazone at 45 mg/day; group 2 received pioglitazone at 45 mg/day for 24 weeks.

Inclusion of a placebo-to-treatment crossover group and a pure pioglitazone-treated group enabled us to perform within-subject and between-subject statistical comparisons. The latter half of the treatment period was conducted in a single-blind manner, and all patients received active treatment during this period.

Sample size estimations were made on the basis of preliminary studies of normal healthy control subjects and age- and BMI-matched patients with type 2 diabetes (17–20). In the reported studies, the peak postischemic hyperemia measured from the fingertip was 38.25 ± 4.71 (mean ± SE) in control subjects and 23.40 ± 3.18 in matched type 2 diabetic subjects. Based on these data, a pooled standard deviation of 24.01 was used to determine power, assuming a 50% attenuation of the blood flow defect in diabetes with pioglitazone treatment (7.5 blood flow units) at 24 weeks of treatment versus placebo. For a power of at least 0.80 at α = 0.05 to detect a treatment effect of 50% attenuation, 10 subjects (5 in each arm) were required to complete the study.

All research subjects enrolled in this study were asked to sign an informed consent document. The research study was reviewed and approved by the Eastern Virginia Medical School Institutional Review Board.

We included subjects with type 2 diabetes (defined by current World Health Organization criteria) between ages 18 and 80 years. Subjects with type 1 diabetes (defined as fasting serum C-peptide concentration <1 ng/ml or diabetes onset at <35 years of age in a nonobese patient), severe retinopathy, significant proteinuria, neuropathy of a nondiabetic origin, amputations or foot ulcers, a past history of stroke or myocardial infarction, heart failure, or the presence of pedal edema, elevated liver enzymes, or any liver disease and any other serious illness that would preclude participation, including being enrolled in any other clinical trial, were excluded. Concomitant use of ACE inhibitors, angiotensin receptor blockers, and insulin was not allowed.

Tests for vascular responses and laboratory parameters were performed at weeks 0, 6, 18, and 24. At baseline a complete medical history was taken, and physical and neurological examinations were performed on each patient. HbA_1c (A1C), serum lipids, C-peptide, and liver function were also measured. These tests were performed at Eastern Virginia Medical School Laboratories, Norfolk, Virginia, or LabCorp, Burlington, North Carolina. Study medication was dispensed at weeks 0 and 12.

Blood flow and skin NO production were measured at baseline and after different stimuli were applied following the protocol shown in Fig. 1. SkBF was measured noninvasively on the dorsal aspect of the left foot by continuous laser Doppler using Periflux small angled thermostatic laser Doppler probes (Perimed, North Royalton, OH), with core diameters of 0.125 mm, laser wavelengths of 780 nm, and fiber separation lengths of 0.25 mm. This system measures skin perfusion as erythrocyte flux. The depths of the measurements, influenced by tissue properties, light source, and probe configuration, were estimated to be between 500 and 1,000 μm. NO was measured adjacent to the laser Doppler probe using a microsensor adapted for the detection of NO via an ion-exchange mechanism in the circulation of humans, which had a detection limit of 5 × 10⁻⁹ mol/l. The sensor was read by the ISO-NO Mark II NO meter (World Precision Instruments, Sarasota, FL), attached to a personal computer, and was calibrated before each use using a three-concentration nitrite standard. These methods have been described in detail elsewhere (2,17,18,20). Limb occlusion was performed with a sphygmomanometer cuff for 5 min; after baseline and ischemia-reperfusion measurements, local warming was maintained by the laser Doppler thermostatic probe at 32°C for 5 min, 40°C for 5 min, and 44°C for 10 min. Clinical safety measures included careful review of each subject before entry into the study and determination of liver enzyme elevation during treatment.

All data, including those for skin perfusion, are presented as means ± SE. The study design was parallel and crossover, with within- and between-subject analyses. The primary dependent variables were SkBF and NO responses to provocative stimuli. The independent variables were the treatment groups (two levels). For demographic data and laboratory results comparisons within and between groups, nonparametric tests (Wilcoxon’s signed-rank test and the Mann-Whitney U test, respectively) were used. Nonparametric tests were also used to examine treatment effects at 6 weeks (within and between groups) and at 24 weeks (within and between groups) on SkBF and skin NO production. Significance was accepted at the P < 0.05 level.

Five patients in group 1 (one dropout) and six patients in group 2 completed the study. Demographic characteristics of and laboratory values for each group at baseline are given in Table 1. There were no differences in any of the variables between the groups.

Table 2 shows the blood pressure responses, BMI change, and laboratory parameter change in each group after 24 weeks of treatment. There were no significant changes. The fall in A1C in pioglitazone-treated patients showed a trend toward significance (P = 0.08). No significant changes were seen at 12 weeks of treatment either (data not shown)

Figure 2A and B demonstrates the blood flow response to ischemia reperfusion (hyperemic response), local heating to 32 and 40°C, and nociceptor stimulated flow (NOCI) with local heating to 44°C. There were no significant effects on ischemia reperfusion or increase in response to warming to 32 or 40°C or to NOCI-mediated vasodilatation at 44°C in the within-group analysis. A between-group analysis was also performed at baseline, 6 weeks, and 24 weeks and showed no differences between the groups (data not shown).

The basal and heat-stimulated NO responses are shown in Figs. 3A and B, comparing baseline NO levels to those after treatment in each group (after 6 and 24 weeks). After 6 weeks, no significant reduction in NO production was seen in any of the groups. In patients treated with placebo for 12 weeks and with pioglitazone for 12 weeks (group 1), there was a decrease in skin NO production after 24 weeks of treatment, but this did not reach statistical significance. In patients treated with pioglitazone for 24 weeks (group 2), the reduction in NO production was statistically significant at baseline and after ischemia reperfusion, local heating, and NOCI stimulation (Table 3). A between-group comparison was performed, but no differences were detected at baseline, 6 weeks, or 24 weeks (data not shown).

To further elaborate on the relationship between glycemic control, insulin sensitivity, and inflammatory markers of diabetes with skin NO production, we analyzed the correlation between the changes of each variable (A1C, C-peptide, WBC count, and other parameters) and the change in NO in the basal state and after heating to 40 and 44°C, in each group after 24 weeks. The reduction in NO production correlated significantly with reductions in A1C, C-peptide, WBC count, and mean corpuscular hemoglobin concentration (MCHC) in the pioglitazone-treated group. This is shown in Table 4. We also evaluated the change in each variable in each group after 6 weeks, but no significant correlations were found at that time (data not shown).

In this study, we examined the effects of treatment with pioglitazone on SkBF and NO production. The cardinal findings were that 24 weeks of treatment did not alter impaired blood flow of the skin of the dorsum of the foot but profoundly reduced the production of skin NO. Improvement in glycemic control and insulin sensitivity (shown by a reduction in A1C and C-peptide) correlated significantly with the reduction in NO production after 24 weeks of treatment in the pioglitazone group. Furthermore, reductions in some parameters of inflammation and microvascular dysfunction in diabetes, such as WBC and MCHC, correlated significantly with the reduction in NO production as well. The lack of change in skin perfusion may suggest that regulation of SkBF is not mediated by NO. Future studies with more patients treated for longer periods with pioglitazone may provide clarification on this relationship.

These findings, at first glance, are somewhat of an anathema, because NO is one of the primary agents eliciting a vasodilatory response in conduit and resistance vessels by relaxing smooth muscle and producing an increase in blood flow (21). Because there is a considerable body of evidence indicating that NO mediates vasodilatation of brachial and forearm vessels, which is compromised in diabetes (22), we sought to determine whether there were abnormalities in NO production or utilization in the skin of people with diabetes. Using direct amperometric measurements of NO in the skin in response to heat and NOCI stimuli, we showed that there was no defect in interstitial NO production. It was produced in nanomolar amounts, as opposed to the picomolar amounts generated by vascular endothelial nitric oxide synthase (eNOS) and was only partially blockable by N^ω-nitro-l-arginine methyl ester (l-NAME). There also was no defect in the vasodilatory response to the NO donor sodium nitroprusside (23). This finding suggested that interstitial NO overproduction (presumably from macrophages, keratinocytes, and dendritic cells) might have been counterproductive and that the regulation of SkBF was not an interstitial NO-mediated phenomenon. There is an active vasodilative mechanism of skin blood vessels that is dependent on C-fiber nociceptors, which accounts for 75% of their dilative capacity (11), and there is an alternate pathway regulating blood flow that is associated with lipid levels, blood pressure, hematocrit, WBC count, and resistance to the action of insulin in diabetes and the metabolic syndrome. This increase in NO production has now been reported by other authors (24) and is thought possibly to be due to an excess of a circulating antagonist (asymmetric dimethylarginine). The suggestion has been made that this unbridled production results in NO reacting with reactive oxygen species (ROS) to form peroxynitrite (ONOO⁻), which is damaging to blood vessel structure and function (25).

Hyperglycemia and oxidative stress deplete NO within the peripheral nerves and endothelium of the microvasculature by reducing eNOS and altering nerve perfusion (23). Depletion of endothelial NO has been observed to lead to low nerve blood flow, hypoxia, and mitochondrial dysfunction (26). Accumulation of ROS during hyperglycemia downregulates neuronal NOS. These observations suggest a role of NO repletion as a potential therapy in experimental diabetic polyneuropathy (DPN). There is, however, a paradoxic response of extravascular NO to hyperglycemia and inflammatory damage to peripheral nerves. Under inflammatory conditions, inducible NOS increases, which leads to nanomolar quantity production of NO as opposed to picomolar amounts production as with eNOS activation (27). High levels of NO causes nerve damage via mitochondrial dysfunction and formation of ONOO⁻ by interaction with ROS. ONOO⁻ nitrosylates proteins, thereby impairing their function. Antioxidants typically defend the cell against this damage. Some studies have shown that patients with diabetes have reduced levels of these antioxidants (28); however, there is conflicting evidence in the literature (29). Hyperglycemia and abnormalities in lipid metabolism generate the formation of ROS. This activates β₂ protein kinase C-β, which increases the action of NAD(P)H oxidase, generating ROS (30). In addition, there is activation of the nuclear transcription factor κB with the induction of inducible NOS and reduction of eNOS (31). The increase in NO coupled with ROS to generate ONOO⁻, which nitrosylates proteins, damages the endothelia of blood vessels and damages DNA. Thus, if pioglitazone affects the NO balance in humans, we would expect to reverse the inability of subjects with diabetes to vasodilate in response to NO-mediated stimulation and reduce neurovascular damage.

Microvascular insufficiency, endoneurial blood flow, and hemodynamic factors lead to nerve damage in patients with DPN (32). Although the sequence of events is not well understood, investigators propose that microvascular vasoconstriction, edema, and ischemia play a role in DPN development (33). Endoneurial edema increases endoneurial pressure, thereby causing capillary closure and subsequent nerve ischemia and damage (34). Diminished regulation of endoneurial blood flow and ischemia may result from decreased nerve density and innervation of vessels (35,36). Nerve ischemia stimulates vascular endothelial growth factor production, exacerbating DPN via overactivation of protein kinase C-β (37). As a result, ischemia and low blood flow reduce both endothelial- and NO-dependent vasorelaxation (38). Vascular defects also result in changes in endoneural vessels. Changes in blood flow correlate with changes in oxygen saturation (39) and reduced sural nerve endoneurial oxygen tension. These changes are potentiated by increased expression or action of vasoconstrictors such as endothelin and angiotensin and decreased activity of vasodilators such as prostacyclin, substance P, calcitonin gene-related peptide, endothelial-derived hyperpolarizing factor, and bradykinin (40). In particular, in patients with DPN there is disruption in vasomotion, the rhythmic contraction exhibited by arterioles and small arteries (19). In type 2 diabetes, SkBF is abnormal, and the loss of neurogenic vasodilative mechanism in hairy skin may precede lower-limb microangiopathic processes and C-fiber dysfunction. Changes in endoneurial blood flow often are reflected by changes in nerve conduction (41). Therefore, both vascular or endoneural alterations may cause damage over time in the peripheral nerves of patients with diabetes. A multifactorial approach is needed to combat the many abnormalities associated with oxidative stress.

This study showed a reduction in skin NO production after treatment with pioglitazone but failed to demonstrate an improvement in SkBF. This may be due to the small amount of patients included and the short duration of the study. It remains to be tested whether the reduction of nitrosative stress by pioglitazone observed here translates to improved neurovascular function in the long term, thus preventing nerve damage and decreasing the predisposition to foot ulceration and amputation.

Takeda Pharmaceuticals provided grant support for this study.

We thank Sandra M. Ribeiro for study subject recruitment and data gathering.