Involvement of Peripheral Polymorphonuclear Leukocytes in Oxidative Stress and Inflammation in Type 2 Diabetic Patients

A total of 18 (9 men and 9 women) normotensive nonsmoking type 2 diabetic patients (duration of disease 36-216 months, mean 96 ± 14), aged 34-73 years (average 51.5 ± 2.5), with normal kidney function, no actual infection, and BMI <30 kg/m² were included in the study.

Patients with known cardiovascular risk factors, significant dyslipidemia(LDL cholesterol ≥ 130 mg/dl, HDL cholesterol ≤ 35 mg/dl, and triglycerides ≥ 400 mg/dl), hypertension (blood pressure > 130/85 mmHg),and severe obesity (BMI ≥ 30 kg/m²), as well as patients with evidence for microvascular or macrovascular diseases, were excluded from the study.

The diabetic patients were compared with an age- and sex-matched normal control group. The characteristics of the subjects are shown in Table 1.

Blood was withdrawn from all of the above subjects and patients after an overnight fast for biochemical and hematological parameters and PMN isolation.

PMN isolation was carried out according to the method of Klebanoff and Clark (28) and modified according to Kristal et al.(24,25). The separated PMNs (>98% pure) were resuspended in a minimal volume(0.1-0.3 ml) of cold Hank's balanced salt solution (HBSS) (Biological Industries, Beit Haemek, Israel), immediately counted, and diluted to a final volume of 1 ml with HBSS with 0.5% glucose, according to the different experimental needs. All sera were saved for in vitro survival assays.

The rate of superoxide release was assayed for 90 min under basal conditions, after stimulation with 0.32 × 10^-7 mol/l PMA (Sigma, St. Louis,MO), and after stimulation with zymosan.

The assay was based on the superoxide dismutase inhibitable reduction of 80μmol/l cytochrome C (Sigma) to its ferrous form(29) with slight modifications(24,25).

Zymosan was prepared by an incubation for 30 min at 37°C of 4 mg zymosan particles (Sigma), with 1 ml sera, pooled from 10 healthy donors. This solution was washed twice with 154 mmol/l NaCl, resuspended in 154 mmol/l NaCl at a concentration of 10 mg/ml, and stored at -80°C until it was used. In our assays, the particle-to-cell ratio was 15(30).

In a separate set of experiments, PMNs from seven randomly chosen diabetic patients and seven age- and sex-matched normal control subjects were preincubated with 100 and 500 mg/dl glucose for 20 min at 37°C, PMA was added, and the assay was carried out as previously described.

GSH determinations were carried out for all subjects. To avoid the spontaneous oxidation of GSH in plasma, the plasma was separated within 4 min of blood withdrawal and acidified (1 ml plasma with 50 μl 50% 5-sulfosalicylic acid)(24,31). At this point, the acidified plasma was frozen for a maximum of 3 weeks. Before the determination of GSH, the plasma was thawed and back-titrated to pH 7 with 0.2 mmol NaOH. The determinations of plasma GSH were carried out according to Adams et al. (32)and Griffith (33), using oxidized GSH (GSSG)-reductase with slight modifications(24). Briefly, two separate evaluations were carried out: assessment of the total GSH level in the plasma[GSH + 2(GSSG)] and assessment of the oxidized form of GSSG. The plasma level of GSH was calculated as the difference between the total GSH and the oxidized 2(GSSG). The total GSH concentration was determined using 10 mmol/l 5-5′-dithiobis, after reduction of the GSSG concentration by GSSG-reductase. The resulting amount of GSH was determined according to a standard curve (0-2.5 mol/l GSSG), recorded in a spectrophotometer (spectronic 1201; Milton Roy, Rochester, NY) at 412 nm, and expressed as the change in absorbency per minute during the enzymatic reduction with GSSG-reductase. To measure GSSG in the plasma, the initial alkylation of the sulfhydryl groups was carried out with 10 mmol/l freshly prepared N-ethylmaleimide(NEM). The excess of NEM was removed by separation on Supelclean (LC18; Sigma)columns, and the GSSG concentration was determined enzymatically in the effluent, as previously described.

The PMN counts of the blood withdrawn in EDTA were performed by using a Coulter STKS (Coulter, Miami, FL) apparatus in both the normal control subjects and the diabetic patients.

Separated PMNs(10⁷/ml) from 10 randomly chosen diabetic patients and 10 age- and sex-matched normal control subjects were incubated with the individual autologous serum (25% vol/vol diluted with HBSS) for 90 min at 37°C. In a second set of experiments, cross incubations were carried out with heterologous sera. PMNs were counted before and after 90 min of incubation. Cell viability was confirmed by trypan blue (0.1% wt/vol) exclusion. Survival was expressed as the ratio of cell counts before and after the 90-min incubation with sera (%).

Data are expressed as means ± SEM. Data are expressed as whiskers(median and range) for the description of the in vitro effect of glucose on the rate of superoxide release. The differences between the study parameters of the two groups were compared by Student's t test. Serum glucose levels, HbA_lc levels, and rate of superoxide release by PMA-stimulated PMNs were correlated by linear regression analysis. P< 0.05 was considered statistically significant.

Table 1 shows that normal control and diabetic patient groups were similar concerning age and sex distribution, blood cholesterol, triglycerides, and normal liver and kidney functions. Blood glucose and HbA_lc were significantly higher in the diabetic subjects than in normal control subjects(Table 1), as expected.

No superoxide release from resting PMNs could be detected under basal conditions during 90 min. The superoxide release from PMA-stimulated PMNs from diabetic patients was significantly greater than the superoxide release of PMA-stimulated PMNs from normal control subjects(Fig. 1). This indicates that the PMNs from diabetic patients are primed. In contrast, stimulation with zymosan, a physiological stimulant used to assess the phagocytic potential of PMNs, resulted in a significant decrease in the rate of superoxide release from diabetic PMNs (Fig. 1).

Figure 2 shows the in vitro effect of two concentrations of glucose on the rate of superoxide release from PMA-stimulated PMNs. No significant effect on glucose was found for either normal control or diabetic PMNs when compared with cells without glucose. It should be emphasized that within each group, the range of superoxide rates of release was similar, with faster rates in the diabetic patient group.

The plasma levels of GSH were significantly lower in diabetic patients than in normal control subjects(Table 1).

Figure 3 shows that in diabetes, no linear correlation could be found between the rate of superoxide release from PMA-stimulated PMNs and the same individual's blood glucose levels (n = 14, r = 0.001, P = 0.995)(Fig. 3A); however,HbA_lc and the rate of superoxide release were positively correlated(n = 11, r = 0.6, P < 0.05)(Fig. 3B).

Diabetic patients had significantly (P <0.001) higher peripheral PMN counts than normal control subjects, although all of the values were within the upper quartile of the accepted normal range(Table 1).

A significant (P = 0.001)reduction in diabetic PMN survival after an incubation of 90 min in diabetic serum was evident when compared with 90 min of incubation of normal control PMNs in normal control sera. In fact, normal control sera promoted the survival of diabetic PMNs (P = 0.001), whereas diabetic sera significantly killed normal control PMNs (P < 0.001)(Fig. 4).

We studied the properties and activities of PMNs and their possible contributions to OS and inflammation in diabetic patients. Our results show that type 2 diabetic patients are exposed to OS and chronic inflammation (partially because of the primed state of their PMNs) before any clinical evidence of angiopathy exists(Fig. 5). When separated,PMA-stimulated diabetic PMNs release superoxide significantly faster than normal control PMNs and concomitantly with a reduced level of the plasma antioxidant GSH, predisposing these patients to OS. Freedman and Hatchell(34) and Cantero et al.(26) have previously shown similar results of significantly higher levels of ROS release from stimulated diabetic PMNs in both cats and patients. Methodological differences may account for the discrepancy between these reports and that of Fuller et al.(27), who reported no changes in the rate of superoxide release from PMNs in type 2 diabetic subjects without vascular disease. In diabetes, the decrease in plasma GSH may be an outcome of greater GSH consumption by ROS. This reduced plasma GSH and the augmented rate of superoxide release showed a negative linear correlation(data not shown), which suggested that superoxide reacts directly with GSH, a phenomenon that was also supported by Thomas et al.(35). It has also been suggested that in diabetes, the antioxidant GSH is largely consumed, mainly because of a regeneration of vitamin C, which is extensively oxidized in diabetic patients (21). Lower amounts of GSH may also be explained by glycation of the enzymeγ-glutamyl cystein synthetase, which generates decreased amounts of GSH in diabetes (36).

Our observations emphasize that the oxidative metabolism of PMNs depends on the nature or type of the stimulant. By using the same assay system, we have shown that the same cell may respond in opposite directions to different stimuli. PMA stimulation of diabetic PMN caused a faster release, whereas zymosan stimulation caused a slower release of superoxide from the same diabetic PMN preparation when compared with a normal control PMN. Because zymosan is used to assess the phagocytic function of PMN, our findings also support the increased susceptibility to infections in diabetic patients, as previously described(37,38,39).

However, glucose per se did not modulate the rate of superoxide release from both diabetic and normal control PMNs in vitro. In addition, no correlation between the rate of superoxide release and blood glucose concentrations of the same individual was found. This implies that in vivo, as well as in vitro, blood glucose concentration does not directly modulate the oxidative metabolism of diabetic PMNs. In this regard, several authors have shown similar findings (40),whereas others have shown that glucose alone modulates PMN activity(41). Nevertheless, we have shown a close relation between OS and the metabolic control of diabetes,estimated by the levels of HbA_lc. Furthermore, we found a significant positive correlation between the rate of superoxide release and blood levels of HbA_lc, showing that they are closely related processes. The causative effect of OS on HbA_lc in diabetes was previously reported by Young et al.(14) and Ruiz et al.(42). The strong positive correlation between the rate of superoxide release and blood levels of HbA_lc, which to our knowledge is reported for the first time,further strengthens the relations between OS and enhanced glyco-oxidation of proteins in diabetes. Rodriguez-Manas et al.(43) have shown a close relation among endothelial dysfunction, levels of HbA_lc, and production of superoxide in an animal model of diabetic rats. Our findings and those of Rodriguez-Manas et al. suggest a close relation between the levels of HbA_lc and diabetes-associated endothelial dysfunction, underlined by OS.

The second important outcome of this study is the PMN contribution to chronic inflammation in diabetes. The existence of chronic inflammation in diabetes is mainly based on the increased plasma concentrations of C-reactive protein, fibrinogen, interleukin-6, interleukin-1, and tumor necrosis factor(44,45). Our in vivo data of increased diabetic PMN counts, combined with the decreased in vitro PMN survival rate, is evidence that primed PMNs contribute to chronic inflammation in diabetes. The leukoclastic effect of the diabetic sera suggests that extracellular factors are involved in the cellular priming of diabetic PMNs. Primed PMNs die by self-necrosis, and at the same time, they actively recruit more PMNs into circulation. The increase in diabetic PMN count, although within the upper quartile of the normal range, adds an indirect inflammatory parameter. Interestingly, the increase in leukocyte counts has been suggested as a predictor of myocardial infarction, even when the counts lie within the normal range(46). We have already described the fate of a primed PMN (in inducing OS concomitantly with cell necrosis) in clinical disorders known to be associated with high prevalence of atherosclerosis and cardiovascular morbidity, such as essential hypertension and uremia(24,25). This study adds diabetes to these disorders, and these relationships are summarized in Fig. 5. The findings of Abu El-Asrar et al.(40), which show that diabetic sera increased superoxide release from PMNs, in combination with our findings,imply that diabetic sera leukoclastic activity may be responsible for the increased superoxide release. The nature of these factors remains unknown. However, intracellular factors should not be ruled out; the magnitude of necrosis in separated PMNs, regardless of their origin, correlated significantly with the rate of their superoxide release, indicating that intracellular factors are also involved in PMN priming(24). High cytosolic calcium,as was found in diabetic PMNs, can explain such priming(47).

An understanding of the contribution of PMNs to OS and inflammation in diabetic patients may illuminate new mechanisms through which endothelial dysfunction evolves and thereby results in angiopathy and atherosclerosis. In the future, a better understanding of the functions of PMNs may be effective in evaluating different treatment modalities, long before clinical evidence of angiopathy appears.

This study was partially supported by a grant donated by the Israel Society of Nephrology and Hypertension and by Technion V.P.R. Fund (W.K. Langfan Diabetes Research Fund).