Exp. E!je RITS. (1992)
55. 767-773
Enhanced Superoxide Radical Polymorphonuclear Leukocytes SHARON Departments
(Received
F. FREEDMAN*
of Ophthalmology Administration Rockville
Production by Stimulated in a Cat Model of Diabetes AND
DIANE
L. HATCHELLT
and Cell Biology, Duke University, and the Durham Medical Center, Durham, NC 27710, U.S.A.
I1 December
1997 and accepted
in revised
form 8 April
Veterans
1992)
This study examines the possibility that polymorphonuclear leukocyte activation, which can cause endothelial injury, may contribute to the capillary closure of diabetic retinopathy. To examine diabetesrelated alterations in polymorphonuclear leukocyte activation. we compared the production of superoxide radical by these cells from normal and from diabetic cats that were maintained hyperglycemic. Polymorphonuclear leukocytes isolated from five diabetic and five normal cats were stimulated with IO ng ml-’ phorbol myristate acetate. and the maximum rate of their superoxide radical production was measured spectrophotometrically. Stimulated polymorphonuclear leukocytes from diabetic cats generated more superoxide radical, at significantly higher rates. than did those from normals (3.32 + 0.3 3 and 2.5O-tO.41 nmol O,- mine’ 10m6cells, respectively: P < 0.02). While addition of insulin or glucagon did not alter stimulated polymorphonuclear leukocyte radical production. glucose in high concentration did mildly impair its production in both groups. The exaggerated respiratory burst of polymorphonuclear leukocytes in diabetes could contribute to microvascular injury in the retina as well as in other tissues. Kq words: diabetic cat : phorbol ester: polymorphonuclear leukocyte : respiratory burst: superoxide.
1. Introduction Capillary closure and endothelial cell loss are primary events in the pathogenesis of diabetic retinopathy (Bresnick, 1989; Frank, 1989; Kohner, 1989). Polymorphonuclear leukocytes (PMNs) are important contributors to capillary resistance (Bagge. Skalak and Atteforo, 19 77 ; Schmid-Schonbein et al., 198 1: Nees and Schonharting. 1987) and, when activated, may play a key role in endothelial injury and microvascular occlusion in disease states (Engler et al., 1986: Worthen et al., 1989). Monocytes and granulocytes have recently been shown to cause capillary obstruction in a rat model of diabetic retinopathy (Schroder, Palinski and SchmidSchonbein, 199 1). Strong correlations existed in each retina between local leukocyte accumulation and other vascular pathology, such as endothelial cell damage, non-perfusion areas, and extravascular leukocytes. Animal models of diabetes provide opportunities for studying disease-related alterations under conditions of intentional chronic hyperglycemia not ethical in human subjects. We have developed a new model of diabetes in the cat, an animal with eyes large enough for surgical and experimental manipulations (Reiser et al., 1987) and in which the lens remains clear (Schall and Cornelius, 1977) allowing continued visualization of the fundus. Histologic evidence of * Hill, t Cell
Current address: Department of Ophthalmology. U.N.C.-Chapel Chapel Hill. NC 27514, U.S.A. For correspondence at: Department of Ophthalmology and Biology, Box 3802. Duke University Eye Center. Durham. NC 27710. 1l.S.A.
0011-~835/92i1107h7+07
$08.00/O
early diabetic microvascular disease has been demonstrated in this cat model (Mansour et al., 1990). PMNs from humans with diabetes have been reported to show alterations in several characteristics, including superoxide (O,-) production (the respiratory burst) and deformability, both of which might contribute to microvascular injury and occlusion. For example, there are reports of decreased filterability of leukocytes (Vermes et al., 1987) and of PMNs (DeVaro et al.. 1988) isolated from diabetic patients as compared with normal controls. There are conflicting reports regarding diabetes-related changes in PMN O,production. In the single. recently published study of the PMN respiratory burst in diabetic animals (Mohsenin and Latifpour, 1990). no significant difference was found in O,- production by PMNs from diabetic rats compared with those of normals. Some studies report an increased respiratory burst by PMNs from diabetic humans (Baranao et al., 1987). but others report no significant difference (Kjersem et al., 1988 : WieruszWysocka et al., 1987, 1989), or even decreased O,production by stimulated PMNs from diabetic humans as contrasted with controls (Herskowitz. Matsutani and Permutt. 1990: Shah, Wallin and Eilen. 1983). O,- production by resting PMNs from diabetic patients has been reported to exceed that by PMNs from normal controls (Wierusz-Wysocka, 198 7 : Shah. Wallin and Eilen, 198 3 ). The effect of elevated ambient glucose on the PMN respiratory burst remains unclear: some investigators find inhibition of the burst (Nielson and Hindson. 1989; Wilson, Tomlinson and Reeves. 1987) while others find no effect (Pickering, 0
1992 Academic
Press Limited
768 Cleary and Getz, 1982: Shah, Wallin and Eilen. 1983). Given the evident disagreement in the literature concerning the effect of the diabetic state on the PMN respiratory burst, we thought it important, in a welldefined animal model. to compare normal and chronically diabetic cats with respect to the respiratory burst of their PMNs. To explore the potential role of PMNs in early diabetic retinopathy, we began by measuring the respiratory burst (an important feature of activation) of PMNs in normal and diabetic cats using the potent, soluble stimulus phorbol 12-myristate 13-acetate (PMA). We also assessed the effect of glucagon, insulin and glucose on 01~- production by these cells. Our findings of an average 3 3 % increase in respiratory burst by stimulated PMNs from the diabetic animals (compared to controls) are reported below.
2. Materials and Methods
Animal care and use in this study were in accordance with the Declaration of Helsinki and Public Health Service Policy on Humane Care and Use of Laboratory Animals (DHEW Publication. NIH 80-2 3 ). Ten adult male cats weighing 4.5-55 kg were used in these studies. Five cats were maintained as normals, while five were rendered diabetic by partial pancreatectomy (Reiser et al.. 198 7). Diabetic cats were given a single morning injection of protamine zinc insulin for at least 6 months prior to study. Insulin doses were adjusted for each cat to keep morning blood glucose between 16.8 and 2 5.2 mM. Morning glucose for the diabetic cats was 22.2 f2.0 mM (mean+S.D.). with a corresponding value of 4.0 + 1.2 mM for the normal cats.
PMN lsolatiorl Venous blood was drawn (into EDTA-containing tubes) from non-fasting cats in the morning prior to daily insulin administration, and 5 min after IM injection of ketamine (15 mg kg-‘). PMNs were isolated under sterile conditions, and all unnecessary manipulations were avoided, to prevent inadvertent PMN activation. Following 1: 1 dilution of blood with 20 mM Hepes-buffered HBSS. free of Ca2+, Mg”+. and phenol red (Hepes/HBSS-, pH 7.4 : Gibco Laboratories, Grand Island, NY. U.S.A.), PMNs were isolated using a Ficoll-Hypaque (Sigma Chemical Co., St Louis, MO, 1J.S.A.) double-density centrifugation method (English and Anderson, 1974). PMNs were then washed with cold Hepes/HBSS and resuspended in Hepes-buffered phenol red-free HBSS (Hepes/HBSS+, pH 7.4, 5.6 mM glucose) to 2-X x 106 cells ml-‘. This method yielded a mixture of PMNs and eosinophils. with some contaminating red blood cells (RBCs), as assessed by
S. F. FREEDMAN
AND
D. L. HATCHELL
Wright’s stain. Cell viability was > 99% by Trypan blue exclusion, Hypotonic lysis was not routinely performed, and RBC contamination averaged 26 y,. Control experiments demonstrated that neither addition of excess RBCs nor hypotonic lysis followed by washing of cells affected 0, production by PMNs in our assays. Cell counts (by phase contrast microscopy J were normalized to obtain equal numbers of leukocytes in each experiment. Eosinophil content of leukocyte preparations was in the normal range for healthy adult cats (Duncan and Prasse. 1988). and averaged 10 -t 6 % for normal cats and I 1 f 4 ‘X, for diabetic cats. One diabetic cat showed persistent hypereosinophilia, with eosinophils constituting 17.73 % of total leukocytes: all data from this cat were excluded from further analysis. For comparison studies, human PMNs were prepared from venous blood of healthy male volunteers (aged 20-3 5 years), using identical methods. RBC contamination averaged 40% in these samples. Eosinophil content was < 3 ‘,${, in these preparations. Cells were maintained on ice prior to assay. and were used within 4 hr of isolation. Cells were manipulated only in plastic vessels.
Measurrvwnt of‘ PMN Suprroxirk Proiluction Superoxide production was measured spectrophotometrically by superoxide dismutase-inhibitable cytochrome c reduction (Cohen and Chovaniec. 1978 ). Assays were performed in duplicate or triplicate at pH 7.4, 37”C, in disposable plastic cuvettes. Standard assays contained 50 &t~ ferricytochrome c (type III. Sigma Chemical Co.) and 0.5 x 10” PMNs ml-’ in 2.5 ml Hepes/HBSS+. After temperature equilibration for 3-5 min in a water bath, absorbance at 550 nm (A& was measured continuously (human cells) or at 5-min intervals (cat cells) for 30 min, in the presence or absence of stimulus. Cuvettes were returned to a 37°C water bath between readings. Control assays were performed in the presence of superoxide dismutase (SOD, 40 pugml-‘), Diagnostic Data, Inc., Mountainview, CA. U.S.A.), which completely inhibited cytochrome c reduction. Phorbol myristate acetate (PMA. Sigma Chemical Co.) was stored as a 1 mg ml-’ stock solution in dimethylsulfoxide at -2O’C and a fresh aliquot was diluted with Hepes/HBSS+ for each day’s assays. Following addition of PMA, cytochrome c reduction (via Ass,,) increased gradually to a limiting linear rate, which was taken to be the maximum rate of 0, production [see Fig. 1(A)]. The lag time was calculated from the intercept of the linear portion of the curve (which persisted for lo-15 min) with the pre-activation baseline (Cohen et al.. 1981). Unstimulated O,- production was estimated in parallel assays. Cytochrome c was not rate-limiting under these assay conditions, and addition of catalase (Boehringer Manheim, Indianapolis, IN. 1J.S.A.) to prevent ac-
SUPEROXIDE
RADICAL
PRODUCTION
IN DIABETES
769
L 0.0
I .o
0.5 Cell
density
Cmtlllon
cells
ml-‘1
FIG. 1. O,- production of cat PMNs stimulated with PMA. A, Increase in A,,, over time following addition of PMA at 1000 ng ml-’ (n), 10 ng ml-’ (@)* 1 ng ml-’ (O), or 0.5 ng ml-l (m). Points shown (corrected for SOD-insensitive changes in A,,,) represent the mean of duplicate determinations performed in a representative experiment on a single cat. Five similar experiments were performed. B, Maximum linear rate of O,- production (0) and corresponding lag (m) are plotted as a function of PMA concentration. Points shown are mean+s.~. of duplicate assays performed in a representative experiment as in (A) above. C, Maximum linear rate of O,- production (0) and lag (m) as a function of cat PMN cell density. Each point represents mean of duplicate determinations as in (A) above.
cumulation of H,O, did not alter measured rates of Ozmproduction. Simultaneous duplicate and triplicate assays exhibited less than a 10% coefficient of variation, but there was considerable day-to-day variation in O,- production by PMNs from a given cat. In all comparisons between PMNs from normal and diabetic cats, each cat is used as the biologic unit of comparison (mean results from two to five different experiments are used for each cat). Unless otherwise stated, values given represent mean f s.D., with n = 4 for diabetic cats, and in = 5 for normal cats. Statistical significance was determined using Student’s unpaired t-test. When indicated, cells were preincubated with additional glucose. insulin, or glucagon for 30 min at 37°C prior to stimulation with PMA (10 ng ml-‘).
Control assays contained 5.6 mM glucose, and lacked insulin and glucagon. Nitroblue Tetrazolium (NBT) reduction NBT reduction was performed by incubating 1 x lo6 PMN ml-’ in Hepes/HBSS+ containing 0~04% NBT and 1 y0 albumin at 3 7°C for O-30 min in the presence or absence of PMA (Cohen et al., 1981). Following termination of the reaction by addition of SOD (20 ,ug ml-l), cells were spun and washed with cold phosphate-buffered saline, resuspended in 5 y0 albumin, then dried and heat-fixed on glass slides. Cells were counterstained with 1 y0 safranin, and 100 consecutive individual cells were examined for formazan deposits.
S. F. FREEDMAN
770
I
TABLE
by PMNs from normal and diabetic cats
Superoxide production
Superoxide production (nmol min’ lo-’ cells) Maximally stimulated
Unstimulated Normal cats (n = 5) Diabetic cats
Lag time (min)
0.12 * om
2,5O+O+kl
6.6kl.O
0.2 1 * 0.0 5
3.32 +03 3*
7.1& 1.4
(n = 4) Values are meansi s.u. * P K 0.02 vs. normal cats.
I
2
3
Normal
4 cats
5
I Normal
I Diabetic
Meons
6
7
8
Diabetic
9 cats
Fro. 2. PMA-stimulated O,- production by PMNs from diabetic and normal cats: maximum rate of O,- production in response to PMA (10 ng ml-‘). Each point (0) represents mean value of duplicate or triplicate determinations. PMNs from each cat were assayed in two to five independent experiments. Group mean + s.n. for normal cats nos 1-5 and diabetic cats nos 6-9 are shown (a): P < 0.02 for the difference between these means.
3. Results Superoxide Production by PMNs from Cats and Humans in Response to PMA Following addition of PMA to a cuvette containing PMNs and cytochrome c. there was an initial time period with little change in A,,,, then a gradual increase in Ajso until a linear rate of change was achieved [Fig. l(A)]. SOD abolished this response. Figures l(A) and (B) show the dose-response relationship between PMA concentration and O,- production. PMNs from both normal and diabetic cats were very sensitive to PMA, demonstrating maximum rates of O,- production in response to as little as 1 ng ml-’ PMA. Lag times shortened as the PMA concentration increased, but did not fall below 5 min even at 1000 ng ml-’ PMA [Fig. l(B)]. The PMA-stimulated rate of O,- production was proportional to the concentration of cells, whereas the lag was indepen-
AND
D L. HATCHELL
dent of this variable up to at least 1 x 10” cells ml-’ [Fig. l(C)]. Using our procedures, human PMNs (from two healthy male volunteers) stimulated with 1000 ng ml-’ PMA produced 6,7+ 3.5 nmol O,- min’ I()-” cells with a lag time of 1.4 & 0.1 min. Superoxide Production by PMNs from Normal ami Diabetic Cats Maximally stimulated (PMA 10 ng ml ’ ) PMNs from diabetic cats generated O,- at significantly higher rates than did those from normals (Table J). Lag times were similar for both groups (Table I). Total 0, generated over 30 min was higher for stimulated PMNs from diabetic than from normal cats (67.0 + 6.5 vs. 5 1.8 + 8.0 nmol O,- lVG cells, respectively, P < 0.05). The differences between PMNs from normal and from diabetic cats, which were evident at 10 ng ml-’ PMA. were not detectable at I ng ml-’ PMA. Thus when PMNs were exposed to 1 ng ml-’ PMA, maximum rates of O,- production by PMNs from diabetic and normal cats were similar (1.84 & 0.88 and 1.9 3 + 0.02 nmol O,- min-’ ZOF cells, respectively), as were the corresponding lag times (14.1 k 3.2 and 1 3.h+ 2.7 min, respectively). There was considerable variability in 0, production among cats within each group, as well as for a given cat on different days (Fig. 2). For example, five serial assays of maximally stimulated PMNs from the same normal cat (no. 3) over a 3-month period. showed a mean O,- production rate of 2.47-_t:O46 nmol 0, min. 1 1o-~li cells with a corresponding lag time of 7.5 1.09 min. Similarly, three assays of stimulated PMNs from the same diabetic cat (no. 8) yielded values of 3.20 & 0.31 nmol O,- min’ 1OF cells and 8.2 i O-5 min, respectively. IJnstimulated PMNs from both normal and diabetic cats showed low levels of O,- production (Table I). Three serial determinations of unstimulated 0, production in a single normal and a single diabetic cat yielded values of 004 k 005 and 022 i 0.1 1 nmol 0, min 1 1OY cells. respectively. When resting 0, production rates were subtracted from total maximally stimulated production rates by PMNs in each respective experiment. PMNs from diabetic cats showed higher net rates of O,- production than did those from normal cats (3.11 kO.37 and 2.44+039 nmol 0, min-’ 1OF cells, respectively, P < 0~05 1. Unstimulated PMNs showed no formazan staining during 30 min of incubation with NBT. In contrast, when PMA ( 10 ng ml-’ ) was added to PMNs in the presence of NBT. 95-97%) of cells were stained with formazan after 15 min. Similar results were obtained with PMNs from normal and those from diabetic cats. Low levels of O,- production by unstimulated PMNs were hence attributed to generalized low-level production rather than to maximal O,- production by a few activated PMNs among many quiescent neighbors.
SUPEROXIDE
RADICAL
PRODUCTION TABLE
IN DIABETES
II
Effect of added glucose. insulin, and glucagon on PMAstimulated rate of superoxide production by cat PMNs
~~ Glucose
22.4 mM 33.6 mM
Insulin
6rnM 400 ng 1 1
Glucagon
Normal cats (“/, of control) ~-~ ~.~~ 97*4 (3) 86k9.t (3) 90222 (3) 98k16 (3)
Diabetic cats (% of control) 92 * 1* (3) 86?c 3-t (3) 97f 16 (3) 93*1 (2)
Values are means i S.D.. with II (number of cats tested) in parentheses * P < 005 vs. control. t 1’ < (Hli for both groups together (n = 6 cats) vs. control.
III Vitro effects 01 Glucose, Insulin, and Glucagon O,- production by PMNs was routinely assayed in the presence of 5.6 mM glucose; however, diabetic cats in our study were consistently hyperglycemic compared with normal cats, We therefore examined the effects of in vitro exposure of PMNs to elevated glucose levels. PMNs were exposed to 56-3 3.6 IIIM glucose for 30 min at 3 7°C prior to addition of PMA (10 ng mll’). As shown in Table II, exposure to 22.4 mM glucose minimally affected PMA-stimulated O,- production by PMNs from both diabetic and normal cats. Exposure to 33.6 mM glucose mildly impaired O,- production by PMNs from both groups of animals. Since cats had been rendered diabetic by partial pancreatectomy. we examined the effects of both insulin and glucagon on O,- production by PMNs. Exposure to insulin at 6 nM or at 6 PM for 30 min did not affect subsequent PMA-stimulated O,- production (Table II). Exposure to 400 ng ll’ glucagon was also without significant effect (Table II). 4. Discussion We have developed a new feline model of diabetes, with the long-term goal of elucidating the pathogenesis of diabetic retinal microvascular disease, specifically the role of PMNs in this process. We began by characterizing the respiratory burst of the PMNs in response to the soluble stimulus PMA. Unlike human PMNs, cat PMNs have been reported to exhibit no chemotaxis towards N-formyl-methionyl-leucylphenylalanine (FMLP) (Gray et al., 1986). We similarly found no PMN respiratory burst in response to FMLP at concentrations up to 1 pM (data not shown). We chose to examine O,- production because it can be conveniently measured in a kinetic mode, and faithfully reflects both oxygen consumption and hydrogen peroxide accumulation which accompany the respiratory burst of PMNs (Makino et al., 1986). Cat PMNs exhibited greater sensitivity to PMA than that reported for human PMNs, and showed longer lag times at saturating PMA concentrations than do human cells (Newburger. Chovaniec and Cohen,
771
1980). When maximally stimulated, cat PMNs produce O,- at rates lower than those reported for human cells (Newburger et al., 1980; Tauber and Babior, 1985). These species-related differences in the PMN respiratory burst are not surprising. since comparable differences in the magnitude of the respiratory burst have been noted for PMNs of other species (Young and Beswick, 1986). The PMNs from our diabetic cats, when stimulated with PMA, demonstrated maximum rates of O,production on average 33 % higher than those of controls. The seemingly remote possibility that this difference derives from trauma experienced at surgery more than 6 months prior to these studies, or was due to daily injections of heterologous insulin in the diabetic animals was not explored. A study of the reversibility of the effect of diabetes on the PMN respiratory burst, by placing diabetic animals in tight glycemic control, would also ultimately be of greater interest. A number of diabetes-related alterations of human PMN function have been identified, changes in the respiratory burst among them. However, the literature reporting these changes is conflicting. Using opsonized zymosan as a stimulus, Baranao et al. (198 7, 1988) found increased luminol-enhanced chemiluminescence as evidence of an increased respiratory burst in PMNs from patients with both insulin-dependent diabetes mellitus (IDDM) and non-insulin-dependent diabetes mellitus (NIDDM) compared with normals. In contrast, using similar stimulus and assay techniques, Wincour et al. (1988) found no difference in the stimulated response of PMNs from normals and subjects with NIDDM. Also using opsonized zymosan (Wierusz-Wysocka et al., 1987) or PMA (Kjersem et al., 1988) to stimulate PMNs, others have found no difference in O,- production between PMNs from patients with IDDM and those of normal controls. Several studies have even shown markedly reduced O,- production by PMNs from patients with IDDM, using either PMA or opsonized zymosan as stimuli (Herskowitz et al., 1990; Shah et al., 1983). The disparate results of these various invesugations cannot easily be reconciled simply by considering differences in patient populations, methods of PMN preparation, or assay methods described. Use of an animal model in our studies eliminated the variable of degree of metabolic control of each patient, because all animals were kept in comparably poor metabolic control throughout the study. To our knowledge, there is a single, recently published study of the PMN respiratory burst in diabetic animals (Mohsenin and Latifpour, 1990). In that study, an increase in total O,production by stimulated PMNs (measured at a single point 20 min after stimulation with high-dose PMA) from diabetic rats (as compared with normal controls) was noted, but did not reach statistical significance. Unstimulated PMNs from our poorly controlled diabetic cats produced more O,- than those from
S. F. FREEDMAN
772
normal cats, but this difference was not statistically significant. The O,- production of unstimulated PMNs from diabetic patients has similarly been reported to moderately exceed that of normal controls (Shah et al., 1983; Wierusz-Wysocka et al., 1987). Exposure of cat PMNs to elevated glucose concentrations, similar to those present in diabetic cat blood (22,4 mM), minimally affected O,- production by stimulated PMNs. The elevated PMN respiratory burst noted in stimulated PMNs from diabetic cats may therefore be due to factors other than that of short term hyperglycemia. While some studies similarly report no effect of elevated glucose on the human PMN respiratory burst in vitro (Pickering et al., 1982; Shah et al.. 1983), others have noted inhibition of the burst with glucose concentrations ranging from 11 mM (Nielson and Hindson. 1989) to 50 mM (Wilson, Tomlinson and Reeves, 1987). The elevated PMN respiratory burst noted in our diabetic cats seems unlikely to be a direct consequence of short term insulin or glucagon deficiency per se, since exposure of diabetic or normal cat PMNs to either of these hormones in vitro did not affect PMAstimulated O,- production. Similar results have been reported in human studies (Shah et al., 1983). We have found enhanced in vitro O,- production by activated PMNs from diabetic cats that have been purposely kept hyperglycemic for at least 6 months. Whether the quantities of superoxide produced by diabetic cat PMNs are sufficient to cause intravascular damage remains unclear: if not O,-, then other toxins produced subsequent to dismutation of O,- (H,O,, HOCl, HO.) may cause tissue injury. Enhancing the potential for microvascular injury by activated PMNs from diabetics is the notion that these PMNs may traverse capillary beds more slowly than PMNs from normal subjects, as suggested by studies of human PMN deformability (Vermes et al., 1987: DeVaro et al., 1988). Our findings are consistent with observations made in the rat model of diabetic retinopathy (Schroder, Palinski and Schmid-Schonbein, 199 1). The percentage of circulating granulocytes and monocytes activated in diabetic rats has been shown to be substantially greater than in control rats. Furthermore, capillary occlusion by granulocytes was documented in perfusion-fixed retinas of only diabetic rats, Interventions aimed at reducing the augmented PMN activation seen in diabetic rats and, now, cats may lead to preventive therapies for diabetic retinopathy. Acknowledgements These studies were supported by Veterans Administration Medical ResearchFunds (Dr Hatchell), National Eye Institute Research Grant EY02903 (Dr Hatchell) and Core Grant EY05722, and a generous gift from Dr and Mrs Harris Vernick.
Dr Freedman
was supported
by Postdoctoral
Research Fellowship Grant PD90036 from the Fight for Sight Research Division of the National Society to Prevent
AND
D. L. HATCHELL
Blindness, awarded in memory of Mary 13.and Alexander P. Hirsch. We are grateful to Professor Irwin Fridovich for providing laboratory space and consultation. and to Dr Gregory Samsa for statistical assistance.
References Armstrong. M.. Jr., Needhamm, D.. Hatchell. U. L. and Nunn, R. S. (1990). Effect of pentoxifylline on the flow of polymorphonuclear leukocytes through a model capillary. Angiology 41, 2 5 3-6 1. Bagge, Ii., Skalak, R. and Atteforo. R. ( 1Y7i). Granulocyte rheology. Experimental studies in an in vitro microflow system. Adv. A4icrocirr. 7, 2948. Baranao. R. I.. Garberi, J. C.. Tesone, P. A. and Rumi. R. S. (1987). Evaluation of neutrophil activity and circulating immune complexes levels in diabetic patients. Hot-m. Metabol. Res. 19. 371-4. Baranao. R. I.. Rumi. L. S.. Tesone, P. A. and Foglia. V. G. (1988). Some characteristics of neutrophils from diabetic patients and their relation to the levels of circulating immune complexes. Acta Diabetol. Lat. 25. 13-23. Bresnick, G. H. (1989). Background diabetic retinopathy. Jn Retina, Vol. II (EdsRyan, S. J.. Schacat, A. P., Murphy, R. P. and Patz, A.). Pp. 327-66. C. V. Mosby : St Louis, MO. [J.S.A. Cohen. H. J. and Chovaniec, M. E. ( 1978). Suoeroxide generation by digitonin-stimulated guinea pip granulocvtes. 1. Clin. Invest. 61. 1081-7. Cohen, H. J.,’ Newburger, P.E.. Chovaniec. M. E.. Whitin. J. C. and Simons, E. R. (1981). Opsonized zymosanstimulated granulocytes-activation and activity of the superoxide-generating system and membrane potential changes. Blood 58, 975-81. Devaro. J., Chumdermpadetsuk, R.. Edelman, K.. Sinclair. S. and MacGregor, R. R. ( 1988). Polymorphonuclear (PMN) filterability in diabetes mellitus. Invest. Ophthalmol. Vis. Sci. 29 (Suppl.), 181. Duncan, J. R. and Prasse. K. W. (1988). Appendix 1. Veterinary Luboratory Medicine. Iowa State IJniv. Press : Ames, Iowa, IJ.S.A. Engler. R. L.. Dahlgren, M. D., Morris. D. D., Peterson. M. A. and Schmid-Schonbein. G. W. ( 1Y86). Role of leukocytes in response to acute myocardial ischemia and reflow in dogs. Am. J. Physiol. 251, H314-H322. English, D. and Andersen. B. R. (1974). Single-step separation of red blood cells, granulocytes and mononuclear leukocytes on discontinuous density gradients of Ficoll-Hypaque. 1. lmmunol. Methods 5. 249-52. Frank, R. N. (1989). Etiologic mechanisms in diabetic retinopathy. In Retina, Vol. II (Eds Ryan. S. J., Schacat. A. P., Murphy, R. P. and Patz, A.). Pp. 301-26. C. V. Mosby : St Louis, MO, U.S.A. Gray, G. D.. Ohlmann, G. M.. Morton, D. K. and Schaub, R. G. ( 1986). Feline polymorphonuclear leukocytes respond chemotactically to leukotriene B, and activated serum but not to F-met-leu-phe. Agents Actions 18. 40-h. Herskowitz, I. C., Matsutani, A. and Permutt. M. A. ( 1990). Leukocyte dysfunction in diabetes is proportional to metabolic control of the diabetic state. Diabetes 39 (Suppl. I ). 27. Kjersem. I-1.. Hilsted. J.. Madsbad. S., Wandall, J. II.. Johansen. S. and Borregaard. N. (1988). Polymorphonuclear leucocyte dysfunction during short term metabolic changes from normo- to hyperglycemia in type I (insulin dependent) diabetic patients. Infection 16, 2 15-20. Kohner. R. H. (1989). Retinal ischemia. In Retina. Vol. II (Eds
SUPEROXIDE
RADICAL
PRODUCTION
IN DIABETES
Ryan, S. J.. Schacat, A. P., Murphy, R. P. and Patz, A.). Pp. 89-97. C. V. Mosby, St Louis, MO, U.S.A. Makino, R.. Tanaka, T., Iizuko. T., Ishimura, Y. and Kanegasake, S. (1986). Stoichiometric conversion of oxygen to superoxide anion during the respiratory burst of neutrophiis. J. Biol. C&m. 261. 11444-47. Mansour, S. Z.. Hatchell. D. L.. Chandler, D., Saloupis, P. and Hatchell. M. C. (1990). Reduction of basement membrane thickening in diabetic cat retina by sulindac. lnvrst. Ophthnlmol. Vis. Sci. 31, 457-63. Markert. M., Andrews, P. C. and Babior, B. M. (1984). Measurement of O,- production by human neutrophils. The preparation and assay of NADPH oxidase-containing particles from human neutrophils. Methods Enzgmol. 105. 358-65. Mohsenin. V. and Latifpour. J. (1990). Respiratory burst in alveolar macrophages of diabetic rats. J. Applied Physiol. 68, 7 384-90. Nees. S. and Schonharting, M. (198 7). Role of different blood cell types in blood rheology. In: Pentoxityllinr and Leukocyte Function (eds Mandell, J. L. and Novack, W. J.). Pp. 105-14, Hoechst-Roussel Pharmaceutied, Inc., Somerville. NJ. Newburger. P. E., Chovaniec, M. E. and Cohen, H. J. (1980). Activity and activation of the granulocyte superoxidegenerating system. Blood 55. 85-92. Nielson, C. I?. and Hindson. D. A. (1989). Inhibition of polymorphonuclear leukocyte respiratory burst by elevated glucose concentrations in vitro. Diabetes 38. 1031-5. Pickering, L. K.. Cleary. T. G. and Getz. S. ( 1982). Effect of glucose concentration and time on polymorphonuclear leukocyte function. J. C/in. Lab. Irnmunol. 8, 31-5. Reiser. H. J., Whitworth. U. G., Hatchell, D. L.. Sutherland. F. S.. Nanda. S.. McAdoo, T. and Hardin, J. R. (1987). Experimental diabetes in cat: partial pancreatectomy alone or combined with local injection of alloxan. Lab. Anirn. Sci. 37, 444-52. Schall, W. D. and Cornelius, I,. M. (1977). Diabetes mellitus. Vet. Clirl. North Am. 7, 613-28. Schmid-Schonbein. G. W., Sung. K. L. P., Tozeren, H..
773 Skalak. R. and Chien, S. (1981). Passive mechanic properties of human leukocytes. Biophys. 1. 36,243-56. Schroder, S.. Palinski, W. and Schmid-Schonbein, G. W. (1991). Activated monocytes and granulocytes, capillary nonperfusion and neovascularization in diabetic retinopathy. Am. 1. Puthol. 139, 81-100. Shah, S. V., Wallin, J. D. and Eilen, S. D. (1983). Chemiluminescence and superoxide anion production by leukocytes from diabetic patients. 1. Clin. Endocrinol. Metab. 57, 402-9. Tauber, A. I. and Babior. B. M. (1985). Neutrophil oxygen reduction: the enzymes and the products. Adv. Free Rad. Biol. Med. 1, 265-307. Vermes, I.. Steinmetz, E. T.. Zeyen, L. J. J. M. and van der Veen, E. A. (1987). Rheologic properties of white blood cells are changed in diabetic patients with microvascular complications. Diabetologia 30. 434-6. Wierusz-Wysocka, B.. Kubicka, M., Wykretowicz, A.. Kasprzak. W. P. & Wysocki, H. (1989). Superoxide anion production and adhesiveness of neutrophils in obese patients. Pol. Arch. Med. Wewn. 82, 2-3, 98-104. Wierusz-Wysocka, B., Wysocki, H., Siekierka. H., Wykretowicz. A., Szczepanik, A. and Klimas, R. (198 7). Evidence of polymorphonuclear neutrophils (PMN) activation in patients with insulin-dependent diabetes mellitus. I. Leuk. Biol. 42. 519-23. Wilson, R. M., Tomlinson. D. R. and Reeves. W. G. ( 198 7). Neutrophil sorbitol production impairs oxidative killing in diabetes. Diabetic Med. 4, 422-34. Wincour, P. H.. Lenton. J.. Puxty. J. A. H. and Anderson. D. C. (1988). Leukocyte microbicidal activity assessed by chemiluminescence in elderly non-insulin dependent diabetes mellitus. Diabetes Res. 9, 73-5. Worthen. S. G., Schwab, B.. III. Elson, E. L. and Downey, G. P. (1989). Mechanics of stimulated neutrophils: cell stiffening induces retention in capillaries. Science 245, 183-6. Young, S. and Beswick, P. (1986). Comparison of the oxidative reactions of neutrophils from a variety of species when stimulated by opsonized zymosan and f-met-leu-phe. 1. Camp. Pathol. 96, 189-96.
55. 767-773
Enhanced Superoxide Radical Polymorphonuclear Leukocytes SHARON Departments
(Received
F. FREEDMAN*
of Ophthalmology Administration Rockville
Production by Stimulated in a Cat Model of Diabetes AND
DIANE
L. HATCHELLT
and Cell Biology, Duke University, and the Durham Medical Center, Durham, NC 27710, U.S.A.
I1 December
1997 and accepted
in revised
form 8 April
Veterans
1992)
This study examines the possibility that polymorphonuclear leukocyte activation, which can cause endothelial injury, may contribute to the capillary closure of diabetic retinopathy. To examine diabetesrelated alterations in polymorphonuclear leukocyte activation. we compared the production of superoxide radical by these cells from normal and from diabetic cats that were maintained hyperglycemic. Polymorphonuclear leukocytes isolated from five diabetic and five normal cats were stimulated with IO ng ml-’ phorbol myristate acetate. and the maximum rate of their superoxide radical production was measured spectrophotometrically. Stimulated polymorphonuclear leukocytes from diabetic cats generated more superoxide radical, at significantly higher rates. than did those from normals (3.32 + 0.3 3 and 2.5O-tO.41 nmol O,- mine’ 10m6cells, respectively: P < 0.02). While addition of insulin or glucagon did not alter stimulated polymorphonuclear leukocyte radical production. glucose in high concentration did mildly impair its production in both groups. The exaggerated respiratory burst of polymorphonuclear leukocytes in diabetes could contribute to microvascular injury in the retina as well as in other tissues. Kq words: diabetic cat : phorbol ester: polymorphonuclear leukocyte : respiratory burst: superoxide.
1. Introduction Capillary closure and endothelial cell loss are primary events in the pathogenesis of diabetic retinopathy (Bresnick, 1989; Frank, 1989; Kohner, 1989). Polymorphonuclear leukocytes (PMNs) are important contributors to capillary resistance (Bagge. Skalak and Atteforo, 19 77 ; Schmid-Schonbein et al., 198 1: Nees and Schonharting. 1987) and, when activated, may play a key role in endothelial injury and microvascular occlusion in disease states (Engler et al., 1986: Worthen et al., 1989). Monocytes and granulocytes have recently been shown to cause capillary obstruction in a rat model of diabetic retinopathy (Schroder, Palinski and SchmidSchonbein, 199 1). Strong correlations existed in each retina between local leukocyte accumulation and other vascular pathology, such as endothelial cell damage, non-perfusion areas, and extravascular leukocytes. Animal models of diabetes provide opportunities for studying disease-related alterations under conditions of intentional chronic hyperglycemia not ethical in human subjects. We have developed a new model of diabetes in the cat, an animal with eyes large enough for surgical and experimental manipulations (Reiser et al., 1987) and in which the lens remains clear (Schall and Cornelius, 1977) allowing continued visualization of the fundus. Histologic evidence of * Hill, t Cell
Current address: Department of Ophthalmology. U.N.C.-Chapel Chapel Hill. NC 27514, U.S.A. For correspondence at: Department of Ophthalmology and Biology, Box 3802. Duke University Eye Center. Durham. NC 27710. 1l.S.A.
0011-~835/92i1107h7+07
$08.00/O
early diabetic microvascular disease has been demonstrated in this cat model (Mansour et al., 1990). PMNs from humans with diabetes have been reported to show alterations in several characteristics, including superoxide (O,-) production (the respiratory burst) and deformability, both of which might contribute to microvascular injury and occlusion. For example, there are reports of decreased filterability of leukocytes (Vermes et al., 1987) and of PMNs (DeVaro et al.. 1988) isolated from diabetic patients as compared with normal controls. There are conflicting reports regarding diabetes-related changes in PMN O,production. In the single. recently published study of the PMN respiratory burst in diabetic animals (Mohsenin and Latifpour, 1990). no significant difference was found in O,- production by PMNs from diabetic rats compared with those of normals. Some studies report an increased respiratory burst by PMNs from diabetic humans (Baranao et al., 1987). but others report no significant difference (Kjersem et al., 1988 : WieruszWysocka et al., 1987, 1989), or even decreased O,production by stimulated PMNs from diabetic humans as contrasted with controls (Herskowitz. Matsutani and Permutt. 1990: Shah, Wallin and Eilen. 1983). O,- production by resting PMNs from diabetic patients has been reported to exceed that by PMNs from normal controls (Wierusz-Wysocka, 198 7 : Shah. Wallin and Eilen, 198 3 ). The effect of elevated ambient glucose on the PMN respiratory burst remains unclear: some investigators find inhibition of the burst (Nielson and Hindson. 1989; Wilson, Tomlinson and Reeves. 1987) while others find no effect (Pickering, 0
1992 Academic
Press Limited
768 Cleary and Getz, 1982: Shah, Wallin and Eilen. 1983). Given the evident disagreement in the literature concerning the effect of the diabetic state on the PMN respiratory burst, we thought it important, in a welldefined animal model. to compare normal and chronically diabetic cats with respect to the respiratory burst of their PMNs. To explore the potential role of PMNs in early diabetic retinopathy, we began by measuring the respiratory burst (an important feature of activation) of PMNs in normal and diabetic cats using the potent, soluble stimulus phorbol 12-myristate 13-acetate (PMA). We also assessed the effect of glucagon, insulin and glucose on 01~- production by these cells. Our findings of an average 3 3 % increase in respiratory burst by stimulated PMNs from the diabetic animals (compared to controls) are reported below.
2. Materials and Methods
Animal care and use in this study were in accordance with the Declaration of Helsinki and Public Health Service Policy on Humane Care and Use of Laboratory Animals (DHEW Publication. NIH 80-2 3 ). Ten adult male cats weighing 4.5-55 kg were used in these studies. Five cats were maintained as normals, while five were rendered diabetic by partial pancreatectomy (Reiser et al.. 198 7). Diabetic cats were given a single morning injection of protamine zinc insulin for at least 6 months prior to study. Insulin doses were adjusted for each cat to keep morning blood glucose between 16.8 and 2 5.2 mM. Morning glucose for the diabetic cats was 22.2 f2.0 mM (mean+S.D.). with a corresponding value of 4.0 + 1.2 mM for the normal cats.
PMN lsolatiorl Venous blood was drawn (into EDTA-containing tubes) from non-fasting cats in the morning prior to daily insulin administration, and 5 min after IM injection of ketamine (15 mg kg-‘). PMNs were isolated under sterile conditions, and all unnecessary manipulations were avoided, to prevent inadvertent PMN activation. Following 1: 1 dilution of blood with 20 mM Hepes-buffered HBSS. free of Ca2+, Mg”+. and phenol red (Hepes/HBSS-, pH 7.4 : Gibco Laboratories, Grand Island, NY. U.S.A.), PMNs were isolated using a Ficoll-Hypaque (Sigma Chemical Co., St Louis, MO, 1J.S.A.) double-density centrifugation method (English and Anderson, 1974). PMNs were then washed with cold Hepes/HBSS and resuspended in Hepes-buffered phenol red-free HBSS (Hepes/HBSS+, pH 7.4, 5.6 mM glucose) to 2-X x 106 cells ml-‘. This method yielded a mixture of PMNs and eosinophils. with some contaminating red blood cells (RBCs), as assessed by
S. F. FREEDMAN
AND
D. L. HATCHELL
Wright’s stain. Cell viability was > 99% by Trypan blue exclusion, Hypotonic lysis was not routinely performed, and RBC contamination averaged 26 y,. Control experiments demonstrated that neither addition of excess RBCs nor hypotonic lysis followed by washing of cells affected 0, production by PMNs in our assays. Cell counts (by phase contrast microscopy J were normalized to obtain equal numbers of leukocytes in each experiment. Eosinophil content of leukocyte preparations was in the normal range for healthy adult cats (Duncan and Prasse. 1988). and averaged 10 -t 6 % for normal cats and I 1 f 4 ‘X, for diabetic cats. One diabetic cat showed persistent hypereosinophilia, with eosinophils constituting 17.73 % of total leukocytes: all data from this cat were excluded from further analysis. For comparison studies, human PMNs were prepared from venous blood of healthy male volunteers (aged 20-3 5 years), using identical methods. RBC contamination averaged 40% in these samples. Eosinophil content was < 3 ‘,${, in these preparations. Cells were maintained on ice prior to assay. and were used within 4 hr of isolation. Cells were manipulated only in plastic vessels.
Measurrvwnt of‘ PMN Suprroxirk Proiluction Superoxide production was measured spectrophotometrically by superoxide dismutase-inhibitable cytochrome c reduction (Cohen and Chovaniec. 1978 ). Assays were performed in duplicate or triplicate at pH 7.4, 37”C, in disposable plastic cuvettes. Standard assays contained 50 &t~ ferricytochrome c (type III. Sigma Chemical Co.) and 0.5 x 10” PMNs ml-’ in 2.5 ml Hepes/HBSS+. After temperature equilibration for 3-5 min in a water bath, absorbance at 550 nm (A& was measured continuously (human cells) or at 5-min intervals (cat cells) for 30 min, in the presence or absence of stimulus. Cuvettes were returned to a 37°C water bath between readings. Control assays were performed in the presence of superoxide dismutase (SOD, 40 pugml-‘), Diagnostic Data, Inc., Mountainview, CA. U.S.A.), which completely inhibited cytochrome c reduction. Phorbol myristate acetate (PMA. Sigma Chemical Co.) was stored as a 1 mg ml-’ stock solution in dimethylsulfoxide at -2O’C and a fresh aliquot was diluted with Hepes/HBSS+ for each day’s assays. Following addition of PMA, cytochrome c reduction (via Ass,,) increased gradually to a limiting linear rate, which was taken to be the maximum rate of 0, production [see Fig. 1(A)]. The lag time was calculated from the intercept of the linear portion of the curve (which persisted for lo-15 min) with the pre-activation baseline (Cohen et al.. 1981). Unstimulated O,- production was estimated in parallel assays. Cytochrome c was not rate-limiting under these assay conditions, and addition of catalase (Boehringer Manheim, Indianapolis, IN. 1J.S.A.) to prevent ac-
SUPEROXIDE
RADICAL
PRODUCTION
IN DIABETES
769
L 0.0
I .o
0.5 Cell
density
Cmtlllon
cells
ml-‘1
FIG. 1. O,- production of cat PMNs stimulated with PMA. A, Increase in A,,, over time following addition of PMA at 1000 ng ml-’ (n), 10 ng ml-’ (@)* 1 ng ml-’ (O), or 0.5 ng ml-l (m). Points shown (corrected for SOD-insensitive changes in A,,,) represent the mean of duplicate determinations performed in a representative experiment on a single cat. Five similar experiments were performed. B, Maximum linear rate of O,- production (0) and corresponding lag (m) are plotted as a function of PMA concentration. Points shown are mean+s.~. of duplicate assays performed in a representative experiment as in (A) above. C, Maximum linear rate of O,- production (0) and lag (m) as a function of cat PMN cell density. Each point represents mean of duplicate determinations as in (A) above.
cumulation of H,O, did not alter measured rates of Ozmproduction. Simultaneous duplicate and triplicate assays exhibited less than a 10% coefficient of variation, but there was considerable day-to-day variation in O,- production by PMNs from a given cat. In all comparisons between PMNs from normal and diabetic cats, each cat is used as the biologic unit of comparison (mean results from two to five different experiments are used for each cat). Unless otherwise stated, values given represent mean f s.D., with n = 4 for diabetic cats, and in = 5 for normal cats. Statistical significance was determined using Student’s unpaired t-test. When indicated, cells were preincubated with additional glucose. insulin, or glucagon for 30 min at 37°C prior to stimulation with PMA (10 ng ml-‘).
Control assays contained 5.6 mM glucose, and lacked insulin and glucagon. Nitroblue Tetrazolium (NBT) reduction NBT reduction was performed by incubating 1 x lo6 PMN ml-’ in Hepes/HBSS+ containing 0~04% NBT and 1 y0 albumin at 3 7°C for O-30 min in the presence or absence of PMA (Cohen et al., 1981). Following termination of the reaction by addition of SOD (20 ,ug ml-l), cells were spun and washed with cold phosphate-buffered saline, resuspended in 5 y0 albumin, then dried and heat-fixed on glass slides. Cells were counterstained with 1 y0 safranin, and 100 consecutive individual cells were examined for formazan deposits.
S. F. FREEDMAN
770
I
TABLE
by PMNs from normal and diabetic cats
Superoxide production
Superoxide production (nmol min’ lo-’ cells) Maximally stimulated
Unstimulated Normal cats (n = 5) Diabetic cats
Lag time (min)
0.12 * om
2,5O+O+kl
6.6kl.O
0.2 1 * 0.0 5
3.32 +03 3*
7.1& 1.4
(n = 4) Values are meansi s.u. * P K 0.02 vs. normal cats.
I
2
3
Normal
4 cats
5
I Normal
I Diabetic
Meons
6
7
8
Diabetic
9 cats
Fro. 2. PMA-stimulated O,- production by PMNs from diabetic and normal cats: maximum rate of O,- production in response to PMA (10 ng ml-‘). Each point (0) represents mean value of duplicate or triplicate determinations. PMNs from each cat were assayed in two to five independent experiments. Group mean + s.n. for normal cats nos 1-5 and diabetic cats nos 6-9 are shown (a): P < 0.02 for the difference between these means.
3. Results Superoxide Production by PMNs from Cats and Humans in Response to PMA Following addition of PMA to a cuvette containing PMNs and cytochrome c. there was an initial time period with little change in A,,,, then a gradual increase in Ajso until a linear rate of change was achieved [Fig. l(A)]. SOD abolished this response. Figures l(A) and (B) show the dose-response relationship between PMA concentration and O,- production. PMNs from both normal and diabetic cats were very sensitive to PMA, demonstrating maximum rates of O,- production in response to as little as 1 ng ml-’ PMA. Lag times shortened as the PMA concentration increased, but did not fall below 5 min even at 1000 ng ml-’ PMA [Fig. l(B)]. The PMA-stimulated rate of O,- production was proportional to the concentration of cells, whereas the lag was indepen-
AND
D L. HATCHELL
dent of this variable up to at least 1 x 10” cells ml-’ [Fig. l(C)]. Using our procedures, human PMNs (from two healthy male volunteers) stimulated with 1000 ng ml-’ PMA produced 6,7+ 3.5 nmol O,- min’ I()-” cells with a lag time of 1.4 & 0.1 min. Superoxide Production by PMNs from Normal ami Diabetic Cats Maximally stimulated (PMA 10 ng ml ’ ) PMNs from diabetic cats generated O,- at significantly higher rates than did those from normals (Table J). Lag times were similar for both groups (Table I). Total 0, generated over 30 min was higher for stimulated PMNs from diabetic than from normal cats (67.0 + 6.5 vs. 5 1.8 + 8.0 nmol O,- lVG cells, respectively, P < 0.05). The differences between PMNs from normal and from diabetic cats, which were evident at 10 ng ml-’ PMA. were not detectable at I ng ml-’ PMA. Thus when PMNs were exposed to 1 ng ml-’ PMA, maximum rates of O,- production by PMNs from diabetic and normal cats were similar (1.84 & 0.88 and 1.9 3 + 0.02 nmol O,- min-’ ZOF cells, respectively), as were the corresponding lag times (14.1 k 3.2 and 1 3.h+ 2.7 min, respectively). There was considerable variability in 0, production among cats within each group, as well as for a given cat on different days (Fig. 2). For example, five serial assays of maximally stimulated PMNs from the same normal cat (no. 3) over a 3-month period. showed a mean O,- production rate of 2.47-_t:O46 nmol 0, min. 1 1o-~li cells with a corresponding lag time of 7.5 1.09 min. Similarly, three assays of stimulated PMNs from the same diabetic cat (no. 8) yielded values of 3.20 & 0.31 nmol O,- min’ 1OF cells and 8.2 i O-5 min, respectively. IJnstimulated PMNs from both normal and diabetic cats showed low levels of O,- production (Table I). Three serial determinations of unstimulated 0, production in a single normal and a single diabetic cat yielded values of 004 k 005 and 022 i 0.1 1 nmol 0, min 1 1OY cells. respectively. When resting 0, production rates were subtracted from total maximally stimulated production rates by PMNs in each respective experiment. PMNs from diabetic cats showed higher net rates of O,- production than did those from normal cats (3.11 kO.37 and 2.44+039 nmol 0, min-’ 1OF cells, respectively, P < 0~05 1. Unstimulated PMNs showed no formazan staining during 30 min of incubation with NBT. In contrast, when PMA ( 10 ng ml-’ ) was added to PMNs in the presence of NBT. 95-97%) of cells were stained with formazan after 15 min. Similar results were obtained with PMNs from normal and those from diabetic cats. Low levels of O,- production by unstimulated PMNs were hence attributed to generalized low-level production rather than to maximal O,- production by a few activated PMNs among many quiescent neighbors.
SUPEROXIDE
RADICAL
PRODUCTION TABLE
IN DIABETES
II
Effect of added glucose. insulin, and glucagon on PMAstimulated rate of superoxide production by cat PMNs
~~ Glucose
22.4 mM 33.6 mM
Insulin
6rnM 400 ng 1 1
Glucagon
Normal cats (“/, of control) ~-~ ~.~~ 97*4 (3) 86k9.t (3) 90222 (3) 98k16 (3)
Diabetic cats (% of control) 92 * 1* (3) 86?c 3-t (3) 97f 16 (3) 93*1 (2)
Values are means i S.D.. with II (number of cats tested) in parentheses * P < 005 vs. control. t 1’ < (Hli for both groups together (n = 6 cats) vs. control.
III Vitro effects 01 Glucose, Insulin, and Glucagon O,- production by PMNs was routinely assayed in the presence of 5.6 mM glucose; however, diabetic cats in our study were consistently hyperglycemic compared with normal cats, We therefore examined the effects of in vitro exposure of PMNs to elevated glucose levels. PMNs were exposed to 56-3 3.6 IIIM glucose for 30 min at 3 7°C prior to addition of PMA (10 ng mll’). As shown in Table II, exposure to 22.4 mM glucose minimally affected PMA-stimulated O,- production by PMNs from both diabetic and normal cats. Exposure to 33.6 mM glucose mildly impaired O,- production by PMNs from both groups of animals. Since cats had been rendered diabetic by partial pancreatectomy. we examined the effects of both insulin and glucagon on O,- production by PMNs. Exposure to insulin at 6 nM or at 6 PM for 30 min did not affect subsequent PMA-stimulated O,- production (Table II). Exposure to 400 ng ll’ glucagon was also without significant effect (Table II). 4. Discussion We have developed a new feline model of diabetes, with the long-term goal of elucidating the pathogenesis of diabetic retinal microvascular disease, specifically the role of PMNs in this process. We began by characterizing the respiratory burst of the PMNs in response to the soluble stimulus PMA. Unlike human PMNs, cat PMNs have been reported to exhibit no chemotaxis towards N-formyl-methionyl-leucylphenylalanine (FMLP) (Gray et al., 1986). We similarly found no PMN respiratory burst in response to FMLP at concentrations up to 1 pM (data not shown). We chose to examine O,- production because it can be conveniently measured in a kinetic mode, and faithfully reflects both oxygen consumption and hydrogen peroxide accumulation which accompany the respiratory burst of PMNs (Makino et al., 1986). Cat PMNs exhibited greater sensitivity to PMA than that reported for human PMNs, and showed longer lag times at saturating PMA concentrations than do human cells (Newburger. Chovaniec and Cohen,
771
1980). When maximally stimulated, cat PMNs produce O,- at rates lower than those reported for human cells (Newburger et al., 1980; Tauber and Babior, 1985). These species-related differences in the PMN respiratory burst are not surprising. since comparable differences in the magnitude of the respiratory burst have been noted for PMNs of other species (Young and Beswick, 1986). The PMNs from our diabetic cats, when stimulated with PMA, demonstrated maximum rates of O,production on average 33 % higher than those of controls. The seemingly remote possibility that this difference derives from trauma experienced at surgery more than 6 months prior to these studies, or was due to daily injections of heterologous insulin in the diabetic animals was not explored. A study of the reversibility of the effect of diabetes on the PMN respiratory burst, by placing diabetic animals in tight glycemic control, would also ultimately be of greater interest. A number of diabetes-related alterations of human PMN function have been identified, changes in the respiratory burst among them. However, the literature reporting these changes is conflicting. Using opsonized zymosan as a stimulus, Baranao et al. (198 7, 1988) found increased luminol-enhanced chemiluminescence as evidence of an increased respiratory burst in PMNs from patients with both insulin-dependent diabetes mellitus (IDDM) and non-insulin-dependent diabetes mellitus (NIDDM) compared with normals. In contrast, using similar stimulus and assay techniques, Wincour et al. (1988) found no difference in the stimulated response of PMNs from normals and subjects with NIDDM. Also using opsonized zymosan (Wierusz-Wysocka et al., 1987) or PMA (Kjersem et al., 1988) to stimulate PMNs, others have found no difference in O,- production between PMNs from patients with IDDM and those of normal controls. Several studies have even shown markedly reduced O,- production by PMNs from patients with IDDM, using either PMA or opsonized zymosan as stimuli (Herskowitz et al., 1990; Shah et al., 1983). The disparate results of these various invesugations cannot easily be reconciled simply by considering differences in patient populations, methods of PMN preparation, or assay methods described. Use of an animal model in our studies eliminated the variable of degree of metabolic control of each patient, because all animals were kept in comparably poor metabolic control throughout the study. To our knowledge, there is a single, recently published study of the PMN respiratory burst in diabetic animals (Mohsenin and Latifpour, 1990). In that study, an increase in total O,production by stimulated PMNs (measured at a single point 20 min after stimulation with high-dose PMA) from diabetic rats (as compared with normal controls) was noted, but did not reach statistical significance. Unstimulated PMNs from our poorly controlled diabetic cats produced more O,- than those from
S. F. FREEDMAN
772
normal cats, but this difference was not statistically significant. The O,- production of unstimulated PMNs from diabetic patients has similarly been reported to moderately exceed that of normal controls (Shah et al., 1983; Wierusz-Wysocka et al., 1987). Exposure of cat PMNs to elevated glucose concentrations, similar to those present in diabetic cat blood (22,4 mM), minimally affected O,- production by stimulated PMNs. The elevated PMN respiratory burst noted in stimulated PMNs from diabetic cats may therefore be due to factors other than that of short term hyperglycemia. While some studies similarly report no effect of elevated glucose on the human PMN respiratory burst in vitro (Pickering et al., 1982; Shah et al.. 1983), others have noted inhibition of the burst with glucose concentrations ranging from 11 mM (Nielson and Hindson. 1989) to 50 mM (Wilson, Tomlinson and Reeves, 1987). The elevated PMN respiratory burst noted in our diabetic cats seems unlikely to be a direct consequence of short term insulin or glucagon deficiency per se, since exposure of diabetic or normal cat PMNs to either of these hormones in vitro did not affect PMAstimulated O,- production. Similar results have been reported in human studies (Shah et al., 1983). We have found enhanced in vitro O,- production by activated PMNs from diabetic cats that have been purposely kept hyperglycemic for at least 6 months. Whether the quantities of superoxide produced by diabetic cat PMNs are sufficient to cause intravascular damage remains unclear: if not O,-, then other toxins produced subsequent to dismutation of O,- (H,O,, HOCl, HO.) may cause tissue injury. Enhancing the potential for microvascular injury by activated PMNs from diabetics is the notion that these PMNs may traverse capillary beds more slowly than PMNs from normal subjects, as suggested by studies of human PMN deformability (Vermes et al., 1987: DeVaro et al., 1988). Our findings are consistent with observations made in the rat model of diabetic retinopathy (Schroder, Palinski and Schmid-Schonbein, 199 1). The percentage of circulating granulocytes and monocytes activated in diabetic rats has been shown to be substantially greater than in control rats. Furthermore, capillary occlusion by granulocytes was documented in perfusion-fixed retinas of only diabetic rats, Interventions aimed at reducing the augmented PMN activation seen in diabetic rats and, now, cats may lead to preventive therapies for diabetic retinopathy. Acknowledgements These studies were supported by Veterans Administration Medical ResearchFunds (Dr Hatchell), National Eye Institute Research Grant EY02903 (Dr Hatchell) and Core Grant EY05722, and a generous gift from Dr and Mrs Harris Vernick.
Dr Freedman
was supported
by Postdoctoral
Research Fellowship Grant PD90036 from the Fight for Sight Research Division of the National Society to Prevent
AND
D. L. HATCHELL
Blindness, awarded in memory of Mary 13.and Alexander P. Hirsch. We are grateful to Professor Irwin Fridovich for providing laboratory space and consultation. and to Dr Gregory Samsa for statistical assistance.
References Armstrong. M.. Jr., Needhamm, D.. Hatchell. U. L. and Nunn, R. S. (1990). Effect of pentoxifylline on the flow of polymorphonuclear leukocytes through a model capillary. Angiology 41, 2 5 3-6 1. Bagge, Ii., Skalak, R. and Atteforo. R. ( 1Y7i). Granulocyte rheology. Experimental studies in an in vitro microflow system. Adv. A4icrocirr. 7, 2948. Baranao. R. I.. Garberi, J. C.. Tesone, P. A. and Rumi. R. S. (1987). Evaluation of neutrophil activity and circulating immune complexes levels in diabetic patients. Hot-m. Metabol. Res. 19. 371-4. Baranao. R. I.. Rumi. L. S.. Tesone, P. A. and Foglia. V. G. (1988). Some characteristics of neutrophils from diabetic patients and their relation to the levels of circulating immune complexes. Acta Diabetol. Lat. 25. 13-23. Bresnick, G. H. (1989). Background diabetic retinopathy. Jn Retina, Vol. II (EdsRyan, S. J.. Schacat, A. P., Murphy, R. P. and Patz, A.). Pp. 327-66. C. V. Mosby : St Louis, MO. [J.S.A. Cohen. H. J. and Chovaniec, M. E. ( 1978). Suoeroxide generation by digitonin-stimulated guinea pip granulocvtes. 1. Clin. Invest. 61. 1081-7. Cohen, H. J.,’ Newburger, P.E.. Chovaniec. M. E.. Whitin. J. C. and Simons, E. R. (1981). Opsonized zymosanstimulated granulocytes-activation and activity of the superoxide-generating system and membrane potential changes. Blood 58, 975-81. Devaro. J., Chumdermpadetsuk, R.. Edelman, K.. Sinclair. S. and MacGregor, R. R. ( 1988). Polymorphonuclear (PMN) filterability in diabetes mellitus. Invest. Ophthalmol. Vis. Sci. 29 (Suppl.), 181. Duncan, J. R. and Prasse. K. W. (1988). Appendix 1. Veterinary Luboratory Medicine. Iowa State IJniv. Press : Ames, Iowa, IJ.S.A. Engler. R. L.. Dahlgren, M. D., Morris. D. D., Peterson. M. A. and Schmid-Schonbein. G. W. ( 1Y86). Role of leukocytes in response to acute myocardial ischemia and reflow in dogs. Am. J. Physiol. 251, H314-H322. English, D. and Andersen. B. R. (1974). Single-step separation of red blood cells, granulocytes and mononuclear leukocytes on discontinuous density gradients of Ficoll-Hypaque. 1. lmmunol. Methods 5. 249-52. Frank, R. N. (1989). Etiologic mechanisms in diabetic retinopathy. In Retina, Vol. II (Eds Ryan. S. J., Schacat. A. P., Murphy, R. P. and Patz, A.). Pp. 301-26. C. V. Mosby : St Louis, MO, U.S.A. Gray, G. D.. Ohlmann, G. M.. Morton, D. K. and Schaub, R. G. ( 1986). Feline polymorphonuclear leukocytes respond chemotactically to leukotriene B, and activated serum but not to F-met-leu-phe. Agents Actions 18. 40-h. Herskowitz, I. C., Matsutani, A. and Permutt. M. A. ( 1990). Leukocyte dysfunction in diabetes is proportional to metabolic control of the diabetic state. Diabetes 39 (Suppl. I ). 27. Kjersem. I-1.. Hilsted. J.. Madsbad. S., Wandall, J. II.. Johansen. S. and Borregaard. N. (1988). Polymorphonuclear leucocyte dysfunction during short term metabolic changes from normo- to hyperglycemia in type I (insulin dependent) diabetic patients. Infection 16, 2 15-20. Kohner. R. H. (1989). Retinal ischemia. In Retina. Vol. II (Eds
SUPEROXIDE
RADICAL
PRODUCTION
IN DIABETES
Ryan, S. J.. Schacat, A. P., Murphy, R. P. and Patz, A.). Pp. 89-97. C. V. Mosby, St Louis, MO, U.S.A. Makino, R.. Tanaka, T., Iizuko. T., Ishimura, Y. and Kanegasake, S. (1986). Stoichiometric conversion of oxygen to superoxide anion during the respiratory burst of neutrophiis. J. Biol. C&m. 261. 11444-47. Mansour, S. Z.. Hatchell. D. L.. Chandler, D., Saloupis, P. and Hatchell. M. C. (1990). Reduction of basement membrane thickening in diabetic cat retina by sulindac. lnvrst. Ophthnlmol. Vis. Sci. 31, 457-63. Markert. M., Andrews, P. C. and Babior, B. M. (1984). Measurement of O,- production by human neutrophils. The preparation and assay of NADPH oxidase-containing particles from human neutrophils. Methods Enzgmol. 105. 358-65. Mohsenin. V. and Latifpour. J. (1990). Respiratory burst in alveolar macrophages of diabetic rats. J. Applied Physiol. 68, 7 384-90. Nees. S. and Schonharting, M. (198 7). Role of different blood cell types in blood rheology. In: Pentoxityllinr and Leukocyte Function (eds Mandell, J. L. and Novack, W. J.). Pp. 105-14, Hoechst-Roussel Pharmaceutied, Inc., Somerville. NJ. Newburger. P. E., Chovaniec, M. E. and Cohen, H. J. (1980). Activity and activation of the granulocyte superoxidegenerating system. Blood 55. 85-92. Nielson, C. I?. and Hindson. D. A. (1989). Inhibition of polymorphonuclear leukocyte respiratory burst by elevated glucose concentrations in vitro. Diabetes 38. 1031-5. Pickering, L. K.. Cleary. T. G. and Getz. S. ( 1982). Effect of glucose concentration and time on polymorphonuclear leukocyte function. J. C/in. Lab. Irnmunol. 8, 31-5. Reiser. H. J., Whitworth. U. G., Hatchell, D. L.. Sutherland. F. S.. Nanda. S.. McAdoo, T. and Hardin, J. R. (1987). Experimental diabetes in cat: partial pancreatectomy alone or combined with local injection of alloxan. Lab. Anirn. Sci. 37, 444-52. Schall, W. D. and Cornelius, I,. M. (1977). Diabetes mellitus. Vet. Clirl. North Am. 7, 613-28. Schmid-Schonbein. G. W., Sung. K. L. P., Tozeren, H..
773 Skalak. R. and Chien, S. (1981). Passive mechanic properties of human leukocytes. Biophys. 1. 36,243-56. Schroder, S.. Palinski, W. and Schmid-Schonbein, G. W. (1991). Activated monocytes and granulocytes, capillary nonperfusion and neovascularization in diabetic retinopathy. Am. 1. Puthol. 139, 81-100. Shah, S. V., Wallin, J. D. and Eilen, S. D. (1983). Chemiluminescence and superoxide anion production by leukocytes from diabetic patients. 1. Clin. Endocrinol. Metab. 57, 402-9. Tauber, A. I. and Babior. B. M. (1985). Neutrophil oxygen reduction: the enzymes and the products. Adv. Free Rad. Biol. Med. 1, 265-307. Vermes, I.. Steinmetz, E. T.. Zeyen, L. J. J. M. and van der Veen, E. A. (1987). Rheologic properties of white blood cells are changed in diabetic patients with microvascular complications. Diabetologia 30. 434-6. Wierusz-Wysocka, B.. Kubicka, M., Wykretowicz, A.. Kasprzak. W. P. & Wysocki, H. (1989). Superoxide anion production and adhesiveness of neutrophils in obese patients. Pol. Arch. Med. Wewn. 82, 2-3, 98-104. Wierusz-Wysocka, B., Wysocki, H., Siekierka. H., Wykretowicz. A., Szczepanik, A. and Klimas, R. (198 7). Evidence of polymorphonuclear neutrophils (PMN) activation in patients with insulin-dependent diabetes mellitus. I. Leuk. Biol. 42. 519-23. Wilson, R. M., Tomlinson. D. R. and Reeves. W. G. ( 198 7). Neutrophil sorbitol production impairs oxidative killing in diabetes. Diabetic Med. 4, 422-34. Wincour, P. H.. Lenton. J.. Puxty. J. A. H. and Anderson. D. C. (1988). Leukocyte microbicidal activity assessed by chemiluminescence in elderly non-insulin dependent diabetes mellitus. Diabetes Res. 9, 73-5. Worthen. S. G., Schwab, B.. III. Elson, E. L. and Downey, G. P. (1989). Mechanics of stimulated neutrophils: cell stiffening induces retention in capillaries. Science 245, 183-6. Young, S. and Beswick, P. (1986). Comparison of the oxidative reactions of neutrophils from a variety of species when stimulated by opsonized zymosan and f-met-leu-phe. 1. Camp. Pathol. 96, 189-96.
