Biochimica et Biophysica Acta, 1055 (1990) 179-185
179
Elsevier BBAMCR 12804
Ferritin-dependent lipid peroxidation by stimulated neutrophils" inhibition by myeloperoxidase-derived hypochlorous acid but not by endogenous lactoferrin Christine C. Winterbourn, Hugo P. Monteiro and Craig F. Galilee Department of Pathology, School of Medicine, Christchurch Hospital, Christchurch (New Zealand)
(Received29 December1989) (Revised manuscript received2 May 1990)
Key words: Lipid peroxidation; Myeloperoxidase;Ferritin; (Human neutrophil)
Human neutrophils stimulated with phorbol myristate acetate or formylmethionylleucylphenylalanine caused superoxide-dependent release of iron from feritin, measured as the formation of a ferrous-ferrozine complex. The stimulated cells also caused ferritin-dependent peroxidation of phospholipid liposomes. Peroxidation was inhibited by lactoferrin, but only at concentrations considerably in excess of what could be achieved by release of endogenous lactoferrin. Peroxidation was enhanced by catalase and methionine, especially when stimulants that release myeloperoxidase were used. Peroxidation was inhibited by added myeloperoxidase. These results are explained by myeloperoxidase catalysing the formation of hypochlorous acid (HOCI) and the HOCI reacting with the lipid to inhibit peroxidation. Thus, neutrophiis are able to use ferritin to promote lipid peroxidation. This may be limited under some conditions by iron binding to iactoferrin or transferrin, and more generally by reactions of the lipid with myeloperoxidase-derived HOCI. However, the latter reactions themselves may be harmful.
Introduction Production of superoxide (O2) and other reactive oxygen species is an essential feature of the microbicidal and cytotoxic actions of neutrophils [1]. Although peroxidation of polyunsaturated lipids is a mechanism by which neutrophils could damage biological targets, relatively few studies of lipid peroxidation by neutrophils have been carried out [2-7]. The main requirements for lipid peroxidation are a transition metal, usually iron, plus a reducing agent such as O 2 [8-11]. Many examples of lipid peroxidation mediated by O 2 and iron have been documented [8,9]. It has recently become clear, however, that neutrophils do not contain significant quantities of endogenous iron that can catalyse either lipid peroxidation or hydroxyl radical production [7,12-14]. Only when exogenous iron is added to the cells has lipid peroxidation been detected [5,7].
Abbreviations: FMLP, N-formylmethionylleucylphenylalanine;PMA, phorbol myristate acetate; BHT, butylated hydroxytoluene; TBA, thiobarbituric acid; TBARs, thlobarbituric acid-reactive substances; DMSO, dimethylsulfoxide. Correspondence: C.C. Winterbourn, Department of Pathology,School of Medicine, Christchurch Hospital, Christchurch, New Zealand.
For neutrophils to promote lipid peroxidation in physiological situations, therefore, they must acquire an exogenous iron catalyst. Iron bound to lactoferrin or transferrin does not promote lipid peroxidation or hydroxyl radical production [15-17], and recent attention has focussed on ferritin as a possible iron source. Biemond et al. [18] were first to demonstrate that neutrophils can release iron from ferritin by an O E-dependent mechanism, and others have shown that 0 2 and other reducing radicals can release ferritin iron [19-21], and thereby promote ferritin-dependent lipid peroxidation [22-25]. Carlin and Arfors [6] reported hpid peroxidation promoted by neutrophils and ferritin. In this paper we examine the efficiency of neutrophil-mediated iron release from ferritin with different neutrophil stimuli, and the requirements for peroxidation of phospholipid liposomes. We show that the cells contain insufficient lactoferrin for its release to influence the catalytic activity of ferritin iron, but when myeloperoxidase is released it modifies the lipid so that peroxidation is inhibited.
Materials and Methods Materials. Ferritin from horse spleen (46% ironsaturated), superoxide dismutase from bovine erythrocytes, catalase from bovine liver (thymol-free), human
0167-4889/90/$03.50 © 1990 ElsevierSciencePublishers B.V. (BiomedicalDivision)
180 lactoferrin and transferrin (both substantially iron-free), butylated hydroxytoluene (BHT), thiobarbituric acid (TBA), phorbol myristate acetate (PMA), N-formylmethionylleucylphenylalanine (FMLP) and cytochalasin B were from Sigma (St Louis, MO). Ficoll was from Pharmacia (Uppsala, Sweden), Hypaque from Sterling Drug Co. (New York, NY), Chelex 100 from Bio-Rad Laboratories (Richmond, CA) and desferrioxamine from Ciba-Geigy (Basle, Switzerland). Other chemicals were from B D H (Poole, U.K.). Myeloperoxidase was purified from h u m a n neutrophils [26]. It had an Aa3o/A28o ratio of 0.67 and its concentration was calculated using ¢430 91000 M - t . c m - 1.
Neutrophils. Neutrophils were isolated from the blood of healthy h u m a n donors by Ficoll-Hypaque centrifugation, dextran sedimentation of red cells and removal of remaining red cells by hypotonic lysis [27]. Neutrophils were suspended in 0.14 M NaC1 containing CaC12 (1 raM), MgC12 (0.5 raM), glucose (1 m g / m l ) and 10 m M phosphate buffer (pH 7.3) (PBS). Reactions were carried out in PBS. Solutions were prepared with distilled deionized water (Milli-Q system) and were treated with Chelex 100 resin (before adding divalent cations). Iron release from ferritin. Solutions containing ferrozine (200 /~M), various concentrations of ferritin and neutrophils ((0-3)- 106/ml) were equilibrated in PBS at 37°C. The cells were stimulated with either 10 # l / m l of P M A (10 /~g/ml in dimethylsulfoxide (DMSO)) or 5 / d / m l of aqueous F M L P ( 2 . 1 0 -5 M) with or without pre-equilibration for 10 min with 1 # l / m l cytochalasin B (5 m g / m l ) . Solutions were then centrifuged in a Minifuge for 1 rain and .4562 of the supernatants was read against a ferrozine plus ferritin blank. Concentrations of FeE+(ferrozine) were calculated using ¢562 27 900 M x . c m - ]. [21]. Lipidperoxidation. Phospholipids were extracted from fresh lamb brains [25]. Multilamellar liposomes were
prepared daily by adding phospholipids (10 m g / m l ) to 0.15 M NaC1 (Chelex-treated) and shaking with glass beads under nitrogen. Liposomes (1 mg phospholipid in a total volume of 1 ml PBS) were incubated at 37°C in air for 30 min with different concentrations of ferritin, neutrophils and other additives as described for each experiment. Neutrophils stimulated with PMA were suspended at 107/ml, PMA (1 /~l/ml of a 100 # g / m l solution in DMSO) was added and the requisite volume of cell suspension was added to the liposome mixture to start the reaction. Neutrophils pretreated with cytochalasin B (1 /~l/ml of a 5 m g / m l solution in DMSO) were incubated at 37°C for 10 min at 107 cells/ml before adding to the liposome mixture. For stimulation with FMLP, 5 #l of aqueous 2 - 10 -5 M F M L P was added to each liposome plus neutrophil mixture. Higher D M S O concentrations were avoided because it was found that > 5 / ~ l / m l affected the yield of lipid peroxidation products. Peroxidation was determined by measuring TBA-reactive substances (TBARs) [25]. E D T A (50/xl of 0.1 M) and B H T (30 #1 of 2% ( w / v ) in ethanol) were added before heating with TBA (0.5 ml) and trichloroacetic acid (1 ml) to prevent further peroxidation due to iron released from ferritin during this step. Solutions were cooled, extracted with butanol (1.25 ml) and centrifuged, then A532 of the butanol phase was measured. Superoxide generation. Rates of 0 2 production by neutrophils treated with the different stimuli were determined by measuring reduction of cytochrome c (reduced-oxidized 21 100 M -1- cm -1) in the presence of 25 /~g/ml catalase. Rates were measured by continuously monitoring .4550 with 50 /~M cytochrome c and 106 cells/ml, or the total amount of 0 2 produced during the course of the reaction was determined from the overall AA550 measured with 150/~M cytochrome c and 0.25 • 106 or 0.5 - 106 cells/ml.
6
B
A
"~
3.0
T ~----_ JT
o "~
2.0
1.0
0.o
0.0
/
oT.....----o
E
5
o ~
4
~
3
I g~
2
=
JO ~
/
O
OJe j
o 1.0
Neutrophlls / ml xlO - 6
2.0
1O0
200 300 [Ferrifin] (~g/rnl)
400
Fig. 1. Iron release from ferritin by PMA-stimulated neutrophils. (A) Effect of increasing cell number; (B) Effect of increasing ferritin concentration. Reaction mixtures all contained 200 #M ferrozine and were incubated at 37°C for 10 rain. In (a) the ferritin concentration was 200 #g/ml and in (b) 2.2-106 neutrophils per ml were used. A562of solutions was measured against a ferrozine plus buffer blank. The blank absorbance was less than 0.006, corresponding to less than 0,2 #M adventitious iron in the buffer. Results are means+ S.D. for three sets of "duplicate assays performed with different neutrophil preparations. Where no error bar is shown, the S.D. fell within the symbol.
181 TABLE I E
Effects of additives on iron release from ferritin by PMA-stimulated neutrophils
|_______.
1 .o
O
The basic reaction mixture contained ferritin (200 ttg/ml), ferrozine (200 ~M) and PMA-stimulated neutrophils (2-106/ml), incubated at 37°C for 10 min. The total superoxide generated in 10 rain, measured separately as cytochrome c reduction, was 150 /tM. Results are the means+ S.D. for triplicate assays using the same neutrophil preparation. Similar results were obtained with other preparations.
v
¢1 r~
0,0-
'
0
'
'
1 2 Neufrophlls / ml x 10 - 6
'
3
Fig. 2. Iron release from ferritin by neutrophils stimulated with F M L P in the absence (o); and presence (O) of cytochalasin B. Conditions were as in Fig. 1A. Points represent means ± S.D. for two sets of duplicates performed with different neutrophil preparations.
Addition
Fe 2+ (ferrozine) Percent formed/10 rain (/xM) of control
None Superoxide dismutase (10/~ g / m l ) Catalase (25 or 60/~g/ml) Myeloperoxidase (30 or 60 nM) Lactoferrin (250 ~ g / m l )
2.77 + 0.49 0.30+0.05 3.68 + 0.25 2.41 + 0.33 2.19 _+0.19
100 11 133 87 81
Results
Iron releasefrom ferritin by stimulated neutrophils Biemond et al. [18] showed that when neutrophils are stimulated with PMA, they cause the reductive release of iron from ferritin. The results in Fig. 1 confirm their findings and show that iron release, measured as the formation of a Fe2+(ferrozine) complex, increases with ferritin concentration and with the number of neutrophils present. N o formation of Fe2+(ferrozine) was observed when ferritin and ferrozine were incubated together, and resting cells released < 10% of the iron seen with PMA-stimulation. Neutrophils stimulated with F M L P in the presence or absence of cytochalasin B also released iron from ferritin (Fig. 2) but in lesser amounts than seen with PMA. This reflects the protracted oxidative burst and decreased 0 2 production that occurs with F M L P compared with the continuous production given by PMA. In the experiments depicted in Figs. 1 and 2, 106 neutrophils produced 85 nmol of 02- in 10 min with PMA, 11 nmol with F M L P and 25 nmol with F M L P
A
plus cytochalasin B. Thus the efficiency of iron release per 0 2 generated was only 0.5-2%. It was higher at the lower rates of 0 2 production and higher ferritin concentrations, but not detectably different for the different stimuli. Under experimentally testable conditions, therefore, neutrophil-mediated iron release from ferritin is not very efficient. However, it is no less efficient than has been observed with 0 2 generated by xanthine oxidase [18,25]. Addition of superoxide dismutase to PMA-stimulated neutrophils suppressed iron release almost completely (Table I). Catalase consistently increased the amount of detectable Fe2+(ferrozine). A similar effect of catalase has been observed with other systems that release iron from ferritin [22,28], and can be attributed to prevention of oxidation of the released iron rather than a direct effect of H202 on iron release [28]. Added myeloperoxidase had little effect on iron release, and a relatively high concentration of lactoferrin was only slightly inhibitory (Table I). This indicates that under
I
O.S
0.3 U3
0.2
/(
o.,
,~J__o ~ ° ~ °
~,~ 0.2
~ ~-
0.1
.........
T
0.0
0.0 0
1 Neufrophlls / rnl x 10 - 6
2
0
100
200 [rerritin] (/zg/ml)
Fig. 3. Lipid peroxidation by PMA-stimulated neutrophils. (A) Effect of increasing cell number in the absence (. . . . . . ) and presence ( ) of ferritin (100 /~g/ml). TBARs were measured as described in the Materials and Methods section. Results are m e a n s + S.D. from three sets of duplicates. Blanks of 0.030 (liposomes only) and 0.085 (liposomes plus ferritin) have been subtracted. Total 02 production/106 cells in 30 min was 200-250 nmol. Additives were: e, ferritin only; A, ferritin+catalase (20 /~g/ml); II, ferritin+methionine (1 mM); o , no ferritin; zx, no ferritin+catalase; 1:3, no ferritin+ methionine. (B) Effect of increasing ferritin concentration. Results are means of duplicates from a typical experiment. No blanks have been subtracted. Incubation mixtures contained: e , 2-106 neutropliils/ml; &, 2.106 n e u t r o p h i l s / m l + 20 / t g / m l catalase; ©, no neutrophils. Other conditions are as in Fig. 1A.
182 the conditions of the assay, lactoferrin did not compete efficiently with ferrozine for the released iron. Lactoferrin is released from neutrophils by PMA and by FMLP in the presence of cytochalasin B. However, the maximum of 3-5 /tg per 106 cells [29] should be much too low to affect the detectable iron.
I 00 ( P~o ~o
-
~
75
o
_6
5o
25
Lipid peroxidation by neutrophils in the presence of ferritin Phospholipid liposomes, when incubated with ferritin and neutrophils stimulated with PMA, underwent lipid peroxidation, monitored by the formation of TBARs, that increased with increasing cell number (Fig. 3A). Peroxidation was several-fold less in the absence of ferritin (Fig. 3A), and increased with increasing ferritin concentration (Fig. 3B). Ferritin alone did not cause significant lipid peroxidation. It increased the TBARs blank (Fig. 3B) but this was due to reactions proceeding during the heating step of the analysis, since ferritin added immediately prior to heating had the same effect as when it was present during the 37°C incubation. The lipid peroxidation seen in the absence of ferritin was inhibited by 100/~M desferrioxamine (not shown), indicating that it was dependent on iron. This was probably present adventitiously, as neutrophils appear not to contain significant levels of low-molecular-weight iron [7,12-14]. Even though the contamination of the buffer was < 0.2/xM, this is probably sufficient to account for the ferritin-independent peroxidation. Catalase enhanced lipid peroxidation. Enhancement was greater at the higher cell concentrations (Fig. 3A), but was relatively independent of ferritin concentration (Fig. 3B). The effect of catalase was mimicked by methionine (Fig. 3A). Methionine is a good scavenger of hypochlorous acid (HOCI) [30], which is produced in the myeloperoxidase-catalysed reaction between H202 and CI-. PMA releases little myeloperoxidase from neutrophils, but enough to convert some of the H202 they produce to HOC1 [31,32]. Catalase prevents this conversion. The results suggest, therefore, that catalase and methionine both acted by influencing reactions of myeloperoxidase-derived HOCI. The effects of other additives on lipid peroxidation by PMA-stimulated neutrophils are shown in Table II. Desferrioxamine inhibited the reaction, indicating a requirement for iron release from the ferritin, as observed in other ferritin-dependent systems [24,25]. Superoxide dismutase inhibited almost totally. Addition of myeloperoxidase, which enhances HOC1 formation [32], inhibited peroxidation, with maximal effect at 30 nM. Methionine prevented inhibition by myeloperoxidase and enhanced lipid peroxidation to the level seen with methionine or catalase alone. Azide, an inhibitor of myeloperoxidase, eliminated its effect, and also slightly enhanced peroxidation by the cells alone.
o
0 0
IO0
200
300
400
500
[Laefoferrin] /.zg/rnl Fig. 4. Effect of exogenous apolactoferrin on lipid peroxidation by PMA-stimulated neutrophils. Reaction mixtures contained 106 neutrophils and 100/~g ferritin per ml. T B A R s were measured as in the Materials and Methods section. Ferritin blanks have been subtracted. Results are means _ S.D. of triplicate assays from two experiments.
Apolactoferrin added to neutrophils stimulated with PMA gave concentration-dependent inhibition of lipid peroxidation (Fig. 4). Inhibition by apotransferrin gave a similar concentration dependence (not shown). The molar concentration of lactoferrin required to inhibit peroxidation (approx 6/~M for 50% inhibition) is of the same order as the concentration of iron released from the ferritin in 30 min. It is sufficiently high to imply that inhibition by released neutrophil lactoferrin (a maximum of 0.1 /~M at 1 0 6 cells/ml [29]) would be negligible under these conditions. If endogenous lactoferrin did inhibit peroxidation, a lag until its binding capacity was exceeded might be expected. No lag was apparent (Fig. 5). However, with 150 # g / m l of apolactoferrin added, inhibition decreased progressively from 26% at 5 rain to 11% at 30 rain, demonstrating
T A B L E II
Effects of oxidant scaoengers and inhibitors on ferritin-dependent lipid peroxidation by PMA-stimulated neutrophils PMA-stimulated neutrophils (2-106), ferritin (100/tg) and liposomes (1 mg) were incubated in 1 ml PBS at 37°C for 30 min before analysing for TBARs. Results are means :t: S.D. of three sets of triplicate assays in which the control A532 values ranged between 0.20 and 0.26 (after subtraction of a liposome and ferritin blank of 0.07). The total production of superoxide during the 30 min was approx 400 nmol. Addition
T B A R s (% of control)
None Desferrioxamine (100 ~M) Superoxide dismutase (10 t~g/ml) Myeloperoxidase (30 nM) Myeloperoxidase (60 nM) Myeloperoxidase + methionine (1 m M ) Methionine Catalase (20 t t g / m l ) Myeloperoxidase + azide (0.5 m M ) Azide
100 9+ 2 5 +_ 5 53 + 10 57 + 10 170 + 24 168 _+17 169 + 15 115 + 20 110+10
183 T A B L E III
~ ~ o
Effects of additives on peroxidation of liposomes by hypoxanthine, xanthine oxidase and FeSO4
02
~o
/o
tO
o/
Liposomes (1 mg) were incubated in PBS (1 ml) with FeSO4 (2 /zM), hypoxanthine (50 ~M), xanthine oxidase (0.007 U / m l ) and additives for 30 rain at 37°C before analysing for TBARs. The control gave an A532 of 0.17 in excess of the liposome blank. Results are m e a n s + S.D. of 4 - 6 assays.
~-=
179
Elsevier BBAMCR 12804
Ferritin-dependent lipid peroxidation by stimulated neutrophils" inhibition by myeloperoxidase-derived hypochlorous acid but not by endogenous lactoferrin Christine C. Winterbourn, Hugo P. Monteiro and Craig F. Galilee Department of Pathology, School of Medicine, Christchurch Hospital, Christchurch (New Zealand)
(Received29 December1989) (Revised manuscript received2 May 1990)
Key words: Lipid peroxidation; Myeloperoxidase;Ferritin; (Human neutrophil)
Human neutrophils stimulated with phorbol myristate acetate or formylmethionylleucylphenylalanine caused superoxide-dependent release of iron from feritin, measured as the formation of a ferrous-ferrozine complex. The stimulated cells also caused ferritin-dependent peroxidation of phospholipid liposomes. Peroxidation was inhibited by lactoferrin, but only at concentrations considerably in excess of what could be achieved by release of endogenous lactoferrin. Peroxidation was enhanced by catalase and methionine, especially when stimulants that release myeloperoxidase were used. Peroxidation was inhibited by added myeloperoxidase. These results are explained by myeloperoxidase catalysing the formation of hypochlorous acid (HOCI) and the HOCI reacting with the lipid to inhibit peroxidation. Thus, neutrophiis are able to use ferritin to promote lipid peroxidation. This may be limited under some conditions by iron binding to iactoferrin or transferrin, and more generally by reactions of the lipid with myeloperoxidase-derived HOCI. However, the latter reactions themselves may be harmful.
Introduction Production of superoxide (O2) and other reactive oxygen species is an essential feature of the microbicidal and cytotoxic actions of neutrophils [1]. Although peroxidation of polyunsaturated lipids is a mechanism by which neutrophils could damage biological targets, relatively few studies of lipid peroxidation by neutrophils have been carried out [2-7]. The main requirements for lipid peroxidation are a transition metal, usually iron, plus a reducing agent such as O 2 [8-11]. Many examples of lipid peroxidation mediated by O 2 and iron have been documented [8,9]. It has recently become clear, however, that neutrophils do not contain significant quantities of endogenous iron that can catalyse either lipid peroxidation or hydroxyl radical production [7,12-14]. Only when exogenous iron is added to the cells has lipid peroxidation been detected [5,7].
Abbreviations: FMLP, N-formylmethionylleucylphenylalanine;PMA, phorbol myristate acetate; BHT, butylated hydroxytoluene; TBA, thiobarbituric acid; TBARs, thlobarbituric acid-reactive substances; DMSO, dimethylsulfoxide. Correspondence: C.C. Winterbourn, Department of Pathology,School of Medicine, Christchurch Hospital, Christchurch, New Zealand.
For neutrophils to promote lipid peroxidation in physiological situations, therefore, they must acquire an exogenous iron catalyst. Iron bound to lactoferrin or transferrin does not promote lipid peroxidation or hydroxyl radical production [15-17], and recent attention has focussed on ferritin as a possible iron source. Biemond et al. [18] were first to demonstrate that neutrophils can release iron from ferritin by an O E-dependent mechanism, and others have shown that 0 2 and other reducing radicals can release ferritin iron [19-21], and thereby promote ferritin-dependent lipid peroxidation [22-25]. Carlin and Arfors [6] reported hpid peroxidation promoted by neutrophils and ferritin. In this paper we examine the efficiency of neutrophil-mediated iron release from ferritin with different neutrophil stimuli, and the requirements for peroxidation of phospholipid liposomes. We show that the cells contain insufficient lactoferrin for its release to influence the catalytic activity of ferritin iron, but when myeloperoxidase is released it modifies the lipid so that peroxidation is inhibited.
Materials and Methods Materials. Ferritin from horse spleen (46% ironsaturated), superoxide dismutase from bovine erythrocytes, catalase from bovine liver (thymol-free), human
0167-4889/90/$03.50 © 1990 ElsevierSciencePublishers B.V. (BiomedicalDivision)
180 lactoferrin and transferrin (both substantially iron-free), butylated hydroxytoluene (BHT), thiobarbituric acid (TBA), phorbol myristate acetate (PMA), N-formylmethionylleucylphenylalanine (FMLP) and cytochalasin B were from Sigma (St Louis, MO). Ficoll was from Pharmacia (Uppsala, Sweden), Hypaque from Sterling Drug Co. (New York, NY), Chelex 100 from Bio-Rad Laboratories (Richmond, CA) and desferrioxamine from Ciba-Geigy (Basle, Switzerland). Other chemicals were from B D H (Poole, U.K.). Myeloperoxidase was purified from h u m a n neutrophils [26]. It had an Aa3o/A28o ratio of 0.67 and its concentration was calculated using ¢430 91000 M - t . c m - 1.
Neutrophils. Neutrophils were isolated from the blood of healthy h u m a n donors by Ficoll-Hypaque centrifugation, dextran sedimentation of red cells and removal of remaining red cells by hypotonic lysis [27]. Neutrophils were suspended in 0.14 M NaC1 containing CaC12 (1 raM), MgC12 (0.5 raM), glucose (1 m g / m l ) and 10 m M phosphate buffer (pH 7.3) (PBS). Reactions were carried out in PBS. Solutions were prepared with distilled deionized water (Milli-Q system) and were treated with Chelex 100 resin (before adding divalent cations). Iron release from ferritin. Solutions containing ferrozine (200 /~M), various concentrations of ferritin and neutrophils ((0-3)- 106/ml) were equilibrated in PBS at 37°C. The cells were stimulated with either 10 # l / m l of P M A (10 /~g/ml in dimethylsulfoxide (DMSO)) or 5 / d / m l of aqueous F M L P ( 2 . 1 0 -5 M) with or without pre-equilibration for 10 min with 1 # l / m l cytochalasin B (5 m g / m l ) . Solutions were then centrifuged in a Minifuge for 1 rain and .4562 of the supernatants was read against a ferrozine plus ferritin blank. Concentrations of FeE+(ferrozine) were calculated using ¢562 27 900 M x . c m - ]. [21]. Lipidperoxidation. Phospholipids were extracted from fresh lamb brains [25]. Multilamellar liposomes were
prepared daily by adding phospholipids (10 m g / m l ) to 0.15 M NaC1 (Chelex-treated) and shaking with glass beads under nitrogen. Liposomes (1 mg phospholipid in a total volume of 1 ml PBS) were incubated at 37°C in air for 30 min with different concentrations of ferritin, neutrophils and other additives as described for each experiment. Neutrophils stimulated with PMA were suspended at 107/ml, PMA (1 /~l/ml of a 100 # g / m l solution in DMSO) was added and the requisite volume of cell suspension was added to the liposome mixture to start the reaction. Neutrophils pretreated with cytochalasin B (1 /~l/ml of a 5 m g / m l solution in DMSO) were incubated at 37°C for 10 min at 107 cells/ml before adding to the liposome mixture. For stimulation with FMLP, 5 #l of aqueous 2 - 10 -5 M F M L P was added to each liposome plus neutrophil mixture. Higher D M S O concentrations were avoided because it was found that > 5 / ~ l / m l affected the yield of lipid peroxidation products. Peroxidation was determined by measuring TBA-reactive substances (TBARs) [25]. E D T A (50/xl of 0.1 M) and B H T (30 #1 of 2% ( w / v ) in ethanol) were added before heating with TBA (0.5 ml) and trichloroacetic acid (1 ml) to prevent further peroxidation due to iron released from ferritin during this step. Solutions were cooled, extracted with butanol (1.25 ml) and centrifuged, then A532 of the butanol phase was measured. Superoxide generation. Rates of 0 2 production by neutrophils treated with the different stimuli were determined by measuring reduction of cytochrome c (reduced-oxidized 21 100 M -1- cm -1) in the presence of 25 /~g/ml catalase. Rates were measured by continuously monitoring .4550 with 50 /~M cytochrome c and 106 cells/ml, or the total amount of 0 2 produced during the course of the reaction was determined from the overall AA550 measured with 150/~M cytochrome c and 0.25 • 106 or 0.5 - 106 cells/ml.
6
B
A
"~
3.0
T ~----_ JT
o "~
2.0
1.0
0.o
0.0
/
oT.....----o
E
5
o ~
4
~
3
I g~
2
=
JO ~
/
O
OJe j
o 1.0
Neutrophlls / ml xlO - 6
2.0
1O0
200 300 [Ferrifin] (~g/rnl)
400
Fig. 1. Iron release from ferritin by PMA-stimulated neutrophils. (A) Effect of increasing cell number; (B) Effect of increasing ferritin concentration. Reaction mixtures all contained 200 #M ferrozine and were incubated at 37°C for 10 rain. In (a) the ferritin concentration was 200 #g/ml and in (b) 2.2-106 neutrophils per ml were used. A562of solutions was measured against a ferrozine plus buffer blank. The blank absorbance was less than 0.006, corresponding to less than 0,2 #M adventitious iron in the buffer. Results are means+ S.D. for three sets of "duplicate assays performed with different neutrophil preparations. Where no error bar is shown, the S.D. fell within the symbol.
181 TABLE I E
Effects of additives on iron release from ferritin by PMA-stimulated neutrophils
|_______.
1 .o
O
The basic reaction mixture contained ferritin (200 ttg/ml), ferrozine (200 ~M) and PMA-stimulated neutrophils (2-106/ml), incubated at 37°C for 10 min. The total superoxide generated in 10 rain, measured separately as cytochrome c reduction, was 150 /tM. Results are the means+ S.D. for triplicate assays using the same neutrophil preparation. Similar results were obtained with other preparations.
v
¢1 r~
0,0-
'
0
'
'
1 2 Neufrophlls / ml x 10 - 6
'
3
Fig. 2. Iron release from ferritin by neutrophils stimulated with F M L P in the absence (o); and presence (O) of cytochalasin B. Conditions were as in Fig. 1A. Points represent means ± S.D. for two sets of duplicates performed with different neutrophil preparations.
Addition
Fe 2+ (ferrozine) Percent formed/10 rain (/xM) of control
None Superoxide dismutase (10/~ g / m l ) Catalase (25 or 60/~g/ml) Myeloperoxidase (30 or 60 nM) Lactoferrin (250 ~ g / m l )
2.77 + 0.49 0.30+0.05 3.68 + 0.25 2.41 + 0.33 2.19 _+0.19
100 11 133 87 81
Results
Iron releasefrom ferritin by stimulated neutrophils Biemond et al. [18] showed that when neutrophils are stimulated with PMA, they cause the reductive release of iron from ferritin. The results in Fig. 1 confirm their findings and show that iron release, measured as the formation of a Fe2+(ferrozine) complex, increases with ferritin concentration and with the number of neutrophils present. N o formation of Fe2+(ferrozine) was observed when ferritin and ferrozine were incubated together, and resting cells released < 10% of the iron seen with PMA-stimulation. Neutrophils stimulated with F M L P in the presence or absence of cytochalasin B also released iron from ferritin (Fig. 2) but in lesser amounts than seen with PMA. This reflects the protracted oxidative burst and decreased 0 2 production that occurs with F M L P compared with the continuous production given by PMA. In the experiments depicted in Figs. 1 and 2, 106 neutrophils produced 85 nmol of 02- in 10 min with PMA, 11 nmol with F M L P and 25 nmol with F M L P
A
plus cytochalasin B. Thus the efficiency of iron release per 0 2 generated was only 0.5-2%. It was higher at the lower rates of 0 2 production and higher ferritin concentrations, but not detectably different for the different stimuli. Under experimentally testable conditions, therefore, neutrophil-mediated iron release from ferritin is not very efficient. However, it is no less efficient than has been observed with 0 2 generated by xanthine oxidase [18,25]. Addition of superoxide dismutase to PMA-stimulated neutrophils suppressed iron release almost completely (Table I). Catalase consistently increased the amount of detectable Fe2+(ferrozine). A similar effect of catalase has been observed with other systems that release iron from ferritin [22,28], and can be attributed to prevention of oxidation of the released iron rather than a direct effect of H202 on iron release [28]. Added myeloperoxidase had little effect on iron release, and a relatively high concentration of lactoferrin was only slightly inhibitory (Table I). This indicates that under
I
O.S
0.3 U3
0.2
/(
o.,
,~J__o ~ ° ~ °
~,~ 0.2
~ ~-
0.1
.........
T
0.0
0.0 0
1 Neufrophlls / rnl x 10 - 6
2
0
100
200 [rerritin] (/zg/ml)
Fig. 3. Lipid peroxidation by PMA-stimulated neutrophils. (A) Effect of increasing cell number in the absence (. . . . . . ) and presence ( ) of ferritin (100 /~g/ml). TBARs were measured as described in the Materials and Methods section. Results are m e a n s + S.D. from three sets of duplicates. Blanks of 0.030 (liposomes only) and 0.085 (liposomes plus ferritin) have been subtracted. Total 02 production/106 cells in 30 min was 200-250 nmol. Additives were: e, ferritin only; A, ferritin+catalase (20 /~g/ml); II, ferritin+methionine (1 mM); o , no ferritin; zx, no ferritin+catalase; 1:3, no ferritin+ methionine. (B) Effect of increasing ferritin concentration. Results are means of duplicates from a typical experiment. No blanks have been subtracted. Incubation mixtures contained: e , 2-106 neutropliils/ml; &, 2.106 n e u t r o p h i l s / m l + 20 / t g / m l catalase; ©, no neutrophils. Other conditions are as in Fig. 1A.
182 the conditions of the assay, lactoferrin did not compete efficiently with ferrozine for the released iron. Lactoferrin is released from neutrophils by PMA and by FMLP in the presence of cytochalasin B. However, the maximum of 3-5 /tg per 106 cells [29] should be much too low to affect the detectable iron.
I 00 ( P~o ~o
-
~
75
o
_6
5o
25
Lipid peroxidation by neutrophils in the presence of ferritin Phospholipid liposomes, when incubated with ferritin and neutrophils stimulated with PMA, underwent lipid peroxidation, monitored by the formation of TBARs, that increased with increasing cell number (Fig. 3A). Peroxidation was several-fold less in the absence of ferritin (Fig. 3A), and increased with increasing ferritin concentration (Fig. 3B). Ferritin alone did not cause significant lipid peroxidation. It increased the TBARs blank (Fig. 3B) but this was due to reactions proceeding during the heating step of the analysis, since ferritin added immediately prior to heating had the same effect as when it was present during the 37°C incubation. The lipid peroxidation seen in the absence of ferritin was inhibited by 100/~M desferrioxamine (not shown), indicating that it was dependent on iron. This was probably present adventitiously, as neutrophils appear not to contain significant levels of low-molecular-weight iron [7,12-14]. Even though the contamination of the buffer was < 0.2/xM, this is probably sufficient to account for the ferritin-independent peroxidation. Catalase enhanced lipid peroxidation. Enhancement was greater at the higher cell concentrations (Fig. 3A), but was relatively independent of ferritin concentration (Fig. 3B). The effect of catalase was mimicked by methionine (Fig. 3A). Methionine is a good scavenger of hypochlorous acid (HOCI) [30], which is produced in the myeloperoxidase-catalysed reaction between H202 and CI-. PMA releases little myeloperoxidase from neutrophils, but enough to convert some of the H202 they produce to HOC1 [31,32]. Catalase prevents this conversion. The results suggest, therefore, that catalase and methionine both acted by influencing reactions of myeloperoxidase-derived HOCI. The effects of other additives on lipid peroxidation by PMA-stimulated neutrophils are shown in Table II. Desferrioxamine inhibited the reaction, indicating a requirement for iron release from the ferritin, as observed in other ferritin-dependent systems [24,25]. Superoxide dismutase inhibited almost totally. Addition of myeloperoxidase, which enhances HOC1 formation [32], inhibited peroxidation, with maximal effect at 30 nM. Methionine prevented inhibition by myeloperoxidase and enhanced lipid peroxidation to the level seen with methionine or catalase alone. Azide, an inhibitor of myeloperoxidase, eliminated its effect, and also slightly enhanced peroxidation by the cells alone.
o
0 0
IO0
200
300
400
500
[Laefoferrin] /.zg/rnl Fig. 4. Effect of exogenous apolactoferrin on lipid peroxidation by PMA-stimulated neutrophils. Reaction mixtures contained 106 neutrophils and 100/~g ferritin per ml. T B A R s were measured as in the Materials and Methods section. Ferritin blanks have been subtracted. Results are means _ S.D. of triplicate assays from two experiments.
Apolactoferrin added to neutrophils stimulated with PMA gave concentration-dependent inhibition of lipid peroxidation (Fig. 4). Inhibition by apotransferrin gave a similar concentration dependence (not shown). The molar concentration of lactoferrin required to inhibit peroxidation (approx 6/~M for 50% inhibition) is of the same order as the concentration of iron released from the ferritin in 30 min. It is sufficiently high to imply that inhibition by released neutrophil lactoferrin (a maximum of 0.1 /~M at 1 0 6 cells/ml [29]) would be negligible under these conditions. If endogenous lactoferrin did inhibit peroxidation, a lag until its binding capacity was exceeded might be expected. No lag was apparent (Fig. 5). However, with 150 # g / m l of apolactoferrin added, inhibition decreased progressively from 26% at 5 rain to 11% at 30 rain, demonstrating
T A B L E II
Effects of oxidant scaoengers and inhibitors on ferritin-dependent lipid peroxidation by PMA-stimulated neutrophils PMA-stimulated neutrophils (2-106), ferritin (100/tg) and liposomes (1 mg) were incubated in 1 ml PBS at 37°C for 30 min before analysing for TBARs. Results are means :t: S.D. of three sets of triplicate assays in which the control A532 values ranged between 0.20 and 0.26 (after subtraction of a liposome and ferritin blank of 0.07). The total production of superoxide during the 30 min was approx 400 nmol. Addition
T B A R s (% of control)
None Desferrioxamine (100 ~M) Superoxide dismutase (10 t~g/ml) Myeloperoxidase (30 nM) Myeloperoxidase (60 nM) Myeloperoxidase + methionine (1 m M ) Methionine Catalase (20 t t g / m l ) Myeloperoxidase + azide (0.5 m M ) Azide
100 9+ 2 5 +_ 5 53 + 10 57 + 10 170 + 24 168 _+17 169 + 15 115 + 20 110+10
183 T A B L E III
~ ~ o
Effects of additives on peroxidation of liposomes by hypoxanthine, xanthine oxidase and FeSO4
02
~o
/o
tO
o/
Liposomes (1 mg) were incubated in PBS (1 ml) with FeSO4 (2 /zM), hypoxanthine (50 ~M), xanthine oxidase (0.007 U / m l ) and additives for 30 rain at 37°C before analysing for TBARs. The control gave an A532 of 0.17 in excess of the liposome blank. Results are m e a n s + S.D. of 4 - 6 assays.
~-=
