Clin. exp. Immunol. (1991) 84, 347-352
AD)ONIS 00099104910015 1J
Identification of a myeloperoxidase inhibitor from normal human serum C. M. YEA, B. DULARAY & C. J. ELSON Department of Pathology, University of Bristol, Bristol, England
(Acceptedfor publication 19 November 1990)
SUMMARY An inhibitor of myeloperoxidase (MPO) has been identified in normal human serum. Initial experiments confirmed that high levels of MPO inhibitory activity are present in sera and that the inhibitor did not act by interfering with the assay. Purification of the inhibitor activity by salt precipitation followed by ion exchange and affinity chromatography revealed the presence of a protein of 150 kD. The purified inhibitory activity displayed dose and time dependency and was not associated with IgG or IgA. It is considered that human serum contains an inhibitor of extracellular MPO capable of protecting against hypohalous acid release in host tissues and that if inhibitor levels are
reduced such protection
may
fail.
Keywords myeloperoxidase inhibitor human
serum
INTRODUCTION
Neutrophils play a key role in defence of the host against microbial attack. They execute this function through two major pathways of killing: dependent and independent on oxygen. After phagocytosis of an invading microbe into an enclosed phagolysosome within the cell, activation of a plasma-membrane-bound enzyme system (NADPH oxidase) occurs and superoxide and other derived radicals are released into the vacuole (Babior, 1978). In conjunction, granules located in the cytoplasm of the neutrophil fuse with the phagolysosome membrane and empty their contents into it, releasing an array of degradative enzymes. One of the major enzymes released during the latter, degranulative process, is myeloperoxidase (MPO) which combines halide ions (usually Cl ) with hydrogen peroxide forming the highly microbicidal hypohalous acid (Klebanoff, 1975). Together these processes result in the death of engulfed bacteria. Release of MPO and other granule enzymes and indeed of superoxide, is initiated before the phagolysosome has completely sealed off (Bagglioni & Dewald, 1985) resulting in the release of granule enzymes into the surrounding milieu. In addition, enzymes are released directly from granules if neutrophils react with surfaces too extensive for engulfment (Henson, 1971; Henson, Johnson & Spiegelberg, 1972). It has been suggested that such extracellular release of granule enzymes occurs in a number of inflammatory diseases including rheumatoid arthritis (RA) (Weismann, 1972). In this condition, high numbers of neutrophils are found in synovial fluids (Palmer, 1968) together with immunoglobulin aggregates capable of stimulating a degranulation response (Morrison,
Pruzanski & Radadive, 1978; Dularay, Dieppe & Elson, 1990) and granule enzymes including MPO (Hadler, Spitznagel & Quinet, 1979; Dularay et al., 1990a). The contribution of these enzymes to the pathogenesis of RA remains unclear since RA synovial fluids also contain inhibitors for many granule enzymes. However, to our knowledge an inhibitor of MPO has not been described. Recently, during an investigation of the degranulation response of RA synovial neutrophils, we demonstrated that RA synovial fluid and sera and normal sera reduced the MPO activity of degranulating neutrophils (Dularay et al., 1990a). These results led us to suggest that the fluids contained a MPO inhibitor. Work reported elsewhere (Dularay, Yea & Elson, 1991) gives credence to this idea by showing that the fluids inhibited the released enzyme rather than the degranulation response of polymorphonuclear leucocytes and also that incubation of synovial fluids with enzyme substrate did not inhibit MPO activity. Moreover, as glucuronidase activity was unaffected by the presence of the fluids, the inhibitor appeared to be peroxidase specific. The purpose of the current work was to determine whether a particular component of the fluids could be identified which exhibited inhibitory activity. MATERIALS AND METHODS Patients Blood and synovial fluids were obtained from RA patients
attending the Rheumatology Outpatients Clinic at the Bristol Royal Infirmary. All patients fulfilled the standard criteria for classical or definite RA. Assay of MPO activity MPO was assayed using a modification of the method of Suzuki et al. (1983) utilizing 3,3',5,5' tetramethyl benzidine (TMBD) in
Correspondence: Dr C. J. Elson, Department of Pathology, University of Bristol, Bristol BS8 ITD, UK.
347
348
C. M. Yea, B. Dularay & C. J. Elson
96-well microtitre plates. MPO was purified as described below and 50 ng (determined spectrophotometrically using a reducedoxidized extinction coefficient of 75 mm- ' at 472 nm; Bos, Wever & Roos, 1978) were added to each well in 20 kl of 0 I M sodium phosphate, pH 7-4. After addition of 30 pl substrate to each well (3-7 mM TMBD, 0-002% H202) in 1:3 mixture of 0-01 M sodium acetate, pH 5-2, and PBS, the plate was incubated at room temperature for 5 min and the absorbance at 690-540 nm was read in a Labsystem automatic plate reader. Blanks contained equal volumes with no MPO. The change of absorbance at 690540 nm was measured. A unit of MPO is the amount causing a change of 1 in 5 min.
Purification of MPO Human polymorphonuclear leucocytes were prepared from buffy coats obtained from the South West Regional Blood Transfusion Centre (Bristol) as described previously (Henderson, Chappell & Jones, 1987). They were resuspended in PBS, sonicated on ice (3 x 1O sec, 50 W) and the membranes prepared by centrifugation (100 000 g, 30 min). After pre-extraction with 0-25% Lubrol PX/0-25% sodium deoxycholate in 10 mm glycine (pH 8 0), they were extracted twice with 0 5% cetyltrimethyl ammonium bromide in 0 I M phosphate buffer, pH 7-4. The supernatant was collected and loaded onto a 12 cm x 12 mm column of S-Sepharose (Pharmacia) pre-equilibrated with extraction buffer (without detergent). After extensive washing the column was eluted with a gradient of 0 2-0 5 M sodium phosphate, pH 7-4, and 3-ml fractions collected. Those displaying absorbance at 340 nm > 0-05 were pooled. Preparations such as these typically had a R, of 0-62-0-68 and displayed one major band (approximately 65 kD) on Coomassie blue stained SDS-PAGE. SDS-PAGE was performed by the discontinuous buffer method of Laemmli (1970) utilizing a 12% separating gel. In the initial experiments supernatants from stimulated neutrophils were used as a source of MPO (MPO-S). Briefly, neutrophils were pretreated with 20 pg/ml cytochalasin B (Sigma; I mg/ml in DMSO) for 10 min at room temperature. The treated cells (2 x 108/ml) were incubated in HBSS containing 10 mg/ml human serum albumin (HSA-HBSS) and formylmethionyl-leucyl-phenylalanine (2 5 x 106 M) for 30 min at 37°C. In our hands these conditions were optimal for maximum MPO release (Dularay, 1989; Dularay et al., 1990b). Inhibition studies The amount of MPO-inhibitory activity present in samples was established using the MPO assay described above. Samples were added to the MPO in the plate (controls contained equal volumes of phosphate buffer) and incubated for 30 min at room temperature prior to addition of substrate, unless otherwise indicated. Amount of inhibitor is expressed such that 1 unit of inhibitor inhibits I unit of MPO.
Purification of MPO inhibitor Pooled normal human serum was prepared from blood collected from healthy volunteers. It was stored at - 70°C without preservatives until used. All procedures were carried out in 0 1 M sodium phosphate buffer (pH 7.4) unless otherwise stated. The serum was centrifuged to remove all particulate matter (100 000 g, I h). The supernatant was collected, made up to 20% ammonium sulphate by the addition of cold saturated ammonium sulphate, stirred for 4 h at 4°C and centrifuged (100 000 g,
*°r-
0-75I 0-5 _IN0co IC:
0
0-251-
z
O
Fig. 1. Purification of the MPO inhibitor by FPLC anion exchange. The ammonium sulphate purified fraction (6 ml) was applied to the Mono Q column and eluted using a salt gradient as shown (--- ). The A280 was ) and 1-ml fractions were collected and continuously monitored ( assayed for inhibitory activity as described in Materials and Methods.
30 min). The supernatant was collected and made up to 50% saturation ammonium sulphate saturation by the further addition of cold ammonium sulphate. After 4-h stirring at 4°C the mixture was centrifuged (1l0 000 g, 30 min) and the supernatant discarded. The precipitate was redissolved in phosphate buffer, centrifuged to remove insoluble material and dialysed overnight against two changes of 10 volumes of phosphate buffer. Approximately 10 mg of the partially purified protein (estimated by absorbance at 280 nm) were loaded onto a 1 ml Mono Q anion exchange column (pre-equlibrated with phosphate buffer) on a Pharmacia FPLC system at a flow rate of 0 1 ml/min and washed with a further 10-column volumes of buffer. Bound proteins were eluted with a non-linear gradient of NaCl (see Fig. 1) were from 0 to 0-5 M. Fractions of I ml collected and assayed for MPO-inhibitory activity as described earlier. Fractions containing inhibitory activity were pooled and dialysed against buffer as described above. Aliquots (I ml) of the inhibitory pooled fractions were loaded onto a l-ml protein-A agarose column at a flow rate of approximately 005 ml/min and 0-25-ml fractions collected whilst the column was washed with three volumes of buffer. All fractions containing unbound protein were pooled, divided into aliquots and stored at - 20°C. ELISA ELISAs were performed essentially as previously described (Day et al., 1989) utilizing Nunc Immulon ELISA plates and alkaline-phosphatase-conjugated sheep anti-human IgG polyclonal antibody. Non-specific adsorption was blocked with 0-05% Tween 20 (Sigma). Protein assay Protein was assayed by the method of Bradford (1976) using bovine serum albumin as standard.
RESULTS
Effect of serum on MPO activity To assess the degree of inhibition of MPO by synovial fluids and sera, free MPO, contained in the supernatant from FMLPstimulated neutrophils, was incubated with varying concentrations of paired synovial fluids and sera from RA patients and normal human sera (Table 1). It can be seen that two out of three of the RA synovial fluids tested were less inhibitory than both
Identification of serum MPO inhibitor
349
Table 1. Effect of rheumatoid sera, synovial fluid (SF) and normal human serum (NHS) on myeloperoxidase (MPO) activity Control response (%) Fluid (%)
SF1
SF2
SF3
SRMI
SRM2
SRM3
NHSI
NHS2
1 10 20
100 15+2 12+5
100 84+3 60+6
100 65+6 30+8
100 16+3 10+2
86+2
100 45+2 15+1
77+5 17+8 10+4
100 20+9 6+5
5+4 5+2
SF1, SRM1; SF2, SRM2; and SF3, SRM3 are paired SF and sera obtained from three patients with rheumatoid arthritis. NHS1, NHS2 are normal sera obtained from normal donors. All results are expressed as a percentage of MPO activity in the absence of serum and SF. Results are the means + s.d. of triplicate determinations.
Table 2. Purification of an inhibitor of myeloperoxidase (MPO) from normal human serum
Fraction tested
Total protein (mg)
Total activity*
Yield (%)
Purification factor
Serum 20-50% NH4SO4 pellet Mono Q eluate Protein A eluate
425-0 145-0 16 5 25
19 6 18 9 9-2 6-9
100 96 47 35
2-8 12 59-8
1
* 1 unit of inhibitor equals the amount that inhibits 1 unit of MPO which causes a change of absorbance of 1 in 5 min in the assay described.
MW
-205
. . . . ;.':j' .}. ..i...,:.
...
.. ...:. ........ -. .
.:. .:::.:..::.. .: .:}:c:. 1 ....
.. ::: ....:..... .::: ....:..
. .~........ :.'~~~~~~~~ ~.' ". ~.~ ;:
......
.A' :....'.
normal and RA sera and consequently normal serum was used as the starting material for the purification. To test the validity of using a single time-point measurement to assess MPO activity, MPO was added to the substrates and absorbance readings were taken at several time intervals. The absorbance increased linearly up to 20 min, the rate being directly proportional to the amount of MPO added. The presence of serum or inhibitory preparations (see below) reduced the linear rate of absorbance change in a dosedependent manner (results not shown) and hence measurement of absorbance at 5 min were used to determine MPO activity. The inhibitory activity of serum in the MPO assay could be due to this effect upon the substrates. To examine this possibility, the assay substrates were incubated with increasing concentrations of serum for 30 min after which MPO-S was added and the enzyme activity compared with substrate incubated with HSA-HBSS. The MPO activity after incubation of substrates in HBSS alone was 1-5+001 MPO units, after incubation of substrates in two (10% v/v) normal sera was, 13+001 and 1-4 + 0-0 units, respectively, and, at 20% v/v serum, was 1-3 + 0-2 and 1-2+0 1 units, respectively.
...
.. ....
1
~~~2
Fig. 2. SDS-PAGE of the purified inhibitor preparation. The MPO inhibitor was purified and subjected to SDS-PAGE (5 yl protein) as described. Lane 1, protein A-agarose void volume eluate; lane 2, molecular weight markers (myosin, 205 kD; ,B-galactosidase, 116 kD).
Effect of albumin on MPO activity Since albumin is prevalent in serum and has known oxidant scavenging properties its effect on detection of MPO activity was assessed. MPO-S was incubated and without increasing concentrations of HSA for 30 min and MPO activity assayed. Incubations containing up to 5 mg/ml HSA showed 100%
C. M. Yea, B. Dularay & C. J. Elson
350 50
40
0C: 0
c
10
0
0
15
20
25
30
35
Molar ratio
Fig. 3. Dose dependency of inhibition of MPO. MPO was incubated at room temperature for 30 min in varying molar ratios with the purified inhibitor preparations and assayed as described. Percentage of inhibition is with respect to control wells containing 0 1 M phosphate (pH 7 4).
control activity and 10 mg/ml displayed 82 + 3% control activity. In another experiment the effect of albumin depletion of serum on MPO-inhibitory activity was investigated. Removal of albumin by cibacron blue affinity chromatography (confirmed by SDS electrophoresis) failed to reduce the MPO inhibitory activity of sera.
Purification of the MPO inhibitor Partial purification of the inhibitory activity was provided by ammonium sulphate fractionation of normal human serum using a 50-20% (sat.) cut which gave a high yield product (Table 2). Application to a Mono Q anion exchange column on a FPLC system and elution by a non-linear salt gradient (Fig. 1) yielded fractions of much higher purity although with low recovery. The bulk of the protein (up to 80%) applied to the mono Q column was not retained and no inhibitory activity was detected in these fractions. Finally, protein A-binding proteins were removed yielding fractions containing a major band of 150 kD (Fig. 2). Two minor bands of 120 kD and 180 kD were often but not always found in these preparations. Attempts to remove the minor bands using further affinity techniques including dye affinity and lectin matrices were unsuccessful although the inhibitory activity had an affinity for two types of the latter matrices (concanavalin A and Triticum vulgaris lectin).
30r 0
20-0
0
c -_
{~ c-
10
0
-1
I
5
10
I
15
20
25
30
Time (min)
Fig. 4. Time dependency of inhibition of MPO. A constant molar ratio of 20: 1 (inhibitor: MPO) was incubated at room temperature for differing lengths of time and assayed for MPO activity. Percentage is shown with respect to controls as for Fig. 3.
Identification of serum MPO inhibitor Gel filtration on a Superose 12 column (300 x 10 mm) also failed to resolve the proteins but indicated a native molecular weight of approximately 150 kD for the inhibitory activity.
Testing for immunoglobulins The presence of the inhibitory activity in the void volume of a protein A column suggests it is not IgG associated. However, the mol. wt of the inhibitor is similar to IgG and since IgG3 does not bind to protein A (Goudswaard et al., 1978), the activity could be due to autoantibodies of this isotype. To examine this possibility, we first tested the ability of human serum to react with MPO by immuno-double diffusion in agar. No precipitin line was detected. Then, the ability of total human IgG (Sigma) to inhibit MPO activity was tested but, inhibition was not observed. Thirdly, MPO was coated on to the wells of microtitre plates (5 pg/well) and the binding of IgG and IgA from normal human serum, total human IgG or purified inhibitor preparations assayed by ELISA. No uptake was detected. Furthermore, no IgG could be detected by ELISA in purified inhibitor preparations plated onto plates directly. Finally, human IgA (Sigma) had no effect upon MPO activity whereas IgA-deficient plasma was as inhibitory as normal serum (plasma IgA deficient, 90% inhibition; normal serum, 87% inhibition; both 5-pl samples). Dose and time dependency of inhibition Incubation of differing concentrations of purified inhibitor with MPO for 30 min revealed a dose-dependent inhibition of MPO activity (Fig. 3) up to 45% at an approximate molar ratio of 30: 1 (inhibitor: MPO). Constant molar ratio incubations (20:1, inhibitor: MPO), set up for varying amounts of time revealed time-dependent inhibition of MPO activity which reached a maximum after 2030 min (Fig. 4).
DISCUSSION A protein has been purified which inhibits MPO activity. The presence of several oxidant scavengers in normal human serum is well documented. However, a scavenging activity appears not to be responsible for the inhibition described here. We have shown previously that the amount of inhibition detected in whole serum (Dularay et al., 1991) and here in purified fractions (Fig. 4) is dependent on the time of incubation of the inhibitory preparations with MPO. Any molecule, such as albumin or ascorbate, which inhibits the assay by scavenging substrate (H202, TMBD) and/or product prior to colour development, e.g. TMBD radical, but not of direct interaction with MPO, would not exhibit such time dependency. The demonstration that pre-incubation of substrate with serum had little effect on MPO activity detected also argues against the MPO inhibitor acting by interfering with the assay. The effect of albumin, a prevalent protein and known antioxidant in serum, on the assay was investigated specifically. The level of inhibition by purified albumin is much less than could account for that detected in serum. Assuming a serum concentration of albumin of approximately 40 mg/ml (Putnam, 1975) the equivalent of 4 mg/ml in the assay (10% serum) inhibited 80% whereas albumin alone failed to inhibit at that level. Indeed, de-albumination of serum by cibacron blue affinity chromatography failed to reduce the inhibitory activity
351
detected in serum. Finally, no band was seen on SDS-PAGE of purified preparations which corresponded to human serum albumin. Comparison of the inhibitory activities of RA SFs and sera and NHS confirmed previous work (Dularay et al., 1991) showing that the latter was consistently more inhibitory and hence it was used as a starting material for fractionation. Purification of the inhibitory activity by FPLC anion exchange after ammonium sulphate precipitation yielded fractions of high purity although the recovery was low. It is possible that the high salt concentrations used caused irreversible denaturation of the inhibitor and future studies will need to clarify this problem. The major protein present in the Mono Q fractions was IgG and consequently immobilized protein A was used to determine whether the inhibitory activity was associated with this protein. While some activity was lost on application to the column (25% of that loaded), the majority of protein was bound (85%) suggesting that the activity was not associated with the bound protein and was not therefore IgG. Results of ELISA tests on the purified preparations support this conclusion since neither MPO-specific nor non-specific IgG was detected. In addition, neither purified IgG nor IgA exhibited MPO inhibitory activity. The fractions obtained from the protein A column showed a major band of 150 kD on SDS-PAGE. Two minor bands were also often found and a role for these cannot be ruled out at this stage. It would appear, however, that they are not structurally associated with the 1 50-kD protein, since the inhibitory activity was found to have a native molecular weight of 150 kD by gel filtration. This protein is likely to be glycosylated as the inhibitor was bound by lectin matrices (result not shown). The time and dose dependence of MPO inhibition by purified preparations, whilst not constituting kinetic characterization, are consistent with the presence of a direct-acting inhibitor in normal human serum. The activity was consistently associated with a 150-kD polypeptide and we suggest that this molecule can reduce the likelihood of extracellular hypohalide ion formation and resultant tissue damage, a postulated mechanism of several inflammatory disorders. In this respect it may be significant that lower levels of MPO inhibitory activity were detected in RA synovial fluids than either RA or normal serum, possibly due to complexing of the activity to MPO prereleased by neutrophils into the synovial fluid. The fact that those RA synovial fluids with the highest MPO activity have the lowest levels of inhibitory activity (Dularay et al., 1991) accords with this suggestion. It is in the joints of these patients, where the levels of free MPO inhibitor are reduced, that tissue damage effected by MPO-derived oxidants may occur.
ACKNOWLEDGMENTS This work was supported by a grant from the Medical Research Council, UK.
REFERENCES BABIOR, B.M. (1978) Oxygen dependent killing by phagocytes. N. Engl. J. Med. 298, 659. BAGGLIONI, M. & DEWALD, B. (1985) The neutrophil. Int. Arch. Allergy appl. Immunol. 76 (Suppl. 1), 13. Bos, A., WEVER, R. & Roos, D. (1978) Characterization and quantification of the peroxidase in human monocytes. Biochem. biophys. Acta, 252, 248.
352
C. M. Yea, B. Dularay & C. J. Elson
BRADFORD, M.M. (1976) A rapid and sensitive method of measuring microgram quantitites of protein utilizing the principle of protein-dye binding. Ann. Biochem. 72, 248. DAY, M.J., RUSSELL, J., KITWOOD, A.J., PONSFORD, M. & ELSON, C.J. (1989) Expression and regulation of erythrocyte auto-antibodies in mice following immunisation with rat erythrocytes. Eur. J. Immunol. 19, 795. DULARAY, B.D. (1989) Mechanisms of inflammation and joint damage in rheumatoid arthritis. Ph.D. thesis, Bristol University, UK. DULARAY, B.D., DIEPPE, P.A. & ELSON, C.J. (1990a) Depressed degranulation response of synovial fluid polymorphonuclear leucocytes from patients with rheumatoid arthritis to IgG aggregates. Clin. exp. Immunol. 79, 195. DULARAY, B.D., YEA, C.M. & ELSON, C.J. (1991) Inhibition of myeloperoxidase by synovial fluid and serum. Ann. rehum. Dis. (In press). DULARAY, B.D., ELSON, C.J., CLEMENTS-JEWERY, S., DAMAIS, C. & LANDO, D. (1990b) Recombinant human interleukin-l beta primes human polymorphonuclear leuckocytes for stimulus-induced myeloperoxidase release. J. Leucocyte Biol. 47, 158. GOUDSWAARSD, J., VAN DER DONK, J.A., NOORDZIJ, A., VAN DAM, R.H. & VAERMAN, J.-P. (1978) Protein A reactivity of various mammalian immunoglobulins. Scand. J. Immunol. 8, 21. HADLER, N.M., SPITZNAGEL, J.K. & QUINET, R.J. (1979) Lysosomal enzymes in inflammatory synovial effusions. J. Immunol. 123, 572.
HENDERSON, L.M., CHAPPELL, J.B. & JONES, O.T.G. (1987) The superoxide-generating NADPH oxidase of human neutrophils is electrogenic and associated with H + channel. Biochem. J. 246, 325. HENSON, P.M. (1971) Interaction of cells with immune complexes adherence and release of constitutents and tissue injury. J. exp. Med. 134, 114s. HENSON, P.M., JOHNSON, H. & SPIEGELBERG, H.L. (1972) The release of granule constituents from human neutrophils. J. Immunol. 109, 1182. KLEBANOFF, S.J. (1975) Antimicrobial systems of the PMN. In The Phagocytic Cell in Host Resistance (ed. by J. A. Bellanti & D. M. Dayton) p. 45. Raven Press, New York. LAEMMLI, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680. MORRISON, A.D., PRUZANSKI, W. & RANADIVE, N.S. (1978) Release of lysosomal enzymes from human polymorphonuclear leucocytes by soluble intermediate immune complexes. Scand. J. Rheumatol. 7, 241. PALMER, D.G. (1968) Total leukocyte enumeration in pathogenic synovial fluids. Am. J. clin. Pathol. 49, 812. PUTNAM, F.M. (1975) In The Plasma Proteins. Structure Function and Genetic Control 2nd edn, p. 64. Academic Press, New York. SUZUKI, K., OTA, H., SASAGAWA, S., SAKATANI, T. & FUJIKURA, T. (1983) Assay method for myeloperoxidase in human polymorphonuclear leucocytes. Ann. Biochem. 132, 345. WEISSMANN, G. (1972) Lysosymal mechanisms of tissue injury in arthritis. N. Engl. J. Med. 286, 141.
AD)ONIS 00099104910015 1J
Identification of a myeloperoxidase inhibitor from normal human serum C. M. YEA, B. DULARAY & C. J. ELSON Department of Pathology, University of Bristol, Bristol, England
(Acceptedfor publication 19 November 1990)
SUMMARY An inhibitor of myeloperoxidase (MPO) has been identified in normal human serum. Initial experiments confirmed that high levels of MPO inhibitory activity are present in sera and that the inhibitor did not act by interfering with the assay. Purification of the inhibitor activity by salt precipitation followed by ion exchange and affinity chromatography revealed the presence of a protein of 150 kD. The purified inhibitory activity displayed dose and time dependency and was not associated with IgG or IgA. It is considered that human serum contains an inhibitor of extracellular MPO capable of protecting against hypohalous acid release in host tissues and that if inhibitor levels are
reduced such protection
may
fail.
Keywords myeloperoxidase inhibitor human
serum
INTRODUCTION
Neutrophils play a key role in defence of the host against microbial attack. They execute this function through two major pathways of killing: dependent and independent on oxygen. After phagocytosis of an invading microbe into an enclosed phagolysosome within the cell, activation of a plasma-membrane-bound enzyme system (NADPH oxidase) occurs and superoxide and other derived radicals are released into the vacuole (Babior, 1978). In conjunction, granules located in the cytoplasm of the neutrophil fuse with the phagolysosome membrane and empty their contents into it, releasing an array of degradative enzymes. One of the major enzymes released during the latter, degranulative process, is myeloperoxidase (MPO) which combines halide ions (usually Cl ) with hydrogen peroxide forming the highly microbicidal hypohalous acid (Klebanoff, 1975). Together these processes result in the death of engulfed bacteria. Release of MPO and other granule enzymes and indeed of superoxide, is initiated before the phagolysosome has completely sealed off (Bagglioni & Dewald, 1985) resulting in the release of granule enzymes into the surrounding milieu. In addition, enzymes are released directly from granules if neutrophils react with surfaces too extensive for engulfment (Henson, 1971; Henson, Johnson & Spiegelberg, 1972). It has been suggested that such extracellular release of granule enzymes occurs in a number of inflammatory diseases including rheumatoid arthritis (RA) (Weismann, 1972). In this condition, high numbers of neutrophils are found in synovial fluids (Palmer, 1968) together with immunoglobulin aggregates capable of stimulating a degranulation response (Morrison,
Pruzanski & Radadive, 1978; Dularay, Dieppe & Elson, 1990) and granule enzymes including MPO (Hadler, Spitznagel & Quinet, 1979; Dularay et al., 1990a). The contribution of these enzymes to the pathogenesis of RA remains unclear since RA synovial fluids also contain inhibitors for many granule enzymes. However, to our knowledge an inhibitor of MPO has not been described. Recently, during an investigation of the degranulation response of RA synovial neutrophils, we demonstrated that RA synovial fluid and sera and normal sera reduced the MPO activity of degranulating neutrophils (Dularay et al., 1990a). These results led us to suggest that the fluids contained a MPO inhibitor. Work reported elsewhere (Dularay, Yea & Elson, 1991) gives credence to this idea by showing that the fluids inhibited the released enzyme rather than the degranulation response of polymorphonuclear leucocytes and also that incubation of synovial fluids with enzyme substrate did not inhibit MPO activity. Moreover, as glucuronidase activity was unaffected by the presence of the fluids, the inhibitor appeared to be peroxidase specific. The purpose of the current work was to determine whether a particular component of the fluids could be identified which exhibited inhibitory activity. MATERIALS AND METHODS Patients Blood and synovial fluids were obtained from RA patients
attending the Rheumatology Outpatients Clinic at the Bristol Royal Infirmary. All patients fulfilled the standard criteria for classical or definite RA. Assay of MPO activity MPO was assayed using a modification of the method of Suzuki et al. (1983) utilizing 3,3',5,5' tetramethyl benzidine (TMBD) in
Correspondence: Dr C. J. Elson, Department of Pathology, University of Bristol, Bristol BS8 ITD, UK.
347
348
C. M. Yea, B. Dularay & C. J. Elson
96-well microtitre plates. MPO was purified as described below and 50 ng (determined spectrophotometrically using a reducedoxidized extinction coefficient of 75 mm- ' at 472 nm; Bos, Wever & Roos, 1978) were added to each well in 20 kl of 0 I M sodium phosphate, pH 7-4. After addition of 30 pl substrate to each well (3-7 mM TMBD, 0-002% H202) in 1:3 mixture of 0-01 M sodium acetate, pH 5-2, and PBS, the plate was incubated at room temperature for 5 min and the absorbance at 690-540 nm was read in a Labsystem automatic plate reader. Blanks contained equal volumes with no MPO. The change of absorbance at 690540 nm was measured. A unit of MPO is the amount causing a change of 1 in 5 min.
Purification of MPO Human polymorphonuclear leucocytes were prepared from buffy coats obtained from the South West Regional Blood Transfusion Centre (Bristol) as described previously (Henderson, Chappell & Jones, 1987). They were resuspended in PBS, sonicated on ice (3 x 1O sec, 50 W) and the membranes prepared by centrifugation (100 000 g, 30 min). After pre-extraction with 0-25% Lubrol PX/0-25% sodium deoxycholate in 10 mm glycine (pH 8 0), they were extracted twice with 0 5% cetyltrimethyl ammonium bromide in 0 I M phosphate buffer, pH 7-4. The supernatant was collected and loaded onto a 12 cm x 12 mm column of S-Sepharose (Pharmacia) pre-equilibrated with extraction buffer (without detergent). After extensive washing the column was eluted with a gradient of 0 2-0 5 M sodium phosphate, pH 7-4, and 3-ml fractions collected. Those displaying absorbance at 340 nm > 0-05 were pooled. Preparations such as these typically had a R, of 0-62-0-68 and displayed one major band (approximately 65 kD) on Coomassie blue stained SDS-PAGE. SDS-PAGE was performed by the discontinuous buffer method of Laemmli (1970) utilizing a 12% separating gel. In the initial experiments supernatants from stimulated neutrophils were used as a source of MPO (MPO-S). Briefly, neutrophils were pretreated with 20 pg/ml cytochalasin B (Sigma; I mg/ml in DMSO) for 10 min at room temperature. The treated cells (2 x 108/ml) were incubated in HBSS containing 10 mg/ml human serum albumin (HSA-HBSS) and formylmethionyl-leucyl-phenylalanine (2 5 x 106 M) for 30 min at 37°C. In our hands these conditions were optimal for maximum MPO release (Dularay, 1989; Dularay et al., 1990b). Inhibition studies The amount of MPO-inhibitory activity present in samples was established using the MPO assay described above. Samples were added to the MPO in the plate (controls contained equal volumes of phosphate buffer) and incubated for 30 min at room temperature prior to addition of substrate, unless otherwise indicated. Amount of inhibitor is expressed such that 1 unit of inhibitor inhibits I unit of MPO.
Purification of MPO inhibitor Pooled normal human serum was prepared from blood collected from healthy volunteers. It was stored at - 70°C without preservatives until used. All procedures were carried out in 0 1 M sodium phosphate buffer (pH 7.4) unless otherwise stated. The serum was centrifuged to remove all particulate matter (100 000 g, I h). The supernatant was collected, made up to 20% ammonium sulphate by the addition of cold saturated ammonium sulphate, stirred for 4 h at 4°C and centrifuged (100 000 g,
*°r-
0-75I 0-5 _IN0co IC:
0
0-251-
z
O
Fig. 1. Purification of the MPO inhibitor by FPLC anion exchange. The ammonium sulphate purified fraction (6 ml) was applied to the Mono Q column and eluted using a salt gradient as shown (--- ). The A280 was ) and 1-ml fractions were collected and continuously monitored ( assayed for inhibitory activity as described in Materials and Methods.
30 min). The supernatant was collected and made up to 50% saturation ammonium sulphate saturation by the further addition of cold ammonium sulphate. After 4-h stirring at 4°C the mixture was centrifuged (1l0 000 g, 30 min) and the supernatant discarded. The precipitate was redissolved in phosphate buffer, centrifuged to remove insoluble material and dialysed overnight against two changes of 10 volumes of phosphate buffer. Approximately 10 mg of the partially purified protein (estimated by absorbance at 280 nm) were loaded onto a 1 ml Mono Q anion exchange column (pre-equlibrated with phosphate buffer) on a Pharmacia FPLC system at a flow rate of 0 1 ml/min and washed with a further 10-column volumes of buffer. Bound proteins were eluted with a non-linear gradient of NaCl (see Fig. 1) were from 0 to 0-5 M. Fractions of I ml collected and assayed for MPO-inhibitory activity as described earlier. Fractions containing inhibitory activity were pooled and dialysed against buffer as described above. Aliquots (I ml) of the inhibitory pooled fractions were loaded onto a l-ml protein-A agarose column at a flow rate of approximately 005 ml/min and 0-25-ml fractions collected whilst the column was washed with three volumes of buffer. All fractions containing unbound protein were pooled, divided into aliquots and stored at - 20°C. ELISA ELISAs were performed essentially as previously described (Day et al., 1989) utilizing Nunc Immulon ELISA plates and alkaline-phosphatase-conjugated sheep anti-human IgG polyclonal antibody. Non-specific adsorption was blocked with 0-05% Tween 20 (Sigma). Protein assay Protein was assayed by the method of Bradford (1976) using bovine serum albumin as standard.
RESULTS
Effect of serum on MPO activity To assess the degree of inhibition of MPO by synovial fluids and sera, free MPO, contained in the supernatant from FMLPstimulated neutrophils, was incubated with varying concentrations of paired synovial fluids and sera from RA patients and normal human sera (Table 1). It can be seen that two out of three of the RA synovial fluids tested were less inhibitory than both
Identification of serum MPO inhibitor
349
Table 1. Effect of rheumatoid sera, synovial fluid (SF) and normal human serum (NHS) on myeloperoxidase (MPO) activity Control response (%) Fluid (%)
SF1
SF2
SF3
SRMI
SRM2
SRM3
NHSI
NHS2
1 10 20
100 15+2 12+5
100 84+3 60+6
100 65+6 30+8
100 16+3 10+2
86+2
100 45+2 15+1
77+5 17+8 10+4
100 20+9 6+5
5+4 5+2
SF1, SRM1; SF2, SRM2; and SF3, SRM3 are paired SF and sera obtained from three patients with rheumatoid arthritis. NHS1, NHS2 are normal sera obtained from normal donors. All results are expressed as a percentage of MPO activity in the absence of serum and SF. Results are the means + s.d. of triplicate determinations.
Table 2. Purification of an inhibitor of myeloperoxidase (MPO) from normal human serum
Fraction tested
Total protein (mg)
Total activity*
Yield (%)
Purification factor
Serum 20-50% NH4SO4 pellet Mono Q eluate Protein A eluate
425-0 145-0 16 5 25
19 6 18 9 9-2 6-9
100 96 47 35
2-8 12 59-8
1
* 1 unit of inhibitor equals the amount that inhibits 1 unit of MPO which causes a change of absorbance of 1 in 5 min in the assay described.
MW
-205
. . . . ;.':j' .}. ..i...,:.
...
.. ...:. ........ -. .
.:. .:::.:..::.. .: .:}:c:. 1 ....
.. ::: ....:..... .::: ....:..
. .~........ :.'~~~~~~~~ ~.' ". ~.~ ;:
......
.A' :....'.
normal and RA sera and consequently normal serum was used as the starting material for the purification. To test the validity of using a single time-point measurement to assess MPO activity, MPO was added to the substrates and absorbance readings were taken at several time intervals. The absorbance increased linearly up to 20 min, the rate being directly proportional to the amount of MPO added. The presence of serum or inhibitory preparations (see below) reduced the linear rate of absorbance change in a dosedependent manner (results not shown) and hence measurement of absorbance at 5 min were used to determine MPO activity. The inhibitory activity of serum in the MPO assay could be due to this effect upon the substrates. To examine this possibility, the assay substrates were incubated with increasing concentrations of serum for 30 min after which MPO-S was added and the enzyme activity compared with substrate incubated with HSA-HBSS. The MPO activity after incubation of substrates in HBSS alone was 1-5+001 MPO units, after incubation of substrates in two (10% v/v) normal sera was, 13+001 and 1-4 + 0-0 units, respectively, and, at 20% v/v serum, was 1-3 + 0-2 and 1-2+0 1 units, respectively.
...
.. ....
1
~~~2
Fig. 2. SDS-PAGE of the purified inhibitor preparation. The MPO inhibitor was purified and subjected to SDS-PAGE (5 yl protein) as described. Lane 1, protein A-agarose void volume eluate; lane 2, molecular weight markers (myosin, 205 kD; ,B-galactosidase, 116 kD).
Effect of albumin on MPO activity Since albumin is prevalent in serum and has known oxidant scavenging properties its effect on detection of MPO activity was assessed. MPO-S was incubated and without increasing concentrations of HSA for 30 min and MPO activity assayed. Incubations containing up to 5 mg/ml HSA showed 100%
C. M. Yea, B. Dularay & C. J. Elson
350 50
40
0C: 0
c
10
0
0
15
20
25
30
35
Molar ratio
Fig. 3. Dose dependency of inhibition of MPO. MPO was incubated at room temperature for 30 min in varying molar ratios with the purified inhibitor preparations and assayed as described. Percentage of inhibition is with respect to control wells containing 0 1 M phosphate (pH 7 4).
control activity and 10 mg/ml displayed 82 + 3% control activity. In another experiment the effect of albumin depletion of serum on MPO-inhibitory activity was investigated. Removal of albumin by cibacron blue affinity chromatography (confirmed by SDS electrophoresis) failed to reduce the MPO inhibitory activity of sera.
Purification of the MPO inhibitor Partial purification of the inhibitory activity was provided by ammonium sulphate fractionation of normal human serum using a 50-20% (sat.) cut which gave a high yield product (Table 2). Application to a Mono Q anion exchange column on a FPLC system and elution by a non-linear salt gradient (Fig. 1) yielded fractions of much higher purity although with low recovery. The bulk of the protein (up to 80%) applied to the mono Q column was not retained and no inhibitory activity was detected in these fractions. Finally, protein A-binding proteins were removed yielding fractions containing a major band of 150 kD (Fig. 2). Two minor bands of 120 kD and 180 kD were often but not always found in these preparations. Attempts to remove the minor bands using further affinity techniques including dye affinity and lectin matrices were unsuccessful although the inhibitory activity had an affinity for two types of the latter matrices (concanavalin A and Triticum vulgaris lectin).
30r 0
20-0
0
c -_
{~ c-
10
0
-1
I
5
10
I
15
20
25
30
Time (min)
Fig. 4. Time dependency of inhibition of MPO. A constant molar ratio of 20: 1 (inhibitor: MPO) was incubated at room temperature for differing lengths of time and assayed for MPO activity. Percentage is shown with respect to controls as for Fig. 3.
Identification of serum MPO inhibitor Gel filtration on a Superose 12 column (300 x 10 mm) also failed to resolve the proteins but indicated a native molecular weight of approximately 150 kD for the inhibitory activity.
Testing for immunoglobulins The presence of the inhibitory activity in the void volume of a protein A column suggests it is not IgG associated. However, the mol. wt of the inhibitor is similar to IgG and since IgG3 does not bind to protein A (Goudswaard et al., 1978), the activity could be due to autoantibodies of this isotype. To examine this possibility, we first tested the ability of human serum to react with MPO by immuno-double diffusion in agar. No precipitin line was detected. Then, the ability of total human IgG (Sigma) to inhibit MPO activity was tested but, inhibition was not observed. Thirdly, MPO was coated on to the wells of microtitre plates (5 pg/well) and the binding of IgG and IgA from normal human serum, total human IgG or purified inhibitor preparations assayed by ELISA. No uptake was detected. Furthermore, no IgG could be detected by ELISA in purified inhibitor preparations plated onto plates directly. Finally, human IgA (Sigma) had no effect upon MPO activity whereas IgA-deficient plasma was as inhibitory as normal serum (plasma IgA deficient, 90% inhibition; normal serum, 87% inhibition; both 5-pl samples). Dose and time dependency of inhibition Incubation of differing concentrations of purified inhibitor with MPO for 30 min revealed a dose-dependent inhibition of MPO activity (Fig. 3) up to 45% at an approximate molar ratio of 30: 1 (inhibitor: MPO). Constant molar ratio incubations (20:1, inhibitor: MPO), set up for varying amounts of time revealed time-dependent inhibition of MPO activity which reached a maximum after 2030 min (Fig. 4).
DISCUSSION A protein has been purified which inhibits MPO activity. The presence of several oxidant scavengers in normal human serum is well documented. However, a scavenging activity appears not to be responsible for the inhibition described here. We have shown previously that the amount of inhibition detected in whole serum (Dularay et al., 1991) and here in purified fractions (Fig. 4) is dependent on the time of incubation of the inhibitory preparations with MPO. Any molecule, such as albumin or ascorbate, which inhibits the assay by scavenging substrate (H202, TMBD) and/or product prior to colour development, e.g. TMBD radical, but not of direct interaction with MPO, would not exhibit such time dependency. The demonstration that pre-incubation of substrate with serum had little effect on MPO activity detected also argues against the MPO inhibitor acting by interfering with the assay. The effect of albumin, a prevalent protein and known antioxidant in serum, on the assay was investigated specifically. The level of inhibition by purified albumin is much less than could account for that detected in serum. Assuming a serum concentration of albumin of approximately 40 mg/ml (Putnam, 1975) the equivalent of 4 mg/ml in the assay (10% serum) inhibited 80% whereas albumin alone failed to inhibit at that level. Indeed, de-albumination of serum by cibacron blue affinity chromatography failed to reduce the inhibitory activity
351
detected in serum. Finally, no band was seen on SDS-PAGE of purified preparations which corresponded to human serum albumin. Comparison of the inhibitory activities of RA SFs and sera and NHS confirmed previous work (Dularay et al., 1991) showing that the latter was consistently more inhibitory and hence it was used as a starting material for fractionation. Purification of the inhibitory activity by FPLC anion exchange after ammonium sulphate precipitation yielded fractions of high purity although the recovery was low. It is possible that the high salt concentrations used caused irreversible denaturation of the inhibitor and future studies will need to clarify this problem. The major protein present in the Mono Q fractions was IgG and consequently immobilized protein A was used to determine whether the inhibitory activity was associated with this protein. While some activity was lost on application to the column (25% of that loaded), the majority of protein was bound (85%) suggesting that the activity was not associated with the bound protein and was not therefore IgG. Results of ELISA tests on the purified preparations support this conclusion since neither MPO-specific nor non-specific IgG was detected. In addition, neither purified IgG nor IgA exhibited MPO inhibitory activity. The fractions obtained from the protein A column showed a major band of 150 kD on SDS-PAGE. Two minor bands were also often found and a role for these cannot be ruled out at this stage. It would appear, however, that they are not structurally associated with the 1 50-kD protein, since the inhibitory activity was found to have a native molecular weight of 150 kD by gel filtration. This protein is likely to be glycosylated as the inhibitor was bound by lectin matrices (result not shown). The time and dose dependence of MPO inhibition by purified preparations, whilst not constituting kinetic characterization, are consistent with the presence of a direct-acting inhibitor in normal human serum. The activity was consistently associated with a 150-kD polypeptide and we suggest that this molecule can reduce the likelihood of extracellular hypohalide ion formation and resultant tissue damage, a postulated mechanism of several inflammatory disorders. In this respect it may be significant that lower levels of MPO inhibitory activity were detected in RA synovial fluids than either RA or normal serum, possibly due to complexing of the activity to MPO prereleased by neutrophils into the synovial fluid. The fact that those RA synovial fluids with the highest MPO activity have the lowest levels of inhibitory activity (Dularay et al., 1991) accords with this suggestion. It is in the joints of these patients, where the levels of free MPO inhibitor are reduced, that tissue damage effected by MPO-derived oxidants may occur.
ACKNOWLEDGMENTS This work was supported by a grant from the Medical Research Council, UK.
REFERENCES BABIOR, B.M. (1978) Oxygen dependent killing by phagocytes. N. Engl. J. Med. 298, 659. BAGGLIONI, M. & DEWALD, B. (1985) The neutrophil. Int. Arch. Allergy appl. Immunol. 76 (Suppl. 1), 13. Bos, A., WEVER, R. & Roos, D. (1978) Characterization and quantification of the peroxidase in human monocytes. Biochem. biophys. Acta, 252, 248.
352
C. M. Yea, B. Dularay & C. J. Elson
BRADFORD, M.M. (1976) A rapid and sensitive method of measuring microgram quantitites of protein utilizing the principle of protein-dye binding. Ann. Biochem. 72, 248. DAY, M.J., RUSSELL, J., KITWOOD, A.J., PONSFORD, M. & ELSON, C.J. (1989) Expression and regulation of erythrocyte auto-antibodies in mice following immunisation with rat erythrocytes. Eur. J. Immunol. 19, 795. DULARAY, B.D. (1989) Mechanisms of inflammation and joint damage in rheumatoid arthritis. Ph.D. thesis, Bristol University, UK. DULARAY, B.D., DIEPPE, P.A. & ELSON, C.J. (1990a) Depressed degranulation response of synovial fluid polymorphonuclear leucocytes from patients with rheumatoid arthritis to IgG aggregates. Clin. exp. Immunol. 79, 195. DULARAY, B.D., YEA, C.M. & ELSON, C.J. (1991) Inhibition of myeloperoxidase by synovial fluid and serum. Ann. rehum. Dis. (In press). DULARAY, B.D., ELSON, C.J., CLEMENTS-JEWERY, S., DAMAIS, C. & LANDO, D. (1990b) Recombinant human interleukin-l beta primes human polymorphonuclear leuckocytes for stimulus-induced myeloperoxidase release. J. Leucocyte Biol. 47, 158. GOUDSWAARSD, J., VAN DER DONK, J.A., NOORDZIJ, A., VAN DAM, R.H. & VAERMAN, J.-P. (1978) Protein A reactivity of various mammalian immunoglobulins. Scand. J. Immunol. 8, 21. HADLER, N.M., SPITZNAGEL, J.K. & QUINET, R.J. (1979) Lysosomal enzymes in inflammatory synovial effusions. J. Immunol. 123, 572.
HENDERSON, L.M., CHAPPELL, J.B. & JONES, O.T.G. (1987) The superoxide-generating NADPH oxidase of human neutrophils is electrogenic and associated with H + channel. Biochem. J. 246, 325. HENSON, P.M. (1971) Interaction of cells with immune complexes adherence and release of constitutents and tissue injury. J. exp. Med. 134, 114s. HENSON, P.M., JOHNSON, H. & SPIEGELBERG, H.L. (1972) The release of granule constituents from human neutrophils. J. Immunol. 109, 1182. KLEBANOFF, S.J. (1975) Antimicrobial systems of the PMN. In The Phagocytic Cell in Host Resistance (ed. by J. A. Bellanti & D. M. Dayton) p. 45. Raven Press, New York. LAEMMLI, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680. MORRISON, A.D., PRUZANSKI, W. & RANADIVE, N.S. (1978) Release of lysosomal enzymes from human polymorphonuclear leucocytes by soluble intermediate immune complexes. Scand. J. Rheumatol. 7, 241. PALMER, D.G. (1968) Total leukocyte enumeration in pathogenic synovial fluids. Am. J. clin. Pathol. 49, 812. PUTNAM, F.M. (1975) In The Plasma Proteins. Structure Function and Genetic Control 2nd edn, p. 64. Academic Press, New York. SUZUKI, K., OTA, H., SASAGAWA, S., SAKATANI, T. & FUJIKURA, T. (1983) Assay method for myeloperoxidase in human polymorphonuclear leucocytes. Ann. Biochem. 132, 345. WEISSMANN, G. (1972) Lysosymal mechanisms of tissue injury in arthritis. N. Engl. J. Med. 286, 141.
