Clinical Science and Molecular Medicine (1974) 41,403414.
NEUTRAL PROTEASE FROM THE
POLYMORPHONUCLEAR LEUCOCYTES OF H U M A N
RHEUMATOID SYNOVIAL F L U I D
R. H. PRYCE-JONES,* J . SAKLATVALAT
AND
G. C. WOOD*
* Department of Pharmaceutical Chemistry, University of Strathclyde, Glasgow, and
f Centrefor Rheumatic Diseases, University Department of Medicine, Royal Inzrmary, Glasgow
(Received 3 May 1974)
SUMMARY
1. The cartilage-proteoglycan-degrading activity of synovial fluid cells from
rheumatoid patients is primarily due to neutral protease activity; hyaluronidase does
not contribute significantly. Evidence for these conclusions is adduced from electrophoresis of degradation products, from comparison of the proteoglycan-degrading
and protease activities of extracts and from the effects of activators and inhibitors.
2. Most of the neutral protease activity is insoluble in buffer of low ionic strength
and may thus be separated from small amounts of other, soluble, proteoglycandegrading enzymes. This ‘insoluble’ protease dissolves in KC1 (1 mol/l) and has
appreciable solubility even at physiological ionic strength.
3. The protease has optimum activity between pH 7.5 and pH 10, with little or no
activity below pH 6-0.
4. The enzyme hydrolyses a number of substrates including elastin and two synthetic substrates for elastase. Hydrolysis of the latter is inhibited by dilutions of
synovial fluids and serum and it thus differs from the elastase-like esterase activity
of synovial fluid and serum.
5. The ‘insoluble’ enzyme fraction caused loss of staining and release of hexuronate
from slices of bovine articular cartilage at salt concentrations near physiological and
these effects were almost completely inhibited by cell-free synovial fluids.
6. The properties of the synovial fluid cell neutral protease are compared with
those of the similar enzyme activity found in granulocytes from blood and it is concluded that the synovial fluid enzyme is probably derived from granulocytes.
Key words : cartilage, granulocytes, protease, proteoglycan, rheumatoid arthritis,
synovial fluid.
Cells present in the synovial fluids of patients with rheumatoid arthritis contain enzymic
activity, optimum at neutral to alkalinepH, which reduces the viscosity of solutions of partially
Correspondence: Dr G. C. Wood, Department of Pharmaceutical Chemistry, University of Strathclyde,
Glasgow G1 1XW.
403
404
R.H. Pryce-Jones, J. Saklatvala and G. C. Wood
purified proteoglycans (chondromucoprotein) extracted from cartilage (Wood, Pryce-Jones,
White & Nuki, 1971). Proteoglycans are structurally important polymers which comprise a
major fraction of the organic material in articular cartilage. They consist of glycosaminoglycan
macromolecules (chondroitin 4-sulphate, chondroitin 6-sulphate and keratan sulphate)
covalently linked to protein. Degradation of these cartilage proteoglycans leads to loss of the
characteristicvisco-elastic properties of the tissue, which are essential to its normal functioning,
and we suggested that the enzyme activity we had observed might contribute to the degradation
of cartilage in rheumatoid arthritis.
The term chondromucoprotein (Malawista & Schubert, 1958)is used to distinguish partially
purified proteoglycans, such as we used in our earlier work, from highly purified proteoglycans.
Unlike the latter, chondromucoprotein is heterogeneous in the analytical ultracentrifuge and
contains small amounts of collagen. Our previous work indicated that the reduction in the
viscosity of chondromucoprotein was not due to degradation of the glycosaminoglycan portion
of the proteoglycans or, primarily, to degradation of collagen but was probably due to hydrolysis of non-collagenous protein. Further evidence for this view is now presented.
Some properties of the synovial fluid cell neutral proteases are described and discussed in
relation to blood granulocyte proteases. Some of these may be implicated in the degradation of
cartilage proteoglycans in rheumatoid arthritis (Ziff, Gribetz & LosPalluto, 1960; Janoff &
Blondin, 1970; Oronsky, Ignarro & Perper, 1973).
MATERIALS AND METHODS
Cells were separated from synovial fluids (10-50 ml) of rheumatoid patients by centrifuging,
washed with NaCl(O.15 mol/l), suspended in 5 ml of Tris-HC1 buffer (0.1 mol/l) at pH 7-5 and
disintegrated by sonication (MSE Ultrasonic Disintegrator, 2 min, 04°C). Cell debris was
removed by centrifuging (2000 g, 15 min, 4°C) and the resulting opalescent supernatant (‘crude
cell extract’) used as a source of enzyme (Wood et al., 1971). In some experiments crude cell
extract (4 ml) was separated into a pellet, fraction P, and a soluble fraction, fraction S, by
centrifuging at 100000 g for 1 h at 4°C (Spinco model L centrifuge, rotor no. 40 E). Fraction P
was suspended in 4 ml of Tris-HC1 buffer (0.1 mol/l) at pH 7.5.
Polymorphonuclear leucocytes (granulocytes) and lymphocytes were isolated from blood
(25 ml) of normal humans or rheumatoid patients by the method of Bayum (1968). Preparations of both types of cell were more than 93% homogeneous. The cells were sonicated in 2 ml
of Tris-HC1 buEer (0.1 mol/l) at pH 7.5 as described above and the crude sonicate was used
without centrifugation.
Chondromucoprotein was prepared from bovine nasal septum as described by Partridge,
Davis & Adair (1961). A purified proteoglycan (PPL-3) was prepared from the same source by
the method of Pal, Doganges & Schubert (1966).
Enzyme assays
The proteoglycan-degrading activity of enzyme preparations was determined by measuring
the reduction of the viscosity of solutions of chondromucoprotein or PPL-3 (Wood et al.,
1971). One unit of activity is that amount of enzyme which causes 20% decrease in specific
viscosity in 10 min when added, in 0.2 ml, to 2 ml of substrate solution (0.4% in Tris-HCl
buffer, 0.1 mol/l, pH 7-5 at 25°C).
Granulocyte neutral protease
405
Protease activity was determined with a number of proteins as substrates. With ureadenatured haemoglobin (Koch-Light Ltd, Colnbrook, Bucks.) the method was that of
Laskowski (1955). With NN-dimethylcasein (BDH Ltd, Poole, Dorset) the method was
adapted from that of Lin, Means & Feeney (1969). The incubation volume was 1.5 ml and 1 ml
of 10% (w/w) Triton X 100 (Sigma, London) was used in place of sodium dodecyl sulphate.
Glycine (BDH Ltd) was used as a standard to calculate the concentration of amino endgroups released. With casein the method of Kunitz (1947) was used.
Hide-powder azure (grade B, Calbiochem Ltd, London), 3 mg in 1-4ml buffer, was used as a
protease substrate essentially as described by Rinderknecht, Silverman, Geokas & Haverback
(1970). The assays were terminated by rapid filtration. Enzyme activity, in milligrams of
of the filtrate
substrate dissolved per hour, was determined by comparing the extinction (ESg5)
with that of hide-powder azure (3 mg/l.4 ml) which had been completely dissolved by chymotrypsin (EC 3.4.21.1 ; Sigma, London).
Elastin, prepared from ox ligamentum nuchae by the method of Partridge & Davis (1955),
was covalently dyed with Remazol Brilliant Blue R (Hoeschst-Castella, Ltd, Manchester)
(RBB) as described by Rinderknecht, Geokas, Silverman, Lillard & Haverback (1968). The
resulting RBB-elastin was used as a protease substrate as described for hide-powder azure
except that the incubation time was 6 h and hog pancreatic elastase (EC 3.4.21.1 1) was used to
determine the absorbance of the substrate which had been completely dissolved.
Calf-thymus histone (Koch-Light Ltd) was used as a protease substrate as described by
Davies, Rita, Krakauer & Weissmann (1971), except that the released end-groups were estimated by using the trinitrobenzene sulphonic acid (Sigma, London) method of Lin et al. (1969)
with the concentration of NaHC03 increased to 95 mmol/l.
The rates of hydrolysis of acetyltyrosineethyl ester, benzoyltyrosineethyl ester, tosylarghine
methyl ester and benzoylarginine ethyl ester at pH 7.9 were determined as described by Humme1 (1959). The rates of hydrolysis of benzoylarginine p-nitroanilide and benzoylarginine
P-naphthylamideat pH 5.0 and of leucine-B-naphthylamideat pH 7.0 were determined by the
methods of Erlanger, Kakowsky & Cohen (1961), and Blackwood & Mandl (1961), respectively
(all these synthetic substrates were from BDH Ltd). The hydrolysis of N-(t-BOC)-L-aIanine
p-nitrophenyl ester (NBA; Sigma, London) and CBZ-L-alanine p-nitrophenyl ester (NCA;
Sigma, London), two synthetic substrates for elastase (Janoff, 1969; Shotton, 1970), was
followed at 37°C by the method of Janoff (1969). To prepare the substrate solutions NBA
(20 pmol) was dissolved in 5 ml of methanol and NCA (1 5 pmol) was dissolved in 10 ml of
methanol; each solution was then diluted to 100 ml with sodium phosphate buffer (0.1 mol/l) at
pH 7.5, containing KC1 (0.2 mol/l).
Protein concentrations were determined by the method of Miller (1959).
Electrophoresis
PPL-3 (4 mg in 1 ml) was incubated at 37°C for 16 h with synovial fluid cell extracts (5
proteoglycan-degrading units in 0.1 ml), chymotrypsin ( 5 pg in 0.1 ml) or hyaluronidase (EC
3.2.1.35 and 36) (Sigma, London) (50 units in 0.1 ml). When synovial fluid cell extract and
chymotrypsin were used the solvent for substrate and enzymes was Tris-HC1 buffer (0.1
mol/l) at pH 7.5. When hyaluronidase was used the solvent was sodium acetate buffer (0.1
mol/l) at pH 5.0. Electrophoresis of the degradation products and of chondroitin sulphate
(Sigma, London) was carried out on polyacrylamide gels (13%, w/v, acrylamide and 0.4%,
406
R. H. Pryce-Jones, J. Saklatvala and G. C. Wood
w/v, bisacrylamide) as described by Davis (1964). The current was 2 mA/tube for 40 min. Gels
were stained with Alcian Blue 8GX (Edward Gurr Ltd, London) or Saffranine 0 (Hopkin and
Williams, Romford, Essex).
Degradation of cartilage
Slices of cartilage (20-40 mg each) were cut from the fore-limb hoof joints of a bullock within
2 h of the death of the animal. In all subsequent operations sterile glassware, buffers and salt
solutions were used. All solutions contained penicillin (Glaxo Ltd; 50 units/ml) and streptomycin (BDH Ltd; 250 mg/l), which did not affect the proteoglycan-degrading activity of
synovial fluid cell extracts. Cartilage slices were washed with KC1 solution (0-2 mol/l) and
samples (60 mg) were incubated with shaking at 37°C for 16 h in buffers ( 5 ml) containing
fraction P of synovial fluid cell extracts (10 units of proteoglycan-degradingactivity/ml). The
buffers (0-1 mol/l+KCl 0.2 mol/l) were acetate (pH 4 9 , phosphate (pH 7.5) and glycine
(pH 9.0). Other cartilage samples were incubated similarly with (a) 5 ml of buffer containing
0-5 ml of cell-free synovial fluid, (b) 5 ml of buffer containing fraction P plus cell-free synovial
fluid and (c) 5 ml of buffer alone. After incubation the uronic acid concentration in the supernatant was determined (Bitter & Muir, 1962). The cartilage slices were examined macroscopically, fixed in 4% buffered formalin at pH 4.0 and sectioned perpendicularlyto the articular surfaces. Serial sections were stained with Alcian Blue and Toluidine Blue (Edward Gurr
Ltd).
RESULTS
Electrophoresis of enzymically degraded proteoglycan
Electrophoresis of PPL-3 degraded by synovial fluid cell extracts or chymotrypsin gave
similar patterns, whose most prominent feature was a sharp band which stained with Alcian
Blue and Saffranine 0 and had the mobility of chondroitin sulphate. Hyaluronidase-degraded
PPL-3 gave only one band, which stained only with Alcian Blue and had a mobility 43% greater
than that of chondroitin sulphate.
Inhibition and activation of the enzyme activity of crude cell extracts
The proteoglycan-degrading and protease activities of synovial fluid cell extracts were affected
similarly by a number of inhibiting and activating substances (Table 1).
Comparison of proteoglycan-degrading and protease activities of crude cell extracts
Determination of the enzyme activities of cells from thirteen samples of synovial fluid
showed a close correlation between proteoglycan-degrading activity and protease activity with
hide-powder azure substrate (Fig. 1). A similar experiment with a further fifteen samples
gave a similar correlation of proteoglycan-degrading activity and protease activity with denatured haemoglobin as substrate, thus extending earlier observations (Wood et al., 1971).
Particle-bound nature of enzyme activity
When crude cell extracts (eleven samples) of widely different enzyme activities were centrifuged 7844% (mean 81%) of the proteoglycan-degrading activity and 7694% (mean 8 1%) of
the protease activity (hide-powder azure as substrate) sedimented in fraction P, which also con-
Granulocyte neutral protease
407
TABL~
1. Effectof inhibitors andactivatorson theproteoglycdgrading andprotease
activities of a crude cell extract
Methods of determining activities are given in the text.
Change of activity (A)
Proteoglycan-degrading
activity(')
p-Chloromercuribenzoate(2-5 mmol/l)
Iodoacetamide (25 mmol/l)
Cysteine (25 mmol/l)
EACA (25 mmol/)
EDTA (25 mmol/l)
DFP(3
(a)
0
Soybean trypsin inhibitor (50 &I)
Heparin (50 Unitslml)
Aprotinin (Trasylol) (100 units/ml)
~~
-29
Protease
activity(*)
-31
-40
-36
-32
-27
-29
-96
-82
-100
-68
-76
+21
+34
-32
-92
-82
-56
- 100
- 100
~
Determined with chondromucoprotein as substrate with 0-5 unit of proteoglycan-degrading activity/ml.
Determined with hidepowder azure as substratewith 20 units ofproteoglycandegrading activity/ml.
c3)DFP[(a) 1 mmol/l; (b) 0.1 m o l / l ] was allowed to react with enzyme (140
proteoglycan-degrading unitslml) for 1 h at 4°C.Mixtures were then diluted for
assay purposes.
e
N
0
3
0
I-
9,
100
200
300
400
500
Proteoglycan degrading activity (units/ml of extract
FIG.1. Relation between proteoglycan-degrading and protease activities of crude synovial fluid
Protease substrate was hidepowder azure.
cell extract (0)and sedimentable fraction P (0).
408
R.H . Pryce-Jones, J. Saklatvala and G. C. Wood
TABLE
2. Recovery ofprofeaseactivities of two cell extracis in fraciion P
Methods of assay are given in the text. Activities with hide-powder
azure and RBB-elastin are expressed as mg dissolved/h and with
dimethyl casein and calf thymus histone as nmol of amino end groups
releasedh.
Substrate
Hide-powder azur?
RBB-elastin
Dimethyl casein
Calf thymus histone
Activity in cell extract
before fractionation
Recovery in
fraction P (%)
548
178
10.2
10.0
7-13
3.02
20.3
13-8
86
77
93
88
88
85
80
77
PH
FIG.2. Effect of pH on proteolytic activity of crude synovial fluid cell extract (m) and sedimentable
fraction P ( 0 ) with hide-powder azure used as substrate. See the Materials and Methods section.
The buffers were: p H 3G6.0, citrate (0.1 mol/l); pH 6.5 and pH 7.0, phosphate (0.1 mol/l);
pH 7.5-85, Tris (0.1 mol/l); pH 9-105, glycine (0.1moI/l); pH 114-12.0, phosphate (0.1 mol/l).
All buffers contained KCl(O.2 mol/l).
Granulocyte neutral protease
409
tained 2140% (mean 29%) of the protein. The remaining activities and protein were recovered
in the supernatant (fraction S). The crude synovial fluid cell extracts also hydrolysed a number
of other protein substrates (Table 2) and in all cases most of the hydrolytic activity was recovered in fraction P. Fraction S was found to contain a complex mixture of enzymes (unpublished experiments) and since fraction P was much more active it was studied further.
Repeated sonication of suspensions of fraction P in Tris-HC1 buffer (0.1 mol/l) at pH 7.5
solubilized (i.e. did not sediment at 100000 g for 1 h) less than 2% of its enzyme activity or
protein content. The relationship between proteoglycan-degrading activity and protease
activity (hide-powder azure) of fraction P is the same as the relationship between these two
activities in crude cell extracts (Fig. 1). The pH-dependence of the protease activity of fraction
P resembles that of crude cell extracts with hide-powder azure (Fig. 2) or denatured haemoglobin (Wood et al., 1971) as substrates, with a peak of activity at pH 7-5-8.0, high activity at
about pH 10 and little activity below pH 6.0.
I
0
I
I
0.4
I
I
0.8
I
I
1.2
.-.
I
1.6
I
I
2.0
C o n c n . of KCL Im o l l L 1
FIG. 3. Effect of concentration of KCI on the proholytic activity of sedimentable fraction P.
Substrates were hide-powder azure (v), RBB-elastin (0)and urea-denatured haemoglobin (W).
Relative activity = activity in the presence of KCljactivity without KCI.
Solubility of fraction P
The rate of hydrolysis of insoluble protein substrates, hide powder and RBB-elastin, by
fraction P, increased markedly (up to thirteenfold) as the salt concentration of the reaction
mixtures was increased to 0.6 mol/l. The rate of hydrolysis of the soluble substrate denatured
haemoglobin, on the other hand, was relatively little affected by KCl concentration (Fig. 3).
To determine the effect of increasing concentration of KCI on the solubility of cell protease, the
mixtures (Fig. 3) containing 0.2,0.5 and 1 mol/l KC1 were centrifuged at 100OOO g for 1 h.
The supernatant solutions contained 47, 80 and 90% of the total activity respectively. Above
0.6 mol/l KCl, when more than 80% of the enzyme had been solubilized,the rates of hydrolysis
of hide-powder azure and the soluble substrate were almost independent of KC1 concentration.
410
R. H . Pryce-Jones, J. Saklatvala and G. C. Wood
For reasons which were not explored the rate of hydrolysis of RBB4astin decreased markedly
at high salt concentrations.
Synthetic substrates
Neither fraction P nor fraction S caused significant hydrolysis of benzoylarginine ethyl ester,
benzoylarginine B-naphthylamide, benzoylargininep-nitroanilide or tosylarginine methyl ester.
Fraction S, but not fraction P, hydrolysed leucine B-naphthylamide. Fraction S hydrolysed
benzoyltyrosine ethyl ester and acetyltyrosine ethyl ester. Fraction P hydrolysed these esters
only very slowly, possibly as a result of contamination with traces of fraction S. Only two of the
synthetic substrates were hydrolysed readily by fraction P, namely NBA and NCA. Typical
activities were: NBA, 1.74pmol hydrolysed min- mg- of protein; NCA, 1-44pmolhydrolysed
min-l mg-I of protein. Fraction S hydrolysed these compounds at about 5-10% of these
rates.
'
Inhibition of elastase-like activity by synovial fluid and serum
Synovial fluid and serum inhibited hydrolysis of hide-powder azure and the elastase substrates, RBB-elastin, NBA and NCA by fraction P. A preparation of fraction P, which had an
RBB-elastin-hydrolysing activity of 6.3 mg h-l mg-l of protein, was completely inhibited by
cell-free rheumatoid synovial fluids and normal sera which had been diluted 50-200-fold to
contain approximately 1.9 and 3-lg of protein/l respectively. A solution of pancreatic elastase
with RBB-elastin-hydrolysing activity of 9.9 mg h-' mg-I of protein was also completely inhibited under these conditions. The diluted synovial fluids and sera alone did not hydrolyse
RBB-elastin in these experiments. Similar concentrations, however, caused appreciable hydrolysis of NBA and NCA. Typical activities were: synovial fluid, 0-021 pmol of NBA min-'
mg-' of protein and 0.026 pmol of NCA min-' mg-I of protein; serum, 0.033 pmol of NBA
min-l mg-' of protein and 0.052 pmol of NCA min-l mg-I of protein. These results were
typical of twelve synovial fluids and three sera and agree with the findings of Janoff & Blondin
(1970), who attributed the activities to the elastase-likeprotease of granulocytes. When synovial
fluids and sera were diluted to concentrations at which their own NBA- and NCA-esterase
activities were less than 1% of the activity of fraction P, they still caused considerable inhibition of the hydrolysis of these substrates by fraction P (Table 3). It is concluded that the NBAand NCA-esterase activities of synovial fluid and fraction P are due to different enzymes.
Cartilage degradation
There was some loss of staining and release of hexuronate from cartilage incubated at pH 4.5
without added cell enzyme. This was probably due to chondrocyte cathepsin D (Dingle, 1969).
These effects were somewhat increased by fraction P. There was no significant loss of staining
or release of hexuronate from cartilage in buffer alone, at pH 7.5 and pH 9.0, but fraction P
caused complete loss of staining and considerable loss of hexuronate at these pH values. Cellfree synovial fluid caused no loss of staining or release of hexuronate at pH 7.5 or 9.0 but
strongly inhibited the effects of fraction P at these pH values. Table 4 gives the results on
hexuronate release.
Proteoglycan-degrading activity of granulocytes and lymphocytes of blood
Signiscant amounts of proteoglycan-degrading activity (16-39 units/mg of protein, mean
Granulocyte neutral protease
41 1
30 units) were observed in extracts of the granulocytes isolated from each of five 25 ml samples
of human blood, three of which were from rheumatoid patients and two from non-rheumatoid
patients. Less than 2 unitlmg of protein (mean 1-4 unitslmg of protein) were observed in extracts of the lymphocytes from the same samples. The activities of neutral protease (hidepowder assay) in two of these preparations were: granulocytes, 2-06 and 1.75 mg of substrate
dissolved h-’ mg-l of protein; lymphocytes, less than 0-025 mg of substrate dissolved h-’
mg-’ of protein.
TABLE
3. Inhibition of NBA- and NCA-esterase activities of the
sedimentable fraction P of cell extracts by dilutwns of celI-fiee
sytwvial fluidr and serum
Methods are given in the text.
Substrate
Activity of
fraction P
@no1 rnin-’
mg-t of protein)
NBA
0,193
NCA
0.073
NCA
0.073
Concentration of
synovial fluid
Inhibition
protein
(%I
(mg/l)
30.5
7.6
17.2
8.6
93.5
98
62
85
59
69
(serum)
4. Release of uronic acidfrom bovine articular cartilage slices by fraction P and
TABLE
inhibition by cell-free synovial fluid
The experimental procedure is given in the text. The amount of uronic released is
given per 60 mg of cartilage in 16 h at 37°C.
Inhibition by
synovial fluid (“A)
Uronic acid released (,umol)
pH
Control
Plus fraction P(l)
Plus fraction P and
synovial fluid(2’
4-5
7.5
9-0
1-00
0.09
0.74
1a70
3.17
Not done
008
0.80
0.31
53
90
Corrected for control values.
c2) Corrected for control values and for synovial fluid values.
This suggests that the proteoglycan-degrading activity from synovial fluid cells is likely to
derive primarily from the granulocytes. Attempts to separate the granulocytes and lymphocytes
from small quantities of synovial fluid, by the method used with blood, proved unsuccessful.
412
R. H. Pryce-Jones, J. Saklatvala and G . C. Wood
Indeed, difficulty was encountered in recovering cells completely by centrifuging some highly
viscous fluids. Probably as a result of this there was only a slight correlation between the total
proteoglycan-degrading activity from the cells of thirty samples of synovial fluid and either the
white blood cell count (Kendall's z = 0.37) or the granulocyte count (Kendall's z = 0.64)
of the fluids.
DISCUSSION
The electrophoretic mobility and staining characteristics of the degradation products of proteoglycan confirm and extend earlier findings with chondromucoprotein (Wood et al., 1971).
They show that the proteoglycan-degrading activity of crude synovial fluid cell extracts is due
to proteolysis and that hyaluronidase does not contribute significantly. This view is supported
by the close correlation between the proteoglycan-degrading activity and the proteolytic
activity (with two substrates) of a range of cell extracts (Fig. l), and by the similarity of the
pH-activity curves obtained with hide-powder azure as substrate, in the present work (Fig. 2),
and with denatured haemoglobin and chondromucoprotein as substrates in earlier work (Wood
et al., 1971). The qualitative similarity of the effects of various activators and inhibitors
(Table l), including serum and cell-free synovial fluids, on the proteoglycan-degrading activity
and protease activity of cell extracts, further support the conclusion that proteoglycandegrading activity is primarily proteolytic. Little significance can be attached to quantitative
comparison of the effects of the activators and inhibitors (Table l), in view of the different
conditions of the two enzyme assays, e.g. enzyme/substrate ratio and the polyanionic nature of
the proteoglycan.
The proteolytic activity of cell extracts with purified proteoglycan as substrate shows that the
activity previously observed with chondromucoprotein as substrate was probably due to direct
action on the proteoglycan protein itself rather than on other proteins responsible for interactions between proteoglycan molecules.
Granulocytes and lymphocytes comprise about 94% of the cells found in rheumatoid synovial
fluids, granulocytes being the predominant type (60-90%). The properties of fraction P of
synovial fluid cell extracts, which contains most of the proteoglycan-degrading activity, are
similar to those of the neutral protease extracted from human blood granulocytes by other
workers (Mounter & Atiyeh, 1960; Ziff et al., 1960; Ohlsson, 1971; Janoff, 1972; Janoff &
Blondin, 1970; Folds, Welsh & Spitznagel, 1972; Leiberman & Kaneshiro, 1972). Thus
activities towards synthetic substrates are similar, as also are responses to activators and
inhibitors.
Fraction P was insoluble in Tris-HC1 buffer (0.1 mol/l) at pH 7.5. In this respect and in
substrate specificity fraction P appears to be quite distinct from the mixture of hydrolases
present in the soluble fraction of cell extracts (fraction S). Fraction P became progressively
more soluble as increasing concentrations of KC1 were added to the buffer and this probably
accounted for the concomitant increase in the rate at which insoluble protein substrates were
hydrolysed (Fig. 3), for solubilization of the enzyme is expected to facilitate approach of the
enzyme to the substrate. Solubilization would be expected to have relatively little effect on the
rate of hydrolysis of soluble protein substrates and the rate of hydrolysis of denatured haemoglobin was in fact little affected by KCI concentration (Fig. 3).
Neutral to alkaline proteases from rabbit granulocytes and rat muscle have also been found
Granulocyte neutral protease
413
to be insoluble at low ionic strength but soluble at high concentrations of KCl (1-2-5 mol/l)
(Davies et al., 1971; Holmes, Parsons, Park & Pennington, 1971). Leiberman & Kaneshiro
(1972) reported that elastase-like esterase activity can be extracted from human purulent
sputum leucocytes by NaCl(1 mol/l) but not by water alone and suggested that solubilization
of the enzyme probably accounted for their results. Despite the insolubility of the synovial
fluid cell proteoglycan-degrading activity at low ionic strength, our results show that the
enzyme can cause release of glycosaminoglycans from cartilage at salt concentrations not far
removed from that of synovial fluid (150 mmol/l).
Throughout this work attempts were made to detect acid-protease activity in cell extracts
with three substrates, but these were uniformly unsuccessful. The small amount of uronic
acid released from cartilage as a result of prolonged incubation with relatively high concentrations of fraction P at pH 4.5 (Table 4) may be due to a trace of acid protease. Folds et al.
(1972) did not detect acid protease in human leucocytes but others (Mounter & Atiyeh, 1960;
Stiles & Fraenkel-Conrat, 1968; Ziff et al., 1960) detected small amounts. However, these
appear to have been smaller than in similar preparations from rabbits (Weissmann & Spilberg,
1968; Cohn & Hirsch, 1960; Wasi, Murray, Macmorine & Movat, 1966) and pigs (Lebez 8c
Kopitar, 1970). Comparison is difficult in view of the variety of extraction methods and assay
procedures that have been used but a recent comparative study (Oronsky et d.,1973) suggested
that human granulocytes contain considerably more proteolytic activity at pH 7.4 than at
pH 5.0 and that the reverse is true for similar preparations from guinea-pig or rabbit.
ACKNOWLEDGMENTS
We gratefully acknowledge financial assistance from the Arthritis and Rheumatism Council
and from Organon Laboratories Ltd. We also thank F.B.A. Pharmaceuticals and HoechstCastella Ltd for gifts of chemicals.
REFERENCES
BITTER,
T. & MUIR,H.M. (1962)A modified carbazole method for uronic acid determination. Analytical
Biochemistry, 4,330-334.
BLACKWOOD,
C. & MANDL,
I. (1961)An improved test for the quantitative determination of trypsin, trypsinlike enzymes and enzyme inhibitors. Analytical Biochemistry, 2,370-379.
WYUM,
A. (1968)Separation of leukocytes from blood and bone marrow: Isolation of mononuclear cells and
granulocytes from human blood :Isolation of mononuclear cells by zone centrifugationand of granulocytes
by combining centrifugation and sedimentation at lg. Scandinavian Journal of Clinical and Laboratory
Invest&ation, Supplement, !37,77-89.
CQHN,Z.A. & HIRSCH,J.G. (1960)The isolation and properties of the specific granules of rabbit polymorphonuclear leukocytes. Journal of Experimental Medicine, 112,983-1009.
DAVIES,
P.,RITA,G.A., KRAKAUER,
K. & WEISMANN,
G. (1971)Characterizationof a neutral protease from the
lysosomes of rabbit polymorphonuclear leucocytes. Biochemical Journal, 123,559-569.
DAVIS,B.J.(1964)Disc electrophoresis. 11. Method and application to human serum proteins. Annals of’theNew
York Academy of Sciences, 121,404-427.
DINGLE,
J.T.(1969)The extracellular secretion of lysosomal enzymes. In: Lysosomes in Biology and Pathobgy,
Vol. 2,pp. 421-436. MS.
Dingle, J.T. & Fell, H.B. North-Holland Publishing Company, Amsterdam and
London.
ERLANGJX,
B.F., KAKOWSKY, M. & COHEN, W. (1961)Preparation and properties of two new chromogenic
substrates of trypsin. Archives of Biochemistry and Biophysics, 95,271-278.
414
R. H . Pryce-Jones, J. Saklatvala and G. C. Wood
FOLDS,
J.D., WELSH,I.R.H. & SPITZNAGEL,
J.K. (1972)Neutral proteases confined to one class of lysosomes of
human polymorphonuclear leukocytes. Proceedings of the Society for Experimental Biology and Medicine,
139,461-463.
HOLMES,
D., PARSONS,
M.E., PARK,
D.C. & PENNINGTON,
R.J. (1971)An alkaline proteinase in muscle homogenates. Biochemical Journal, 125, 98~.
H ~ M E B.C.W.
L,
(1959)A modified spectrophotometricdetermination of chymotrypsin, trypsin and thrombin.
CanadianJournal of Biochemistry and Physiology, 37,1393-1399.
JANOFF, A. (1969)Alanine pnitrophenyl esterase activity of human leucocyte granules. Biochemical Journal,
114,157-159.
JANOFF, A. (1972)Human granulocyte elastase. American Journal of Pathology, 68,579-591.
JANOFF,
A. & BLONDM,J. (1970)Depletion of cartilage matrix by a neutral protease fraction from human
leukocyte lysosomes. Proceedings of the Society for Experimental Biology and Medicine, 135,302-306.
KUNITZ,M. (1947)Crystalline soybean trypsin inhibitor. 11.General properties. Journal of General Physiology,
30,291-310.
LASKOWSKI,
M. (1955)In: Methods in Enzymology, Vol. 2, pp. 26-36. Eds. Colowick, S.P. & Kaplan, N.O.
Academic Press, New York.
LEBEZ,D. & KOPITAR,
M. (1970)Leukocyte proteinase. I. Low molecular weight cathepsins of F and G types.
Enzymologia, 39,271-283.
LEIBERMAN,
J. & KANESHIRO,
W. (1972) Inhibition of leukocyte elastase from purulent sputum by alpha-lantitrypsin. Journal of Laboratory and Clinical Medicine, 80,88-101,
LIN,Y., MEANS,G.E. & FEENEY,
R.E.(1969)The action of proteolytic enzymes on NN-dimethylated proteins.
Basis for a micro assay for proteolytic enzymes. Journal of Biological Chemistry, 244, 789-793.
MALAWISTA,
I. & SCHUBERT,
M. (1958)Chondromucoprotein:New extractionmethod and alkaline degradation.
Journal of Biological Chemistry, 230,535-544.
MILLER,G. (1959)Protein determinations using Folin and Ciocalteu's reagent. Analytical Chemistry, 31,964.
MOUNTER,
LA. & ATIYEH,
W.(1960)Proteases of human leukocytes. Blood, 15,52-59.
OHLSSON,
K. (1971) Neutral leukocyte proteases and elastase inhibited by alpha-1-antitrypsin. Scandinavian
Journal of Clinical and Laboratory Investigation, 28,251-253.
ORONSKY,
A.L., IGNARRO,
L.J. & PERPER,
R.J. (1973) An in uitro model for estimating chondromucoprotein
degradation by human polymorphonuclear leukocytes utilising 35S labelled cartilage. Federation Proceedings, 323,292 (Abstract no. 429).
PAL,S., DOGANGES,
P.T. & SCHUBERT,
M. (1966) The separation of new forms of protein-polysaccharide of
bovine nasal cartilage. Journal of Biological Chemistry, 241,42614266.
PARTRIDGE,
S.M. & DAVIS,H.F.(1955)The chemistry of connective tissue. 3. The composition of the soluble
proteins derived from elastin. Biochemical Journal, 61,21-30.
PARTRIDGE,
S.M.,DAVIS,H.F. & ADAIR,G.S. (1961)The chemistry of connective tissue. 6. The constitution of
the chondroitin sulphate-protein complex in cartilage. Biochemical Journal, 79, 15-26.
RINDERKNECHT,
H.,GEOKAS,
M.C., SILVERMAN,
P., LILLARD,
Y. & HAVERBACK,
B.J. (1968)New methods for
determination of elastase. Clinica Chimica Acta, 19, 327-339.
RINDERKNECHT,
H.,SILVERMAN,
P., GEOKAS,
M.S. & HAVERBACK,
B.J. (1970)Determination of trypsin and
chymotrypsin with RemazoI briIliant blue hide. CZinica Chimica Ada, 28,239-246.
SHOTTON,
D.M. (1970)In: Methods in Enzymology, Vol. 19, pp. 113-140. Ed. Perlmann, G.E. & Lorand, L.
Academic Press, New York.
STILES,
M.A. & FRAJNCEL-CONRAT,
J. (1968)The subcellular distribution of human leukocyte cathepsins. Blood,
32,119-133.
WASI,S., MURRAY,
R.K., MACMORINE,
D.R.L. & MOVAT,
H.Z. (1966) The role of polymorphonuclear leukocyte lysosomes in tissue injury, intlammation and hypersensitivity. XI. Studies on proteolytic activity of
polymorphonuclear leukocyte lysosomes of rabbit. British Journal of Experimental Pathology, 47,411423.
WHWANN,G. & SPILBERG,
I. (1968)Breakdown of cartilage protein-polysaccharide by lysosomes. Arthritis
and Rheumatism, 11,162-169.
WOOD,G.C., PRYCE-JONES,
R.H., WHITE,D.D. & NUKI, G. (1971) Chondromucoprotein-degrading neutral
protease activity in rheumatoid synovial fluid. Annals of the Rheumatic Diseases, 30,73-77.
ZIFF,M., GRIBETZ,
H.J. & LOSPALLUTO,
J. (1960)The effect of leukocyte and synovial membrane extracts on
cartilage mucoprotein. Journal of Clinical Investigation, 39,405-412.