Biochem. J. (1987) 246, 599-603 (Printed in Great Britain)
Collagen activates superoxide anion production by human
polymorphonuclear neutrophils
Jean-Claude MONBOISSE,* Georges BELLON,* Jean DUFER,t Alain RANDOUX* and Jacques-Paul BOREL*
*Laboratory of Biochemistry, CNRS UA 610, UFR Medicine, 51 rue Cognacq Jay, F 51095 Reims, and tLaboratory of
Hematology, Institut J. Godinot, 45 rue Cognacq Jay, F 51100 Reims, France
Human polymorphonuclear neutrophils (PMNs), purified on Ficoll-Hypaque cushions, were incubated for
5 min with calf skin acid-soluble collagen and the released superoxide anions (02) measured spectro-
photometrically by reduction of ferricytochrome c or by chemiluminescence analysis. This collagen
stimulated the release of 02- unless it had been treated with pepsin. The stimulatory activity remained in
denatured collagen, was contained only in the a l(I) chain and was present in the aI1(I)-CB 6 (CNBr-cleaved)
peptide, which is C-terminal. The activity was linearly dependent on the collagen concentration up to about
200 /tg/ml. In addition, this collagen induced a release offl-glucuronidase and N-acetyl-,-glucosaminidase
from PMNs.
INTRODUCTION
Native soluble collagen and collagen degradation
products are chemotactic for PMNs in vivo (Chang &
Houck, 1970; Riley et al., 1984). Laskin et al. (1986)
showed recently that low-Mr synthetic polypeptides
(consisting oftriplet units ofproline, hydroxyproline and
glycine), bovine dermal collagen digested with bacterial
collagenase and type I calf skin collagen fragments
produced with CNBr are all potent chemoattractants for
human peripheral-blood neutrophils. As many factors,
such as complement components C3a and C5a, ara-
chidonic acid and bacterial polypeptides, are both
chemotactic factors and activators for the generation of
02- by PMNs, type I collagen itself should induce the
formation of oxygen free radicals by PMNs. We report
here that bovine acid-soluble collagen and one of the
peptides obtained on its digestion with CNBr do trigger
superoxide anion (02) formation by human PMNs
in vitro. This activation varies with the collagen concen-
tration, is modulated by the extracellular Ca2+ concen-
tration and is located in a sequence of the al(I) chain
comprising a helical segment and the C-terminal
telopeptide.
MATERIALS AND METHODS
Reagents
Ferricytochrome c (type VI), FMLP, SOD from
bovine erythrocytes, NEM, PMSF and p-nitrophenyl
fl-D-glucuronide were purchased from Sigma Chemical
Co. (St. Louis, MO,-U.S.A.).p-Nitrophenyl 2-acetamido-
2-deoxy-fl-D-glucopyranoside and 3-aminophthalhydra-
zide (luminol) were bought from Aldrich (Strasbourg,
France).
Acetonitrile (h.p.l.c. grade) was obtained from BDH
Chemicals (Poole, Dorset, U.K.). Heptafluorobutyric
acid (Sequenal grade) was purchased from Pierce
Chemical Co. (Rockford, IL, U.S.A.) and trifluoroacetic
acid (Uvasol grade) was from Merck (Darmstadt,
Germany). Deionized water was further purified by
filtration through a MilliQ water system (Millipore
Corp., Bedford, MA, U.S.A.). All the usual reagents
(analytical grade) were obtained from Prolabo (Paris,
France) or Merck (Darmstadt, Germany).
PBS (Ca2+-free) and PBS containing 1.3 mM-CaCl2
were prepared as described by Curnutte et al. (1979).
FMLP was dissolved at a concentration of 1 mg/ml in
DMSO. The stock solution was stored at -20 °C and
further diluted in PBS before use. Cytochrome c was
dissolved in PBS (final concn. 1.0 mmol/ml) and SOD
was diluted in PBS to a concentration of 0.1 mg/ml,
corresponding to an activity of60 units/ml; units ofSOD
are defined as described by McCord & Fridovich (1969).
Preparation of collagens
Acid-soluble collagen was obtained from calf skin by
extraction with 0.5 M-acetic acid and then precipitation
with 0.7 M-NaCl at acidic pH by the method of Piez
et al. (1963).
The a-chains from calf skin acid-soluble (non-pepsin-
treated)collagen were purified by gel-filtration chromato-
graphy on Agarose A5M (Bio-Rad, Richmond, CA,
U.S.A.) in 0.05 M-Tris/HCI buffer, pH 7.4, containing
1 M-CaCl2. The fraction containing the a-chains was
further chromatographed on CM-cellulose under de-
naturing conditions (Pontz et al., 1973) in order to
prepare pure acl(I) and a2(I) collagen chains. The TCA
and TCB fragments obtained in native form by tadpole-
collagenase digestion ofbovine type I collagen were a gift
from Dr. Y. Nagai (Medical Research Institute, Tokyo,
Japan).
CNBr cleavage ofa1 (I) chains prepared from calf skin
acid-soluble collagen was done by the method of Epstein
Abbreviations used: CM-cellulose, carboxymethylcellulose; CB-peptide, CNBr-cleaved peptide; DMSO, dimethyl sulphoxide; NEM, N-ethyl-
maleimide; FMLP, N-formylmethionyl-leucylphenylalanine; PMSF, phenylmethanesulphonyl fluoride; PBS, phosphate-buffered saline
(135 mm-NaC1/5 mM-KCl/lI mM-potassium phosphate, pH 7.4); PMN, polymorphonuclear neutrophil; PAGE, polyacrylamide-gel electrophoresis;
SOD, superoxide dismutase; O2- superoxide anion; TCA and TCB, tadpole-collagenase fragments A and B.
Vol. 246
599
J.-C. Monboisse and others
et al. (1971) in 70% (v/v) formic acid under an N2
atmosphere for 4 h. The resulting CB-peptides were
thoroughly dialysed against distilled water. The re-
maining non-diffusible peptides were separated by
reverse-phase h.p.l.c. in a system comprising two pumps
(Altex A 110), a gradient programming system controlled
by an Apple Ile computer and a Rheodyne 7125
injection valve, all bought from Touzart et Matignon
(Paris, France). The column was an Aquapore RP 300
C18 (particle size 10,um) (4.6 mm x 250 mm) from
Brownlee Laboratories (Santa Clara, CA, U.S.A.). The
eluate was monitored at 214 nm by a model SPD-2A
spectrophotometer (Shimadzu, Kyoto, Japan). The
method was a modification of that described by Van der
Rest & Fietzek (1982).
The purity of collagen fractions and a-chains was
checked by PAGE performed by the method of Laemmli
(1970) in slab gels containing 5% (w/v) polyacrylamide.
The CB-peptides were characterized by the same
technique, with the percentage of polyacrylamide raised
to 12.5 (w/v).
Acid-soluble collagen was submitted to limited pepsin
digestion, at an enzyme/substrate ratio of 1:1000 (w/w)
in 0.5 M-acetic acid at 4 °C for 18 h, by the method of
Chung et al. (1974). The pepsin was then inactivated by
dialysis against PBS. Denatured acid-soluble collagen
was obtained by heating the collagen solution at 60 °C
for 45 min. Several preparations ofacid-soluble collagen
were also delipidated by extraction with chloro-
form/methanol (2:1, v/v) at 4 °C for three periods of
24 h each.
Acid-soluble collagen was degraded with Clostridium
histolyticum collagenase (CLSPA; Worthington, Free-
hold, NJ, U.S.A.) purified by the method of Peterkofsky
& Diegelman (1971) and treated with proteinase
inhibitors (NEM and PMSF). Collagenase was then
removed by ultrafiltration with microcollodion bags
(Sartorius, Gottingen, Germany). The amounts of
collagen hydrolysates added to test tubes were calculated
so as to yield the same hydroxyproline concentration as
in whole collagen.
Preparation of the neutrophil suspension
Human PMNs were purified by the method of
Markert et al. (1984), with minor modifications. As
described by Homan-Muller et al. (1975), the con-
taminating erythrocytes were lysed by a hypo-osmotic
buffer (10 mM-KHCO3, 15 mM-NH4Cl, 0.14 mm-EDTA,
pH 7.2). A suspension of 107 PMNs/ml in balanced
Hanks solution was used for the experiments. It
contained 96+4% PMNs, with a viability greater than
95% as assessed by Trypan Blue exclusion. The
contaminating cells were lymphocytes.
Assessment Of 02- released by ferricytochrome c
reduction
02- release was measured by the method of English
et al. (1981). The cell suspension (100 /sl) was prewarmed
to 37 °C for 5 min and transferrred to 13 mm x 100 mm
glass tubes containing 0.85 ml of PBS and 0.1 ml of
cytochrome c solution. Activation was started by adding
0.1 ml of solution ofthe stimulating agent, either FMLP
(final concn. 0.5 ,M) or collagen (final concn.
12.5-400 jug/ml). For multiple determinations, the re-
quired numbers of replicates (five) were prepared,
containing identical quantities of every reagent. As a
control in each experimental category, SOD (50 1) was
added to one ofthe test tubes before the addition ofcells.
This was utilized as the spectrophotometric blank. At
5 min after the start of incubation, SOD (50,ul) was
added to stop the reduction of cytochrome c by °2- in
all except the control tubes. The tubes were then cooled
to 4 °C and centrifuged. Reduced cytochrome c was
assayed in supernatants by measuring the A549.5 in a
Beckman DU 40 spectrophotometer against the SOD-
inhibited blank. 02- concentrations were calculated by
using an cm. value of 15.5 mm-" cm-' (ferrocytochrome
c minus ferricytochrome c) according to the method of
Margoliash & Frohwirt (1959) and expressed in mmol of
02-/5 min per 106 cells.
Assessment of 02- release by luminol-dependent
chemiluminescence analysis
PMN chemiluminescence assays were performed by
the method of Tenner & Cooper (1982). Portions of
PMN suspension (0.1 ml) containing 106 cells, 0.1 ml of
collagen suspension and 0.02 ml of 1 mM-luminol
solution in 0.1 M-NaOH were added to 0.78 ml of
PBS/0.02 M-Hepes/0.25% bovine serum albumin in
6 ml polypropylene Picovials (Packard, Rungis, Paris,
France). Chemiluminescence was monitored every
0.5 min for 30 min in a Packard 4430 liquid-scintillation
counter. Chemiluminescence values were expressed as
c.p.m. per 106 cells after subtraction of background
chemiluminescence.
Enzyme release from PMNs
Portions (0.1 ml) of acid-soluble collagen solution
were added to 0.2 ml of PMN suspension containing 106
cells in PBS supplemented with 1.3 mM-CaCI2 and
incubated for 20 min at 37 'C. The tubes were rapidly
cooled, centrifuged at 800 g for 10 min, and the enzyme
activities measured in supernatant and in the remaining
cell pellet, repeatedly frozen and thawed.
Lactate dehydrogenase was measured by the method
described by Buhl et al. (1978), 8-glucuronidase and
N-acetyl-fl-glucosaminidase respectively with
p-nitrophenyl fl-D-glucuronide and p-nitrophenyl
2-acetamido-2-deoxy-/J-D-glucopyranoside as substrates
(Troost et al., 1976). Results (means of four determi-
nations) were expressed as percentages of the total
(cell +supernatant) enzymic activity.
RESULTS
Acid-soluble collagen stimulation of PMNs
The contact of PMNs with acid-soluble collagen
molecules induced the production of 02-, which,
monitored by the ferricytochrome c reaction, was linear
for 8 min (Fig. 1). The amount of 02- formed in
response to acid-soluble collagen at 0.3,umol/I was
comparable with that elicited by 0.5,1M-FMLP
(11.85+0.85 as against 12.55 +0.57 nmol of 02-75 min
per 106 cells). Pepsin-treated collagen was devoid of
activity.
The generation of 02 by PMNs increased with the
concentration of non-pepsin-treated collagen up to a
maximum at 200,ug/ml. Heat-denatured collagen had
the same effect (Table 1). At concentrations over
200,ug/ml the collagen molecules began to precipitate
1987
600
Collagen activates 02- production by neutrophils
into fibrils. Activation was more efficient when Ca2+ at
1.3 mmol/l was added to the solution.
The luminol-dependent chemiluminescent response of
PMNs in the presence of 0.3 ,uM acid-soluble collagen is
shown in Fig. 2.
Enzyme release from PMN under collagen stimulation
The percentage of activity of lactate dehydrogenase
released in supernatant was unchanged after stimulation
0.4
0.2
0.1
0 4 6
Time (min)
Fig. 1. Continuous monitoring of 02 production by collagen-
activated PNINs evaluated by ferricytochrome c
reduction
*, Non-activated PMNs; *, pepsin-treated acid-soluble
collagen (0.3 pM); V, acid-soluble collagen (0.3 gM); A,
FMLP (0.5 /zM). Collagen or FMLP solution (0.1 ml) was
added to the incubation medium. The reduction of
ferricytochrome c was monitored at 549.5 nm against a
spectrophotometric blank supplemented with SOD before
the addition of the cell suspension.
by acid-soluble collagen, whereas percentages of ,-
glucuronidase and N-acetylglucosaminidase were sig-
nificantly increased (Table 2).
Determination of the active fragment of acid-soluble
collagen
Heat-denatured and delipidated calf skin acid-soluble
collagens were active. They triggered roughly the same
amounts of 02- as did the non-pepsin-treated calf skin
collagen. Both pepsin digestion and bacterial-collagenase
treatment suppressed the induction of 02- production.
When the fragments TCA and TCB prepared from type
I collagen by the action of tadpole collagenase were
tested, the TCB fragment was active, whereas the TCA
fragment was not. The alI chain purified from acid-
soluble collagen was active, and the a2 chain was not.
The non-diffusible CB-peptides induced the production
ofeven more 02- than did the collagen itself(19.65 + 2.09
as against 1 1.98 + 0.85 nmol of 02-/5 min per 106 cells).
40
E
6.
Q 30
c
CD
E 20
3
.E
6-)
x
n 10
0
0
Time (min)
Fig. 2. Luminol-dependent chemiluminescent response of PMNs
activated by collagen
*, Non-activated PMNs; *, pepsin-treated acid-soluble
collagen (0.3 /tM); V, acid-soluble collagen (0.3 /eM); A,
FMLP (0.5 #M).
Table 1. Dose-dependent 2- generation by neutrophils activated
extracellular calcium
Results are means + 1 S.D. (n = 4).
by acid-soluble collagen in the presence or in the absence of
Production of °2- (nmol/5 min per 106 cells)
Concn. of collagen Acid-soluble collagen Gelatin Acid-soluble collagen Gelatin
or gelatin (mg ml-') without Ca2+ without Ca2+ with Ca2+, 1.3 mM with Ca2+, 1.3 mM
0
12.5
25
50
100
200
300
400
1.27+0.73
1.76+0.24
2.63+0.19
5.01 +0.25
8.63+0.34
10.23 +0.22
2.05+0.41
1.84+0.31
1.27+0.73
1.57 +0.25
1.86+0.19
3.06+0.15
7.30 +0.53
9.18+0.32
8.23+0.63
8.31 +0.33
1.74+0.52
2.40+0.21
3.79 +0.23
6.15+0.10
10.51 +0.50
15.12+0.25
4.45+0.26
2.56+0.39
1.74+0.52
2.10+0.47
2.76+0.10
5.55 +0.34
10.54+0.58
15.48+0.56
12.69 +0.44
11.90+0.96
Vol. 246
601
J.-C. Monboisse and others
Table 2. Enzyme activities released in the supernatant after activation of PMNs by collagen, expressed as percentages of total
(cell+supernatant) activity-
Differences from non-activated PMNs significant for: *P < 0.01; **P < 0.001. NS, not significant.
Lactate N-Acetyl-fl-
dehydrogenase fl-Glucuronidase glucosaminidase
Non-activated PMN
FMLP (0.5 1uM)
Acid-soluble collagen
(0.3 /SM)
Pepsin-treated acid-
soluble collagen (0.3 ,UM)
CB-peptides from
acid-soluble collagen
(100 jug/ml)
3.12+ 1.15
4.82 +0.92 NS
2.25 + 1.90 NS
5.13 +0.77 NS
2.67+1.19 NS
3.80+2.27
12.75 + 1.63*
17.50+ 3.11*
2.75 + 1.20 NS
16.35 +0.35**
5.98 +0.45
11.31 +0.83**
16.34+ 1.15**
5.46+0.63 NS
12.79 + 0.87**
al(I)-CB 7
al(I)-CB 6
al(M)-CB 3
T 1 2 3 4 5 6 7
ma
a b
Fig. 3. SDS/PAGE of the non-diffusible CB-peptides from al(l) chain: separation by two successive reverse-phase h.p.l.c. steps
T, non-diffusible CB-peptides of the collagen al(I) chain; a, lanes 1, 2 and 3, material from the three peaks of the first h.p.l.c.
step: b, lanes 4, 5, 6 and 7, material from the four peaks of the second h.p.l.c. step. The fractions in lanes 2 and 5 activated
02- production by PMNs. Lane 6 induces a very slight activation, owing to contamination.
We isolated the active CB-peptide by reverse-phase
h.p.l.c. of the non-diffusible CB-peptides. The first step,
in trifluoroacetic acid, separated three peaks; the activity
was contained in the second one. SDS/PAGE showed
that this peak contained peptides l1(I)-CB 6 and
a-1(I)-CB 7 plus some higher-Mr components (Fig. 3, a).
The active peak was rechromatographed by reverse-phase
h.p.l.c. in the same column, but this time in hepta-
fluorobutyric acid, four fractions being separated. The
02--stimulating activity was found in the second fraction,
which contained the peptide cc1(I)-CB 6 (Fig. 3, b).
DISCUSSION
The activation of PMNs results in several -distinct
morphological and metabolic events, all leading to
-phagocytosis and degradation of foreign substances.
Among them is the formation and secretion of oxygen
free radicals contemporaneous with a sudden rise in the
consumption ofoxygen, known as the 'respiratory burst'
and triggered by many stimuli. The liberation of 02- iS
easily monitored by a technique using ferricytochrome c
and-by measurement of chemiluminescence.
The family of collagen molecules, particularly of the
interstitial type such as type I, is known for its
chemoattractant properties (Chang & Houck, 1970;
Riley et al., 1984; Laskin et al., 1986). Most other
chemoattractant substances, such as FMLP, also stimu-
late the activation of PMNs. Some authors (Hoffstein
et al., 1985) have suggested that any effect of collagen
as trigger of the PMN respiratory burst could be
ruled out on the basis of experiments conducted with
pepsin-treated collagen. However, we show here that
acid-soluble collagen, or its main constituent, type I
1987
602
........
...........
Collagen activates °2- production by neutrophils 603
collagen, when purified without the use of pepsin, is
actually as efficient an activator as FMLP. This
stimulatory effect depends on the presence of the
telopeptides of collagen, as demonstrated by its disap-
pearance when acid-soluble collagen had been treated
with pepsin. The effect is dose-dependent up to a limit
which appears to correspond to the collagen concen-
tration above which the fibrils precipitate spontaneously.
In addition to this precipitation, large amounts of
collagen serve as scavengers for 02- (Monboisse et al.,
1986).
It was found that only the al1(I) chain of collagen was
responsible for the effect. Denatured collagen was still
active, suggesting that the primary structure of the
molecule was responsible for the activity. The telopeptide
sequence liberated by treatment of collagen by purified
bacterial collagenase was not capable of triggering the
activation, but the TCB peptide was. In order to locate
the active region more precisely, we fragmented the
collagen molecule with CNBr and assayed the resulting
CB-peptides for their ability to stimulate PMN. A
modified h.p.l.c. method of CB-peptide separation (Van
der Rest & Fietzek, 1982) was used to prepare the
peptide ocl(I)-CB 6 and purify it from the contaminating
al(I)-CB 3 and al(I)-CB 7 peptides. al(I)-CB 6 was the
only active CB-peptide.
The location ofthe active region at the C-terminal end
of the type I collagen molecule may be fairly close to the
region that controls platelet binding (Fauvel et al., 1978),
but the latter site remained active after pepsin treatment
of the collagen. In addition, Morton & Barnes (1986)
recently demonstrated the presence of several other
platelet-activating regions in the collagen molecule.
How type I collagen activates PMNs is not yet known.
It seems to trigger 02- and lysosomal enzyme liberation
simultaneously, but it remains to be demonstrated
whether there is an effect on the formation of
eicosanoids.
The physiological significance of this effect may be of
interest with regard to inflammation. The contact of
PMNs with altered connective tissues would trigger the
degradation reactions characteristic of the first stages of
any inflammation reaction. It would provide a link
between the accumulation of PMNs in inflamed tissues,
owing to the chemotactic effect of several molecules
(among them products of degradation of collagen) and
the cleaning effect of oxygen free radicals produced by
these PMNs.
We are indebted to Professor Y. Nagai (Medical Research
Institute, Tokyo, Japan), who kindly provided us with samples
of TCA and TCB fragments. This work was supported by
grants from the University of Reims and from the CNRS (UA
610). We are grateful to Mrs. R. Platzek and Mrs. C. Leroux
for technical assistance and Mrs. S. Etienne and Mrs.
M. Debref for typing the manuscript.
REFERENCES
Buhl, S. N., Jackson, K. Y. & Graffunder, B. (1978) Clin.
Chem. 24, 261-266
Chang, C. & Houck, J. C. (1970) Proc. Soc. Exp. Biol. Med.
134, 22-26
Chung, E., Kelle, E. M. & Miller, E. J. (1974) Biochemistry 13,
3459-3464
Curnutte, J. T., Babior, B. M. & Karnovsky, M. L. (1979) J.
Clin. Invest. 63, 637-647
English, D., Roloff, J. S. & Lukens, J. N. (1981) J. Immunol.
126, 165-171
Epstein, E. H., Scott, R. D., Miller, E. J. & Piez, K. A. (1971)
J. Biol. Chem. 246, 1718-1724
Fauvel, F., Legrand, Y. J. & Caen, J. P. (1978) Thromb. Res.
12, 273-285
Hoffstein, S. T., Gennaro, D. E. & Manzi, R. M. (1985) Lab.
Invest. 52, 515-522
Homan-Miiller, J. W., Weening, R. S. & Ross, D. (1985) J.
Lab. Clin. Med. 85, 198-207
Laemmli, U. K. (1970) Nature (London) 227, 680-685
Laskin, D. L., Kimura, T., Sakakibara, S., Riley, D. J. & Berg,
R. A. (1986) J. Leukocyte Biol. 39, 255-266
Margoliash, E. & Frohwirt, N. (1959) Biochem. J. 71, 570-572
Markert, M., Andrews, P. C. & Babior, B. M. (1984) Methods
Enzymol. 105, 358-365
McCord, J. M. & Fridovich, I. (1969) J. Biol. Chem. 244,
6049-6055
Monboisse, J. C., Gardes-Albert, M., Randoux, A., Ferradini,
C. & Borel, J. P. (1986) in Superoxide and Superoxide
Dismutase (Rotilio, G., ed.), pp. 53-55, Elsevier,
Amsterdam
Morton, L. F. & Barnes, M. J. (1986) Proc. FECTS Meet. 8th,
Manchester, 137
Peterkofsky, B. & Diegelman, R. (1971) Biochemistry 10,
988-994
Piez, K. A., Eigner, E. A. & Lewis, M. S. (1963) Biochemistry
2,58-66
Pontz, B. F., Muller, P. K. & Heigel, W. N. (1973) J. Biol.
Chem. 248, 7558-7564
Riley, D. J., Kerr, J. S., Laskin, D. S., Guss, H. N., Curran,
S. F. & Berg, R. A. (1984) Clin. Res. 32, 563A (abstr.)
Tenner, A. J. & Cooper, N. R. (1982) J. Immunol. 128,
2547-2552
Troost, J., Van der Heijden, M. C. M. & Staal, G. E. J. (1976)
Clin. Chim. Acta 73, 329-346
Van der Rest, M. & Fietzek, P. P. (1982) Eur. J. Biochem. 125,
491-496
Received 24 November 1986/27 April 1987; accepted 28 May 1987
Vol. 246