From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
Release of Elastase From Monocytes Adherent
to a Fibronectin-Gelatin Surface
By D.L. Xie, I?. Meyers, and G.A. Homandberg
Fibronectin (Fn) is a circulatingand extracellular matrix glycoprotein that may serve to facilitate phagocytosis because
of its ability to bind many inflammatory ligands and to a
monocyte receptor. Fn fragments have been shown in many
systems to have augmented propertiesover those of native
Fn. We show in this report that although Fn fragments did
not cause elastase release from monocytes in suspension,
fragments did cause elastase release from monocytes that
were first bound to Fn-gelatinsurfaces. An amino-terminal
29-Kd and a 140-Kd integrin-bindingfragment were halfmaximally active at 1 0 0 nmol/L, whereas the Arg-Gly-Asp-
P
LASMA FIBRONECTIN (Fn) is a circulating Fn isoform
that can bind to integrin receptors on monocyte cell
surfaces' and that, as an opsonin, may present various types
of macromolecules to monocytes for ingestion.2 Interaction
of monocytes with Fn surfaces may be an important regulator
of inflammatory properties because monocyte interactions
with Fn surfaces have been shown to augment the activities
of the complement receptor^^.^ and the Ig Fc receptors' and
to cause expression of characteristics of inflammatory macrophages, such as protease release.'
In certain diseases such as rheumatoid arthritis, enhanced
Fn deposition is observed on the articular cartilage or synovial
tissue
This could lead to enhanced monocyte attachment to the surface and consequent monocyte activation.
However, enhanced levels of fibronectin fragments (Fn-f) are
also found,"." and these could interfere with either normal
or pathologic Fn function. A further complication is that Fnf often have greatly enhanced or altered properties over those
of native Fn,' and therefore their effect on normal Fn function
cannot be predicted. Therefore, we have studied the effect of
Fn-f on an important event in monocyte activity; i.e., protease
release. We show here that Fn-f can cause release of elastase
from Fn surface-bound monocytes, whereas native Fn cannot
and actually blocks the effect of the Fn-f. Because Fn-f are
found in the synovial fluid,"." the outcome could be enhanced deposition onto the cartilage or synovial tissue of
elastase from tissue-bound monocytes.
From the Department ofBiochemistry, Rush-Presbyterian-StLuke's
Medical Center, Chicago, IL.
Submitted May 4, 1992; accepted September 8, 1992.
Supported by the Rush University Committee on Reseurch, the
Greater Scleroderma Foundation of Chicago, and National Institute
of Arthritis and Musculoskeletaland Skin Diseases Specialized Center
of Osteoarthritic Research Grant No. AR39239-03.
Address reprint requests to Gene A . Homandberg, PhD. Department
of Biochemistry, Rush-Presbyterian-St Luke's Medical Center, 1653
West Congress Parkway, Chicago, I L 60212-3864.
The publication costs of this article were defiayed in part by page
charge payment. This article must therefore be hereby marked
"advertisement" in accordance with 18 U.S.C. section 1734 solely to
indicate this fact.
0 I993 by The American Society of Hematology.
0006-4971/93/8101-0035$3.00/0
186
Ser integrin-recognitionpeptide was half-maximallyactive
at 1 0 0 pmol/L. Fluid-phase Fn was ineffective yet blocked
the activity of the Fn fragments. Complexing of Fn with
gelatin or with heparin partially removed the blocking effect
of Fn. Similar results were obtained with U-937 cells. Substitution of the Fn-gelatinsurface with bovine articular cartilage also promoted elastase release. Therefore, in conditions in vivo in which monocytes bind to tissue surface,
a high ratio of Fn fragments to native Fn may upregulate
certain monocyte activities such as protease release.
0 1993 by The American Society of Hematology.
MATERIALS AND METHODS
All common chemicals were of the highest quality available. Except
where noted, all protease substrates, proteases, chemical reagents,
and chromatography resins were from Sigma Chemical Co (St Louis,
MO). Human sputum elastase and human sputum cathepsin G (41 15
U/mg) were from Elastin Products. The peptide, GRADSPK, were
from ICN (Costa Mesa, CA). Bovine spleen cathepsin D; human
plasma thrombin (1,300 NIH U/mg); human kidney urokinase; human plasma plasmin; human glu-plasminogen; type 1 porcine gelatin;
Escherichia coli lipopolysaccharide;concanavalin A from Cunavalis
ensifomis; deoxyglucose; Streptomyces actinomycin D cytochalasin
B from Helminthosporiumdematioideum; cycloheximide; the elastase
substrate: MANA (N-methoxysuccinimidyl-L-Ala-L-Ala-L-Pro-LVal-gnitroanilide); the cathepsin G substrate: N-succinimidyl-LAlaL-Ala-L-Pro-L-Phe-p-nitroanilide;
the plasmin substrate: S-225 1
(H-D-Val-Leu-Lys-para-nitroanilide);
the peptides: RGDS and
GRGLSLSR the endotoxin test kit: E-toxate with limulus amebocyte
lysates; DMEM (Dulbecco's modified Eagle's medium); RPMl 1640;
Hanks' buffer solution; Hypaque-Ficoll (Sigma Histopaque 1077);
and gentamicin were from Sigma. Electrophoresis supplies were from
Bio-Rad (Richmond, CA).
Isolation of Fn and Fn-1: Fn was isolated from plasma by adsorption to gelatin-Sepharose and Fn-f were isolated from cathepsin
D and thrombin digests of Fn as described.12The Fn-f studied here
have been
and the amino-terminal sequences of some
of the Fn-f p~b1ished.l~
The Fn-f from amino-terminus to carboxylterminus are a 29-Kd heparin-binding Fn-f, a 50-Kd gelatin-binding
domain, a mixture of 1 IO-Kd to 150-Kd (denoted 140-Kd here) integrin-binding central Fn-f, a 35-Kd carboxyl-terminal heparinbinding Fn-f, and a mixture of 160-Kd to 190-Kd Fn-f containing
the gelatin-binding and integrin-binding domains, which we term the
190-Kd Fn-f. Quantitation of Fn-f was performed by use of extinction
~0efficients.l~
Immediately before use, Fn-f solutions were dialyzed
against 50 mmol/L of phosphate without protease inhibitors or against
DMEM and sterile filtered. Periodic testing for endotoxin with the
endotoxin kit from Sigma Chemical Co showed insignificant levels
of less than 50 pg/mg protein or less than 50 pg/mL.
Isolation of human mononuclear cells and of peripheral blood
monocytes und characterization. Human mononuclear cells were
isolated from human buffy coats by Hypaque-Ficoll centrif~gation'~
as modified.' Viability of cells was assessed by trypan blue and was
typically over 90%. Monocytes were visualized by use of the a-naphthy1 butyrate esterase kit from Sigma. Esterase staining suggested that
30% of the mononuclear cells were monocytes. Based on this determination, the preparations were diluted to 5 million monocytes per
milliliter for use. To isolate purer monocytes, mononuclear cells at
10 million/mL were allowed to attach to Fn-gelatin of 35-mm culture
wells and nonadherent cells washed away.' The attached monocytes
were then studied for protease release without removal and reattachBlood, Vol 81,No 1 (January l), 1993:pp 186-192
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
FIBRONECTIN FRAGMENTS CAUSE ELASTASE RELEASE
ment. Such preparations were over 95% monocytes as shown by esterase staining.
Fn-gelatin-coated tissue culture wells. Plastic 24-well tissue culture wells (1.6-cm diameter) were coated with 500 pL of a 30 mg/
mL solution of type I gelatin in PBS-Mg2+(0.1 mol/L phosphate,
0.15 mol/L NaCI, 10 mmol/L Mg2+,pH 7.4) at 40°C. After 2 hours,
the excess reagent was removed by aspiration and the well dried at
37°C for 2 hours, and 500 pL of a solution of 1 mg/mL Fn in PBSMg2+added. After I hour at 37”C, the excess was aspirated away,
washed three times with PBS-Mg2+,and the plate kept in a humidified
atmosphere.
The amount of Fn bound to gelatin surfaces was determined by
addition of 1Z51-labeled
Fn. The amount of Fn bound was dependent
on both gelatin concentration and Fn fluid phase concentrations.
With our optimal conditions of 30 mg/mL gelatin and 1 mg/mL Fn,
the maximal amount of Fn bound. This amount was 9 (+) 1.1 (n =
8) pg/cm2. With 5 mg/mL gelatin and I mg/mL Fn, 4.5 (+) 0.7 (n
= 5) pg of Fn bound/cm’. With 30 mg/mL gelatin and 160 pg/mL
Fn, 4.5 (+) 1.2 (n = 5) pg of Fn bound/cm2. The number ofmonocytes
that bound were dependent on the amount of surface-bound Fn.
With 9 pg of Fn/cm’ and addition of 5 million monocytes in I mL,
the maximal attachment was 140,000 (+) 32,000 (n = 12) monocytes/
cm’, or approximately 700,000 cells/well. This represents the binding
of about 14% of added monocytes.
Culture of U-937 cells. This was performed in suspension culture
in 10%fetal bovin serum-Roswell Park Memorial Institute medium
(FBS-RPMI) 1640 containing 25 pg/mL gentamicin in a Corning
no. 251 11 75-cm’ canted neck polystyrene flask (Corning Glass
Works, Corning, NY) in 5% C02-95% air at 37°C. Every 3 days, or
when the cells reached a density of 1 million cells/mL, the cells were
subcultured at 0.25 million cells/mL. Before the cells were used to
test for elastase release, suspensions were washed three times with
50-fold dilutions of RPMI 1640 to remove serum that would inhibit
released elastase. The washed cells were used within 2 hours.
Bovine articular cartilage slices. These were obtained by removal
of cartilage from the metacarpophalangeal joints of adult bovine animals (1 8 to 20 months old) obtained from the slaughter house. Joints
were opened under aseptic conditions, slices cut from the exposed
articular surface, minced with a surgical scalpel, washed with DMEM,
and frozen to inactive all cellular processes.
Protease assays. Elastase activity was measured spectrophotometrically at 406 nm by hydrolysis of 0.5 mmol/L MANA at 22°C
in a I-mL volume. Aliquots of less than 50 pL of test solution or
standard and 50 pL of a 10 mmol/L solution of MANA in dimethyl
sulfoxide (DMSO) were added to 50 mmol/L tris-HC1, 150 mmol/
L NaCI, pH 7.4. A 0.1 mol/L pg aliquot of human sputum elastase
typically gave a rate of absorbance change at 406 nm of 0.01 U/min/
mL through a I-cm path, with 0.5 ng being the minimal amount
detectable in a 30-minute assay.
Cathepsin G activity was assayed with 3 mmol/L N-succinimidylL-Ala-L-Ala-L-Pro-L-Phe-p-nitroanilide
in mL of 50 mmol/L tris, I
mol/L NaC1, pH 7.4 at 22°C. A 0.02-pg/mL solution of human
sputum cathepsin G had an absorbance change at 410 nm of0.0005/
min/mL through a I-cm light path. The minimal detectable level
was 1 ng.
Plasminogen activator activity was measured at 22”C, as described,
by addition of test solution or 0.1 pg human urokinase to I mL of
50 mmol/L tris, 150 mmol/L NaCI, pH 7.4, containing 25 &mL
Glu-plasminogen and the plasmin-substrate S-225 I . I 6 The absorbance
change at 406 nmol/L was monitored. The exponential rate of increase
of absorbance caused by increasing plasmin activity was calculated
and plotted as absorbance change versus time’. The minimal detectable activity over a 30-minute assay period corresponded to 5 ng
of urokinase.
187
RESULTS
Fn-fdid not cause elastase releasefrom monocytes in suspension. We first investigated the ability of Fn-f to cause
elastase release from mononuclear cells. A suspension of
mononuclear cells in DMEM-containing I O mmol/L of
MgC12 in a sterile polypropylene tube was adjusted to 5 million monocytes/mL and l pmol/L in either native Fn, the
29-Kd, 50-Kd, 140-Kd, 35-Kd, or 190-Kd Fn-f or 1 mmol/
L in the RGDS peptide. Incubation was continued at 37°C
in 95% air-5% C 0 2 .After 2 hours, 100-pL aliquots of these
solutions were assayed. Such solutions were found to contain
less than 0.5, 1, and 5 ng, respectively, of elastase, cathepsin
G, and plasminogen activator activity per 1 million monocytes. These amounts were not considered significant. Four
different mononuclear isolations gave similar results.
Fn-fdid not cause elastase releasefrom monocytes attached
to plastic. The effect of a plastic tissue culture surface on
Fn fragment-mediated release was investigated next. Similar
mononuclear solutions, as above, were added to each well of
a 24-well tissue culture dish. After 1 hour, the density of
bound monocytes was similar to that on Fn/gelatin surfaces;
between 120,000 and 160,000 monocytes/cm2. The fraction
of cells that attached was about 16% of applied. Fn or the
Fn-fwere then added as described above. After an additional
2 hours, the incubations were assayed. None of the experimental wells showed detectable elastase, cathepsin G, or activator release when four different mononuclear preparations
were assayed in this manner.
Fn-f did cause elastase release from monocytes bound to
Fn-gelatin-surfaces. The effect of a Fn-gelatin surface was
investigated next. Aliquots (1 mL) of similar mononuclear
suspensions with 5 million monocytes per milliliter were
added to each well of a 24-well tissue culture dish that had
been first coated with gelatin and then with native Fn. After
1 hour, the number of monocytes bound ranged between
120,000 and 160,000 cells/cm2 or averaged 700,000 cells
bound per well in a I-mL volume. Then, I-pmol/L concentrations of Fn or the Fn-f were added. After 2 hours, 100-pl
aliquots of these solutions were assayed for elastase activity.
In this case, the 29-Kd, 140-Kd, 190-Kd Fn-f, the Arg-GlyAsp-Ser (RGDS) peptide (Fig 1; RGD), and a mixture of the
29-Kd, 50-Kd, 35-Kd, and 140-Kd Fn-f (MIX), respectively,
caused release of 20 (+) 4, 32 (k)5 , 35 (k)6, 15 (+) 5, and
28 (+) 4 ng, respectively, of elastase per million monocytes
(Fig 1). The monocyte count included both bound and unbound monocytes. These data were based on seven different
mononuclear isolations with two different Fn-f preparations
and showed up to 17-fold stimulation of elastase release
compared with bovine serum albumin (BSA) or the control
that was only a gelatin surface. Incubations of cells on a Fngelatin surface with added fluid phase Fn (Fig 1: FN)caused
only a slight release of elastase above control levels. Also, the
RGDS mutant peptides, GRGLSLSR and GRADSPK, at
0.2 mmol/L did not cause elastase release above control levels
(data not shown). Goat antibodies to human Fn, at 50 pg/
mL final concentration, decreased the 0.1 pmol/L 140-Kd
Fn-f-mediated elastase release by 70 (k)9% (n = 3).
The elastase released was not de novo synthesized. To
determine if the elastase released was presynthesized and
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
XIE, MEYERS, AND HOMANDBERG
45
40
w 35
30
-
3 2025 0
15 -
’0
t
5k
0
C
BSA FN 29 35 50 140 190 RGD MIX
FRAGMENT WITH FN:GEL
Fig 1. Releaseof elastase from monocytesadherent to Fn-gelatin
surfaces. Mononuclear cells were prepared by Hypaque-Ficollcentrifugation. The cells were suspended in DMEM-containing 1 0
mmol/L MgC12at 5 million monocytes per milliliter and allowed to
bind to a Fn-gelatin surface of a 24-well tissue culture dish at 37°C
in 5% C02-95% air. After 30 minutes, suspensions were adjusted
to 1 rmol/L Fn or Fn-f or 1 mmol/L RGDS peptide. Incubation was
continued at 37°C in 95% air-5% C02. Aliquots of 1 0 0 pL. which
represented 500,000 monocytes, were assayed after 2 hours for
activity against MANA. Elastase quantities are nanograms per million monocytes, with means and standard deviations based on seven
different mononuclear preparations.
packaged or de novo synthesized, we subjected mononuclear
suspensions to lysis with 4 mol/L urea-50 mmol/L tris- 150
mmol/L NaC1, pH 7.4 for 30 minutes and assayed lysates.
The lysates showed 42 ng elastase/million monocytes as
compared with the maximal amount of release by 1 pmol/
L 140-Kd Fn-f of 35 ng/million monocytes. Further, adjustment of cell suspensions to I O pg/mL actinomycin D for 30
minutes before Fn-f addition did not block release. Therefore,
the elastase released by Fn-f was not de novo synthesized
during the incubation.
Cathepsin G release. Cathepsin G was also released in
140-Kd Fn-f-conditioned media at a level of 5 ng per million
monocytes, about fivefold over control levels. Plasminogen
activator activities could not be detected in any of the solutions.
Fn-factivity. The Fn-f were half-maximally active at 0.1
pmol/L, whereas the RGDS peptide was half-maximally active at 100 pmol/L (Fig 2). The elastolytic activity released
was not inhibited by 20 mmol/L EDTA or by 1 U/mL of
the serine protease inhibitor, trasylol, but was inhibited by a
1 mmol/L concentration of the serine protease inhibitor,
phenylmethylsulfonyl fluoride. A 50 U/mL final concentration of trasylol did inhibit 50% of the activity. This inhibition
spectrum is similar to that found for monocytic and neutrophil elastase.”
The individual eflects of the Fn-f were not additive. The
effect of inactive Fn-f on possibly blocking the active Fn-f
was tested next. Mixtures of 0.1 pmol/L 140-Kd Fn-f, or 0.1
pmol/L 29-Kd Fn-f with 0.1 pmol/L 35-Kd, or with 50-Kd
Fn-f did not diminish the effect of the 29-Kd or 140-Kd Fnf by greater than 5% for three observations. A solution of 0.1
pmol/L 29-Kd Fn-f and 0.1 pmol/L 140-Kd Fn-f did not
increase the elastase release over that caused by the 140-Kd
Fn-f alone.
Native soluble Fn and an anti-integrin antibody blocked
Fn-f-mediated elastase release. The activity of Fn in causing or blocking Fn-f activities was investigated as well as investigation of similar sites of interaction. A similar experiment
was performed as in Fig 1 except in some samples equimolar
( 1 pmol/L) fluid phase Fn was included with monocyte incubations with I pmol/L 140-Kd or I pmol/L 29-Kd Fn-f.
In this case, soluble Fn blocked 66% of the 140-Kd activity,
80% of the 29-Kd activity, and 80% of the RGDS activity.
However, 0.2 pmol/L Fn blocked only 33% of the activity
of a I pmol/L 140-Kd Fn-f solution. Therefore, the Fn-f bind
to similar sites that native Fn can bind, native Fn can block
the Fn-f activities, and native Fn may regulate the Fn-f effect
in vivo.
Next, the ability of an antibody to a Fn receptor to block
release was tested. Addition of 10 pg/mL of anti-integrin,
anti-a,p, (Chemicon Corp, Temecula, CA) to a 0. I-pmol/L
140-Kd or 29-Kd monocyte incubation on a Fn-gelatin surface decreased elastase release by 67 (+) 9% (n = 3) and 32
(+) 7% (n = 3), respectively.
Fn-ligand complexes had a lower blocking activity than
native Fn. Because Fn-ligand complexes are found in vivo
and may not have similar activities as native Fn, we tested
whether two types of complexes are active in blocking elastase
release as is native Fn. We found that addition of complexes
of Fn-heparin, to a final I-pmol/L concentration, to incubations of 1 pmol/L 140-Kd Fn-f with Fn surface-bound
monocytes did not block elastase release (1 10%activity of a
1-pmol/L 140-Kd control) as did fluid phase native Fn (10%
of the 140-Kd activity). A complex of Fn-gelatin, at 1 pmol/
L, blocked 60% of the 140-Kd Fn-f-induced elastase release.
Gelatin and heparin alone, at 2 pmol/L, did not cause elastase
40
T
30
W
In
2
2 20
W
0
C
10
0
0.01
0.1
10
1
100
1000
uM
Fig 2. Effect of Fn-f concentration on elastase released from
monocytes adherent to Fn-Gelatin surfaces. Incubations were performed as described in Fig 1, with various 140-Kd, 29-Kd, and RGDS
concentrations. Elastase quantities are nanograms per million
monocytes, with means and standard deviations based on seven
different mononuclear preparations.
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
FIBRONECTIN FRAGMENTS CAUSE ELASTASE RELEASE
release. Therefore, complexes of Fn may not block the Fn-f
effect in vivo.
Tests of other components. We then investigated whether
the Fn-f effect was specific and whether other components
could also cause elastase release from Fn surface-bound
monocytes. Concanavalin A (1 mg/mL), a monocyte activator, did not cause elastase release, although lipopolysaccharide (100 pg/mL), also an activator, caused release of 8 I %
of the elastase activity of 140-Kd Fn-f treatment. Goat antirabbit IgG (50 pg/mL) did not cause release. Tissue components such as gelatin ( 1 mg/mL), heparin (50 pg/mL), and
chondroitin sulfate (50 pg/mL) did not induce elastase release.
However, prostaglandin E2 ( I pg/mL) caused over 80% of
the release observed with the 140-Kd Fn-f. Therefore, the
elastase release from Fn surface-bound cells is not specific to
Fn-f.
We also tested various components for their ability to block
the 1 pmol/L 140-Kd Fn-f-mediated elastase release from
Fn surface-bound monocytes. The glucocorticoid dexamethasone (0.14 pmol/L) completely blocked release, whereas antibodies to Fn were less effective. However, the metabolic
inhibitors deoxyglucose (3 mmol/L) and azide (0.05%)
blocked elastase release by over 95% as well as did cytochalasin B (10 pg/mL). Therefore, the elastase release was energy
dependent and may require microfilament structure.
The released elastaseactivated unattached monocytes. To
investigate how a system in which only 14% to 16% of the
cells attached to Fn surfaces could release as much elastase
as found in urea lysates of the same number of cells, we
investigated the possible role of elastase as a soluble mediator.
We found that addition of 5 ng of elastase to a 1-mL incubation of 5 million monocytes in a 1.5-mL polypropylene
tube caused release of 38 (+) 19 ng (n = 7) of elastase per
million monocytes within 1 hour of incubation. Because this
is approximately the amount of elastase found in urea lysates
of monocytes and is, therefore, the maximal amount of resident elastase that can be liberated, these results suggest that
elastase released from the small fraction of Fn surface-bound
monocytes also could activate suspended monocytes to release
elastase.
In a second experiment, cells were added to a Fn-gelatin
surface under conditions in which only a fraction of the cells
could bind the surface and after 30 minutes elastase was added
(5 ng/nL final). The addition of elastase caused release of 3 1
(k) 6 ng (n = 3) elastase per million monocytes within 2
hours, confirming that elastase may be a mediator of Fn-fmediated monocytic elastase release in the earlier Fn-gelatin
surface experiments. A similar experiment was then performed to determine if plastic surfaces could replace the Fngelatin surfaces. Cells were first added to wells of a 24-well
plastic tissue culture plate, with no Fn present, followed by
addition of elastase. In this case, after 2 hours, only 2 (k) 1
ng elastase/million cells were detected, whereas controls
showed undetectable levels. The elastase concentration was
not significantly greater than in the control wells after 14
hours. It was expected that full elastase release would have
occurred within 2 hours as with cells in the polypropylene
tubes. We have no explanation for this difference. A similar
experiment was then performed to test whether Fn-f would
189
potentiate the effect of elastase on cells bound to plastic. A
solution of the 140-Kd Fn-f ( I pmol/L final) and elastase (5
ng/mL final) were simultaneously added to the cells previously added to tissue culture wells. By 2 hours, 2 ng elastase/
million cells were detected, but by 14 hours, the levels had
increased to about 36 ng elastase/million cells, about fourfold
over controls at that time point. Therefore, the 140-Kd
Fn-f did potentiate the effect of elastase, but the overall effect
occurred much more slowly than on the Fn-gelatin surface.
To determine whether elastase had to be enzymatically
active to promote release of monocytic elastase, elastase test
solution (0.1 mg/mL) was reacted with 0.1 mmol/L phenylmethyl sulfonyl fluoride (PMSF) and exhaustively dialyzed.
The elastase derivative retained only 1% of its initial activity.
This derivative (5 ng/mL final) was then added to a cell suspension on a Fn-gelatin surface, and the suspension was assayed after 2 hours. There was no increased elastase activity,
even after 14 hours, suggesting that enzymatically active
elastase was required for release of endogenous elastase.
Elastase release also occurred with purified monocytes and
17-937cells. We next investigated if the elastase release from
monocytes could occur in the absence of other mononuclear
cells. Monocytes were purified from mononuclear suspensions by allowing the cells to adhere to Fn-gelatin surfaces.
The unbound cells were washed away and Fn-f added as above
to 1 pmol/L final concentrations. After a 2-hour incubation,
elastase release from these bound purified monocytes was
found to occur at the level of 43 (+) 4.5 ng and 27 (+) 4.3
ng of elastase release/million monocytes (n = 3 different
mononuclear isolations) for the 140-Kd and 29-Kd Fn-f, respectively, whereas the other Fn-f were less than 20% as active.
These cell counts were based on the number of attached cells;
14% of total applied. U-937 cells were also tested. One-milliliter suspensions from a 5 million cells/mL suspension were
allowed to attach to Fn-gelatin surfaces to test for elastase
release. After 1 hour, approximately 32 (k) 5% of the cells
bound. The excess cells were washed away. Addition of
Fn-f resulted in release of 32 (2)4.8 ng/million (n = 5) of
attached U-937 cells (data not shown).
Cartilage slices could substitute for the Fn-gelatin surface
and promote elastase release. Because articular cartilage
tissue has Fn bound to the surface, and because the Fn deposition can increase in rheumatoid a r t h r i t i ~ , we
~ . ~tested
whether normal bovine articular cartilage would promote
elastase release as did Fn-gelatin surfaces. Mononuclear suspensions were adjusted to 1 pmol/L Fn-f or Fn or 1 mmol/
L synthetic RGDS (Arg-Gly-Asp-Ser) peptide or 1 pmol/L
BSA with an approximately IO-mg freeze-killed bovine articular cartilage slice. The cartilage tissue was first frozen to
block any possible active responses from the tissue. Figure 3
shows the results from three different cell preparations and
shows that the same Fn-f that caused elastase release from
Fn-gelatin surfaces did the same in the presence of cartilage
and at approximately the same levels. Native Fn, at 1 pmoll
L, also blocked the Fn-f-mediated elastase release here as
well, and addition of 1 mg/mL gelatin removed 72% of the
blocking activity of soluble Fn. Without cartilage there was
insignificant elastase release after 2 hours, and no elastolytic
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
190
XIE, MEYERS, AND HOMANDBERG
50
45
40
w 35
30
t
25
20
0
15
10
5
0
C
BSA F N 29 35 50 140 1 9 O R G D M I X
FRAGMENT WITH CARTILAGE
Fig 3. Release of elastase from monocytesadherent to a bovine
articular cartilage slices. Experiments were performed by addition
of a 5-mg to 10-mg freeze-killedcartilage slice to a 1-mL suspension
of mononuclear cells at 5 million monocytes per milliliter. After 30
minutes, the suspensions were adjusted to 1 Wmol/L Fn or Fn-f.
After an additional 2-hour incubation at 37°C in 5% C02-95% air,
with occasional gentle agitation, a 100 CL aliquot was assayed for
activity against MANA. Elastase quantities are nanograms per million monocytes, with means and standard deviations based on three
different mononuclear preparations.
activity was detected in Fn-flcartilage incubations without
monocytes after 2 to 6 hours.
DISCUSSION
Our objectives here were to investigate whether or not the
presence of Fn-f, such as are found in arthritic synovial
fluid,'0,'' and which are likely found as products of inflammation at sites oftissue injury, can modulate protease release
from monocytes. We studied the effects of Fn-f both on suspended monocytes and on surface-bound monocytes because
we considered that attachment may activate monocytes. A
Fn surface was considered relevant for these studies because
Fn may bind to sites of injury and present an attachment
surface to monocytes in tissue injury. For example, in rheumatoid arthritis, enhanced levels of Fn are found on the articular cartilage surface, on the synovial membrane, and on
the surface of pannus tissue near attached mononuclear
cell^.^.^ To provide a Fn surface for in vitro studies, we used
a Fn-gelatin surface because such a surface has been shown
to promote monocyte binding in a Fn concentration-dependent fashion.' With such a surface the gelatin does not likely
serve a specific function other than to concentrate the Fn to
enhance the affinity for the attachment molecules/receptors.
The rationale for investigating the effect of Fn-f was because
they are found in rheumatoid arthritic synovial fluids,"." in
plasma after thermal injury,I8 and will likely be found in
body fluids after or during any type of tissue injury in which
protease activities are expressed. Because Fn-f may block
normal Fn function,"." and they often have altered properties
over native Fn, their activities may be of relevance in many
types of pathologies. Our means of analysis of their effect
was to measure certain proteases that have been shown to be
released by monocytes, such as elastase,' cathepsin G,I9 and
plasminogen activator.20
The Fn-f did cause elastase release but only when monocytes were first bound to a Fn-gelatin or cartilage surface.
The effect of cell attachment on protease release has been
reported before; for example, suspended neutrophils stimulated with soluble ligands have been reported to release very
little collagenase as compared with neutrophils bound to IgGcoated surfaces or cartilage surfaces.2' Further, Wright et a1'
found that soluble Fn does not activate complement receptors
of monocytes but surface-bound Fn does. Also, phorbol esters
have been shown to stimulate elastase and cathepsin G release
from monocytes but only when Fn is surface bound.22
Many of the activities of macrophages, including chemotaxis, adherence, differentiation, secretion, and phagocytosis,
can be attributed to the binding of Fn by integrin receptors
on the macrophage cell surface. Four integrin receptors on
mononuclear phagocytes have been documented, including
VLA-5 (asbl)and IIa/IIIb (a2b3), which can bind the RGDS
and VLA-3 (a3b,)and VLA-4 (Gbl), which binds
independently of the RGDS s e q ~ e n c e . ~The
~ . ~effect
'
of the
140-Kd integrin-binding Fn-f observed here may occur
through the VLA-5 or IIa/lIIb receptors. However, the activity
of the 29-Kd Fn-f must occur through the other receptors or
to other undiscovered receptors. A likely candidate may be
a recently described 67-Kd macrophage surface protein with
affinity for the amino-terminal Fn domain.26 This protein
had an affinity of 20 nmol/L for the 29-Kd Fn-f, and its
interaction with the 29-Kd Fn-f was reported to be blocked
by native Fn but not by the 140-Kd integrin-binding Fn-f.26
One of the activities of this protein may be to mediate interaction of fibrin with cells of the mononuclear phagocyte lineage.z6There may be such cell surface-binding proteins on
many types of cell because the 29-Kd Fn-f can bind fibroblast~,~'
endothelial cell^,'^-'^ and chondrocytes.28Or the 29Kd Fn-f may bind heparin-like surface components in a process termed matrix-driven transl~cation.~~
Such an event may
trigger secondary responses such as elastase release. This 29Kd Fn-f has several cellular activities and has been shown to
be chemotactic for monocytes,30 to inhibit endothelial cell
g r o ~ t h , ' ~and
- ' ~to stimulate protease expression in chondrocytes.28Further, the 29-Kd Fn-f is found in vivo in normal
human plasma at about 5 pg/mL3' and is elevated in plasma
of diabetic patients, indicating endothelial d y s f ~ n c t i o n . ~ ~
Much further work will be required to determine what
receptors are involved in the activities of the 29-Kd and 140Kd Fn-f and how these Fn-f cause effects different than does
native Fn and why monocytes must be attached to Fn surfaces, likely through Fn receptors, for the Fn-f to be effective.
We could propose that normally binding of monocytes to Fn
surfaces or to soluble Fn through Fn receptors does not allow
elastase release. However, when the Fn-f are added and interaction of the 140-Kd Fn-f with RGDS receptors and the
29-Kd Fn-f occurs, with perhaps the recently reported 67Kd protein discussed above, this may stimulate the cells to
undergo different changes in the cytoskeletal structure than
those caused by native Fn. These changes may enhance secretion of elastase containing granules by removing the inhibitory effect of Fn and Fn receptors at the basal surface.
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
FIBRONECTIN FRAGMENTS CAUSE ELASTASE RELEASE
The addition of native soluble Fn may block these Fn-f effects
by displacing the bound 140-Kd and 29-Kd Fn-f and allowing
a readjustment of the cytoskeletal structure.
Such receptor-mediated events in mononuclear cells may
require the participation of or regulation by multiple receptors. For example, a 120-Kd cell-binding Fn-f or native Fn
was shown to potentiate collagen-induced secretion of interleukin-1 (IL-1) from human blood mononuclear cells by
binding to the a5bl Fn i ~ ~ t e g r iThus,
n . ~ ~occupancy at the a,b,
integrin by Fn-f or Fn apparently regulates the effects of collagen interaction at the collagen receptor, the a2bl integrin.
In our system reported here, native Fn interaction appeared
to potentiate the effect of Fn-f. However, we have no evidence
for involvement of a collagen receptor.
The differential effect of Fn-f as compared with native Fn
could be because the Fn-f may simply bind to Fn receptors
differently than native Fn because of their lesser multivalent
interactions at the cell surface. They may be incapable of
causing receptor clustering or the normal activities of native
Fn. Or they may introduce new postbinding receptor responses different than does Fn. Therefore, Fn-f may act as
partial or incomplete agonists of Fn and may either activate
certain processes or remove the inhibitory effects of native
Fn. For example, it has been shown that when monocytes
bind to Fn surfaces, complement receptor activity is enhanced, but the enhancement is removed when RGDS peptides are subsequently added.23In this case, the RGDS peptides may block normal Fn function, whereas in our work
reported here, the Fn-f may either activate or simply remove
the normal blocking effect of native Fn. We found that saturation with soluble native Fn actually blocks elastase release,
consistent with our suggestion that the normal effect of interaction of Fn with Fn receptors may be to disallow elastase
release. The addition of Fn-f may compete with and reverse
the inhibitory effect of the Fn at the basal surface.
The differential effect of native Fn and Fn-f on protease
release also has been demonstrated in a neutrophil system in
which Fn-f promote degranulation but native Fn does not34
and in a synovial fibroblast system in which only Fn-f promote
stromelysin release.35 We also have recently reported that
Fn-f can augment metalloproteinase release from chondrocytes in cartilage, whereas native Fn does not.*’
The addition of fluid phase Fn blocked the effect of the
Fn-f, suggesting that elastase release is not dependent only
on the concentration of Fn-f but rather on the ratio of Fn-f
to native Fn. However, addition of heparin and gelatin ligands
to Fn partially removed the ability of Fn to block elastase
release, likely because these ligands blocked the integrinbinding region of Fn. These observations suggest that high
ratios of Fn-f and Fn-ligand complexes to Fn may favor elastase release and that Fn ligand complexes formed in vivo
may not be effective in activating monocytes.
It should be noted that in our system, only about 14%of
the added monocytes actually bound to the Fn surface. Because the maximal amount of elastase caused to be released
by Fn-f is similar to the amount found in 4 mol/L urea lysates
of the cells, the elastase must be released from suspended
cells as well. Therefore, a diffusible factor(s) such as elastase
may be released from bound monocytes and, in turn, augment
191
elastase release from suspended cells. We did find that addition of elastase to suspended monocytes or to monocytes
on a Fn-gelatin surface augmented elastase release. Enzymatically inactive elastase did not augment release. This general observation on the effect of elastase is similar to the observation that kallikrein stimulates release of elastase from
neutrophils during c o a g ~ l a t i o nAlso
. ~ ~ of relevance is the observation that addition of elastase or cathepsin G to monocytes has been shown to enhance release of oxygen metabolites
and to enhance the oxidative metabolic re~ponse.~’
The release of elastase was blocked by deoxyglucose, sodium azide, and cytochalasin B, suggesting that metabolic
energy and microfilament activity were required for the release. Actinomycin did not block, suggesting that mRNA
synthesis was not required. Further, 4 mol/L urea lysates
showed the same level of elastase as measured on addition
of Fn-f to Fn surface-bound monocytes. Therefore, the elastase was probably packaged and secreted in an energy-dependent process when Fn-f were added. Although it is generally assumed that proteases are not packaged in cytoplasmic
granules in monocytes as they are in neutrophils, it has been
shown by immunoelectron microscopy that cytoplasmic
granules that show peroxidase activity also contain elastase
and cathepsin G activity.38The existence of elastase in such
granules is consistent with our data presented here.
We substituted the Fn-gelatin surface with bovine articular
cartilage slices that contain endogenous Fn on the surface to
mimic certain events that might occur in rheumatoid arthritis.
The Fn-f promoted elastase release in this system as well.
Therefore, higher levels of Fn on damaged cartilage or synovial tissue may promote monocyte binding, and Fn-f in
synovial fluid may promote release of elastase directly onto
the tissue surface.
An interesting feature of promotion of elastase release by
the Fn-f is that the proteolysis may be upregulated even further because Fn-f that indirectly cause proteolysis may originate from the induced proteolysis. This may occur in vivo
because Fn-f have been shown to arise by the action of activated leukocytes on endothelial cells.39 We conclude that
in conditions in which monocytes attach to Fn surfaces and
in which Fn fragmentation or formation of Fn ligands is relatively high, elastase release would be greatly enhanced.
REFERENCES
1. Bevilacqua MP, Amrani DL, Mosesson MW, Bianco D Re-
ceptors for cold-insoluble globulin (plasma fibronectin) on human
monocytes. J Exp Med 153:42, 1981
2. Ruoslahti E: Fibronectin and its receptors. Annu Rev Biochem
51:315, 1988
3. Pommier CB, Inada S, Fries LF, Takahashi T, Frank MM, Brown
El: Plasma fibronectin enhances phagocytosis of opsonized particles by
human peripheral blood monocytes. J Exp Med 157:1844, 1983
4. Wright SD, Craigmyle LS, Silverstein SC: Fibronectin and
serum amyloid P component stimulate C3b- and C3bi-mediated
phagocytosis in cultured human monocytes. J Exp Med 158:1338,
1983
5 . Wright SD, Licht RM, Craigmyle LS, Silverstein SC: Communication between receptors for different ligands on a single cell:
Ligation of fibronectin receptors induces a reversible alteration in
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
192
the function of complement receptors on cultured human monocytes.
J Cell Biol 99:336, 1984
6. Wright SD, Meyer BC: Fibronectin receptor of human macrophages recognizes the sequence arg-gly-asp-ser. J Exp Med 162:
762, 1985
7. Bianco C, Edelson PJ: Disease, in Siskind G, Weksler M (eds):
New York, NY, Grune & Stratton, 1977, p I
8. Lavietes BB, Carsons S, Diamond HS, Laskin RS: Synthesis,
secretion and deposition of fibronectin in cultured human synovium.
Arth Rheum 28:1016, 1985
9. Scott DL, Wainwright AC, Walton KW, Williamson N: Significance of fibronectin in rheumatoid arthritis and osteoarthritis.
Ann Rheum Dis 40: 142, I98 I
IO. Carsons S, Lavietes BB, Diamond HS, Kinney SG: The immunoreactivity, ligand and cell binding characteristics of rheumatoid
synovial fluid fibronectin. Arth Rheum 28:601, 1985
1 1. Carsons S, Lavietes BB, Diamond HS: Role of fibronectin in
rheumatic diseases in fibronectin, in Mosher D (ed): Biology of Extracellular Matrix: A Series. San Diego, CA, Academic, 1989, p 327
12. Homandberg GA, Williams JE, Schumacher B, Grant D,
Eisenstein R: Heparin-binding fragments of fibronectin are potent
inhibitors of endothelial cell growth. Am J Pathol 120327, 1985
13. Homandberg GA, Kramer J, Grant D, Christianson G,
Eisenstein R Heparin-binding fragments of fibronectin are potent
inhibitors of endothelial cell growth: Structure-function correlates.
Biochim Biophys Acta 874:61, 1986
14. Homandberg GA, Erickson J: A model of fibronectin tertiary
structure based on studies of interactions between fragments. Biochemistry 25:69 17, 1986
15. Boyum A: Isolation of mononuclear cells and granulocytes
from human blood: Isolation of mononuclear cells by one centrifugation, and of granulocytes by combining centrifugation and sedimentation at I g. Scand J Clin Lab Invest 21:77, 1968 (suppl 97)
16. Homandberg GA, Wai T Reduction of disulfidesin urokinase
and insertion of a synthetic peptide. Biochim Biophys Acta 1038:
209, 1990
17. Baugh RJ, Travis J: Human leukocyte granule elastase: Rapid
isolation and characterization. Biochemistry 15336, 1976
18. LaCelle P, Blumenstock FA, Saba TM: Blood-borne fragments
of fibronectin after thermal injury. Blood 77:2037, 1991
19. Werb Z, Gordon S: Elastase secretion by stimulated macrophages. Characterization and regulation. J Exp Med 142:361, 1975
20. Gordon S, Unkeless JC, Cohn ZA: Induction of macrophage
plasminogen activator by endotoxin stimulation and phagocytosis:
Evidence for a two-stage process. J Exp Med 140:995, 1974
21. Chatham WW, Heck LW, Blackburn WD Jr: Ligand-dependent release of active neutrophil collagenase. Arth Rheum 33:228,
I990
22. Pantalone RM, Page RC: Lymphokine induced production
and release of lysosomal enzymes by macrophages. Proc Natl Acad
Sci USA 72:2091, 1975
23. Brown EJ, Goodwin JL: Fibronectin receptors of phagocytes.
Characterization of the arg-gly-asp binding proteins of human
monocytes and polymorphonuclear leukocytes. J Exp Med 167:777,
1988
XIE, MEYERS, AND HOMANDBERG
24. Takada Y, Wayner EA, Carter WG, Helmer ME: Extracellular
matrix receptors, ECMRII and ECMRI, for collagen and fibronectin
correspond to VLA-2 and VLA-3 in the VLA family of heterodimers.
J Cell Biochem 37:385, 1988
25. Guan JI, Hynes RO: Lymphoid cells recognize an altematively
spliced segment of fibronectin via the integrin receptor alpha 4 beta
I . Cell 60:53, 1990
26. Blystone SC, Kaplan JE: Isolation of an amino-terminal fibronectin-binding protein on human U937 cells and rat peritoneal
macrophages. J Biol Chem 267:3968, 1992
27. McDonald JA: Extracellular matrix assembly. Annu Rev Cell
Biol4: 183, 1988
28. Homandberg GA, Meyers R, Xie DL: Fibronectin fragments
cause chondrolysis of bovine articular cartilage slices in culture. J
Biol Chem 267:3597, 1992
29. Newman SA, Frenz DA, Hasegawa E, Akiyama SK: Matrixdriven translocation; dependence of interaction of the amino-terminal
domain of fibronectin with heparin-like surface components of cells
or particles. Proc Natl Acad Sci USA 84:479, 1987
30. Lohr KM, Kurth CA, Xie DL, Homandberg GA, Kurth CA:
The amino-terminal 29 and 72 kilodalton fragments of fibronectin
mediate selective monocyte recruitment. Blood 76:2 1 17, 1990
3 1. Skrha J, Richter H, Hormann HT: Evidence for the presence
of a free N-terminal fibronectin 30-kDa domain in human plasma
by quantitative determination with an indirect immunosorbent. Anal
Biochem I73:228, 1988
32. Skrha J, Vackova I, Kvasnicka J, Stibor V, Stolba P, Richter
H, Hormann H: Plasma free N-terminal fibronectin 30-kDa domain
as a marker of endothelial dysfunction in type I diabetes mellitus.
Eur J Clin Invest 20:171, 1990
33. Pacifici R, Basilico C, Roman J, Zutter MM, Santoro SA,
McCracken R Collagen-induced release of interleukin 1 from-human
blood mononuclear cells. Potentiation by fibronectin binding to the
alpha,beta, integrin. J Clin Invest 89:61, 1992
34. Wachtofogel YT, Abrams W, Kucich U, Weinbaum G, Shapria
M, Colman RW: Fibronectin degradation products containing the
cytoadhesive tetrapeptide stimulate human neutrophil degranulation.
J Clin Invest 81:1310, 1988
35. Werb Z, Tremble PM, Behrendtsen 0, Crowley E, Damsky
CH: Signal transduction through the fibronectin receptor induces
collagenase and stromelysin gene expression. J Cell Biol 109:877,
1989
36. Wachtofogel YT, Kucich U, James HL, Scott CF, Schapira
M, Zimmerman M, Cohen AB, Colman RW: Human plasma kallikrein releases neutrophil elastase during blood coagulation. J Clin
Invest 72:1672, 1983
37. Speer CP, Pabst MJ, Hedegaard HB, Rest RF, Johnston RB
Jr: Enhanced release of oxygen metabolites by monocyte-derived
macrophages exposed to proteolytic enzymes: Activity of neutrophil
elastase and cathepsin G. J Immunol I33:2 15I , 1984
38. Kargi HA, Campbell EJ, Kuhn C 111: Elastase and cathepsin
G of human monocytes: Heterogeneity and subcellular localization
to peroxidase-positivegranules. J Histochem Cytochem 38: 1 179, I990
39. Vincent PA, Del Becchio PJ, Saba TM: Release of Fibronectin
fragments from endothelial cell monolayers exposed to activated leukocytes; relationship to plasma fibronectin levels after particle infusion. Exp Mol Pathol 48:403, 1988
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
1993 81: 186-192
Release of elastase from monocytes adherent to a fibronectin-gelatin
surface
DL Xie, R Meyers and GA Homandberg
Updated information and services can be found at:
http://www.bloodjournal.org/content/81/1/186.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.