Injury-induced gelatinase and thrombin-like activities in
regenerating and nonregenerating nervous systems
IGOR FRIEDMANN, ANAT FABER-ELMAN, ETI YOLES, AND MICHAL SCHWARTZ1
Department of Neurobiology, The Weizmann Institute of Science, 76100 Rehovot, Israel
ABSTRACT
It is now widely accepted that injured
nerves, like any other injured tissue, need assistance
from their extracellular milieu in order to heal. We
compared the postinjury activities of thrombin and
gelatinases, two types of proteolytic activities known
to be critically involved in tissue healing, in nonregenerative (rat optic nerve) and regenerative (fish
optic nerve and rat sciatic nerve) neural tissue.
Unlike gelatinases, whose induction pattern was comparable in all three nerves, thrombin-like activity
differed clearly between regenerating and nonregenerating nervous systems. Postinjury levels of this
latter activity seem to dictate whether it will display
beneficial or detrimental effects on the capacity of
the tissue for repair. The results of this study further
highlight the fact that tissue repair and nerve regeneration are closely linked and that substances that
are not unique to the nervous system, but participate
in wound healing in general, are also crucial for
regeneration or its failure in the nervous system.—
Friedmann, I., Faber-Elman, A., Yoles, E., Schwartz,
M. Injury-induced gelatinase and thrombin-like activities in regenerating and nonregenerating nervous
systems. FASEB J. 13, 533–543 (1999)
Key Words: thrombin z matrix metalloproteases z CNS z PNS
Neurons in the mammalian central nervous system
(CNS)2 are unable to regenerate after axonal injury,
whereas both the mammalian peripheral nervous
system (PNS) and the CNS of lower vertebrates are
capable of regeneration (1). PNS axons fail to elongate more than 1 mm into optic nerve grafts (2). In
contrast, neurons from most regions of the mammalian CNS can elongate their axons into PNS grafts,
presumably aided by the supportive cellular environment of the peripheral nerve (3). It therefore seems
that CNS neurons possess, at least in part, the
potential to regenerate, but lack the supportive and
permissive environment that characterizes regenerative systems after injury. Our research group recently
proposed that the processes of tissue repair and
regeneration in the nervous system have features in
common with those of any recovering tissue (4, 5).
Injury of any tissue triggers a complex cascade of
events that modifies the environment and induce a
0892-6638/99/0013-0533/$02.25 © FASEB
process of tissue repair. In effecting these changes,
molecular components such as proteases play a
critical role (6 – 8). Failure of the CNS to regenerate
after injury might be the result of impairment of the
tissue repair mechanism at one or more stages of the
healing cascade, including those involving protease
participation.
One of the proteases known to influence the
nervous system is thrombin, a key enzyme in blood
coagulation (9). Thrombin acts as a potent mitogen
that modulates the morphology of astrocytes (10, 11)
and can induce the secretion of nerve growth factor
(NGF) by astrocytes (12). It inactivates acidic fibroblast growth factor, known to promote neuritic outgrowth and astrocyte proliferation (13–15). It inhibits outgrowth from neuritic neuronal cells due to
interaction with the thrombin receptor PAR-1 (16,
17), which is proteolytically activated by thrombin
(7). Thrombin at high concentrations was shown to
induce apoptotic cell death in astrocytes and neurons cultured under normal conditions, whereas at
moderate concentrations it protected those cells
from a variety of metabolic insults (18 –20). Both
effects were shown to be due to the activation of the
PAR-1 receptor. The processes of neuroprotection
and apoptotic cell death display similar characteristics in the signal transduction pathways (19, 21, 22),
and it was suggested that different concentrations of
thrombin might result in different activation levels of
the same pathway (21).
Another group of proteases involved in wound
healing is the family of matrix metalloproteases
(MMPs, collagenases). Two of these, gelatinase A
(MMP-2, 72 kDa) and gelatinase B (MMP-9, 92 kDa),
belong to the subfamily of gelatinases. MMPs are
largely responsible for the degradation of extracellular
matrix (ECM) components such as collagen and pro1
Correspondence: E-mail: bnschwar@weizmann.weizmann.ac.il
Abbreviations: CNS, central nervous system; DMEM, Dulbecco’s modified Eagle’s medium; ECM, extracellular matrix;
IL, interleukin; MMP, matrix metalloprotease; MMP-2, gelatinase A; MMP-9, gelatinase B; NGF, nerve growth factor; OD,
optical density; PA, plasminogen activator; PN-I, glial-derived
protease nexin I; PNS, peripheral nervous system; SDS-PAGE,
sodium dodecyl sulfate-polyacrylamide gel electrophoresis;
TIMPs, tissue inhibitors of matrix metalloprotease; TNF,
tumor necrosis factor.
2
533
teoglycans in several normal and pathological processes, including tissue remodeling (23), and MMP
genes are among the most abundant of those expressed by cells in inflammatory lesions (24). Remodeling of the ECM by proteolytic activity is known to be
crucial for growth-cone motility (8). Whereas thrombin
is known to produce the matrix component fibrin (9,
25, 26), MMPs are responsible for matrix degradation.
Fibrin fibronectin-containing matrix promotes the
growth of cultured septal and hippocampal neurons
from newborn rats (27). The MMPs are activated by the
fibrin-degrading plasmin, which is itself activated by
plasminogen activators (PA) (24). Cultured neurons
from the PNS release at least two proteases capable of
degrading ECM components, a calcium-dependent
MMP and a PA (28). In addition, NGF-induced expression of gelatinase A by dorsal root ganglionic neurons
occurs in correlation with the ability to degrade ECM
and to extend neurites (29), and NGF production in
astrocytes is induced not only by thrombin but also by
gelatinases (12). In contrast to the above findings,
which point to beneficial effects of MMPs in the
nervous system, several studies hint at the involvement
of MMPs in processes detrimental to the brain. Intracerebral injection of bacterial collagenase as well as of
tumor necrosis factor (TNF) -a induces gelatinase B
production, which causes delayed opening of the
blood–brain barrier (30, 31). Patients with inflammatory neurological disorders, including multiple sclerosis, show increased gelatinase B in the cerebrospinal
fluid (32). Gelatinase B is also present in the cerebrospinal fluid of mice with experimental allergic encephalomyelitis, the animal model of multiple sclerosis,
where it cleaves myelin basic protein (33).
To investigate the relationship of gelatinase and
thrombin activities to regenerative and degenerative
processes in the nervous system in vivo, we compared
their activities in response to axonal injury in two
regenerative nervous systems (rat sciatic nerve and
fish optic nerve) with a nonregenerative nervous
system (rat optic nerve). We show that the induction
of gelatinase activity shortly after injury is comparable in the three nerve types. In contrast, the postinjury induction of thrombin-like activity in the nonregenerative optic nerve differs clearly from that in
both regenerative nerves in a way that could explain
its detrimental and beneficial effects, respectively, on
regeneration.
Animals were used according to the regulations formulated
by IACUC (Institutional Animal Care and Use Committee).
Crush injury
Rats were anesthetized with 10 mg/kg xylazine (Vitamed,
Israel) and 50 mg/kg ketamine (Fort Dodge Laboratories,
Fort Dodge, Iowa). For the optic nerve crush, a lateral
canthotomy was performed in the right eye under a binocular
operating microscope as described previously (34). The conjunctiva was incised laterally to the cornea, the retractor bulbi
muscle was separated, and the optic nerve was exposed. The
dura was left intact. For the sciatic nerve crush, a small
incision was made in the thigh to expose the nerve. The
exposed optic or sciatic nerve was crushed by calibrated
forceps for 30 s. After the sciatic nerve crush, the skin was
sutured. The rats were allowed to recover.
Fish were deeply anesthetized with 0.05% 3-aminobenzoic
acid ethyl ester (Sigma, St. Louis, Mo.). As in the rat, the
conjunctiva was incised laterally to the cornea and the exposed optic nerve was crushed intraorbitally with forceps for
30 s. The fish were then returned to their tanks.
Preparation of conditioned media
At specified times after crush injury, the animals were again
anesthetized and their crushed nerves were dissected out,
cleaned with Kimwipes, and washed in Dulbecco’s modified
Eagle’s medium (DMEM) without phenol red. Each of the
three types of nerves was pooled and incubated separately for
1.5 h at room temperature in DMEM without phenol red. The
nerves were then removed and the three resulting media, now
termed ‘conditioned media’, were collected, centrifuged at
15000 3 g for 15 min in an Eppendorf centrifuge at 4°C, and
stored at 270°C. Before use, the conditioned media were
kept on ice to limit proteolysis. Protein concentrations were
determined by Bradford analysis (Bio-Rad, Hercules, Calif.).
For some experiments, nerves were separated into proximal
and distal segments before incubation. As controls, conditioned media were similarly prepared from uninjured nerves.
Preparation of nerve extracts
As in preparation of the conditioned media, the animals were
reanesthetized at specified times after crush injury. The
crushed nerves were dissected from their sheaths and some
were separated, as before, into proximal and distal segments.
The pooled nerves or nerve segments were then homogenized and left overnight at 4°C in extraction buffer containing 0.5% Triton X-100, 0.1 M Tris-HCl, pH 8.0. After centrifugation at 15,000 3 g for 15 min at 4°C, the protein
concentrations of the supernatants were determined by Bradford analysis (Bio-Rad). Nerve extracts were stored at 270°C
and prior to the activity assays were kept on ice to limit
proteolysis. As controls, nerve extracts were similarly prepared from uninjured nerves.
Thrombin activity assay
MATERIALS AND METHODS
Animals
Carp (Cyprinus carpio) were purchased from Tnuva, Israel.
Sprague-Dawley rats, 8 wk old, were purchased from the
Weizmann Institute of Science Experimental Animal Center.
534
Vol. 13
March 1999
An assay for thrombin activity was designed according to a
previously described procedure (35). Samples of the conditioned media, each containing 5 mg of total protein, were
incubated at 37°C in a flat-bottomed, 96-well plate with 50
mM Tris-HCl, 5 mM EDTA, pH 8.4, and 150 mM of N-p-TosylGly-Pro-Arg-p-nitroanilidine (Chromozym-TH; Sigma), a specific substrate for thrombin, in DMEM without phenol red in
a final volume of 200 ml. In some experiments, the thrombin
The FASEB Journal
FRIEDMANN ET AL.
inhibitor hirudin (Sigma) or thromstop (American Diagnostica, Greenwich, Conn.) was added. Optical density (OD) at
405 nm was measured at the times specified. Wells containing
all of the above except for the conditioned media were used
as blanks, and signals in the tested samples were expressed as
the difference between the measured OD and the OD recorded at the beginning of the assay. Only those measurements obtained when enzymatic activity was in the linear
range are shown. In some cases, pure thrombin from bovine
plasma (Sigma) was assayed in order to quantify the signals of
the assay.
Thrombin activation by ecarin
Samples of the conditioned media, each containing 5 mg of
total protein, were incubated for 2 h at 37°C in a flatbottomed, 96-well plate with 50 mM Tris-HCl, 5 mM EDTA,
pH 8.4, in DMEM without phenol red, in a final volume of
100 ml with 0.1 U or 1 U of ecarin (American Diagnostica), a
specific prothrombin activator derived from snake venom.
Controls were treated in the same way, but without ecarin.
Thrombin-like activity was then assayed as described above.
Most of the existing prothrombin was activated by 0.1 U of
ecarin, as increasing the concentration to 1 U was found to
have no further effect on the signal.
Gelatin zymography
Conditioned media or nerve extracts containing equal
amounts of protein were electrophoresed through an 8%
sodium dodecyl sulfate-polyacrylamide gel electrophoresis
(SDS-PAGE) gel (Bio-Rad) that was polymerized in the presence of 1 mg/ml gelatin. To produce a linear polyacrylamide
gradient gel, a 7% and a 15% gel solution— each containing
1 mg/ml gelatin—were mixed by a gradient mixer and
poured with a peristaltic pump at a speed of 2 ml/min, as
described (36). The gels were washed once for 30 min in
2.5% Triton X-100 to remove the SDS, and once for 30 min in
the reaction buffer (50 mM Tris-HCl, 200 mM NaCl, 10 mM
CaCl2, pH 7.5). The gels were then incubated in fresh
reaction buffer at 37°C for 3 days. Gelatinolytic activity was
visualized by staining of the gels with 0.5% Coomassie brilliant blue.
Statistics
The significance of differences was calculated using the
Welch alternate t test. This test assumes Gaussian populations,
but does not assume that the compared populations have
equal standard deviations, and is therefore more conservative
than the conventional unpaired Student’s t test.
RESULTS
Thrombin-like activity in rat optic and sciatic
nerves
Thrombin activities were assayed in conditioned
media prepared from uninjured rat optic and sciatic
nerves and from nerves excised 1, 4, and 7 days after
the injury (Fig. 1A). Samples of conditioned media
were then incubated for 1 h with the chromogenic
substrate Chromozym-TH and their OD values were
recorded. Conditioned media derived from optic
Figure 1. Thrombin-like activities in rat optic and sciatic
nerves. Nerves were subjected to crush injury, the rats were
killed at the indicated time points after the injury, and
conditioned media were collected. A) Thrombin-like activities in rat optic nerve-conditioned medium (ONCM, white)
and sciatic nerve-conditioned medium (SNCM, black). Activity-generated elevations in OD at 405 nm after 1 h of
thrombin activity assay are shown. The bar graph shows the
mean 6sem of OD values obtained from four different
preparations. The increase in thrombin-like activity on day 1
after injury is extremely significant in the optic nerve
(***P,0.0001) and very significant in the sciatic nerve
(P,0.001) when compared to that obtained in uninjured
nerves. Insert: Thrombin-like activity in rat sciatic nerveconditioned medium. Activity-generated elevations in OD at
405 nm after 10 h of thrombin activity assay are shown. Each
bar graph shows the mean 6sem of OD values obtained from
four different preparations. The increase in thrombin-like
activity on day 1 after injury is very significant compared to
that obtained in uninjured nerves (**P,0.001). B) Thrombin-like activitiy in rat optic nerve-conditioned medium
(ONCM). Activity-generated elevations in OD at 405 nm after
1 h of thrombin activity assay are shown. The graph shows the
mean 6sem of values obtained from different preparations,
whose numbers (n) are indicated at each point. The increase
in thrombin-like activity on day 1 after the injury is extremely
significant (***P,0.0001) compared to that obtained in
uninjured nerves.
GELATINASE AND THROMBIN-LIKE ACTIVITIES IN NERVOUS SYSTEMS
535
of the synthetic thrombin inhibitor thromstop (Fig.
2) or hirudin (not shown). At the different time
points examined, similar dose-dependent patterns of
inhibition were observed in the optic and the sciatic
nerves: the presence of thromstop at a concentration
of 0.1 mM inhibited activity by approximately 60%,
and at a concentration of 1 mM by about 90%.
Thrombin-like activity is potentially associated with
retrograde degeneration
Figure 2. Thrombin-like activity in the presence of the specific
thrombin inhibitor thromstop in rat optic (gray circles) and
sciatic (black squares) nerve. Rats were killed on the day after
injury and conditioned media were collected. Thrombin-like
activity was assayed in the presence of different concentrations of thromstop. Mean values 6sd of representative examples are shown. Calculation of the inhibition was based on OD
elevation at 405 nm after 1 h of thrombin activity assay in the
optic nerve and 10 h of the assay in the sciatic nerve.
nerves excised 1 day after the injury showed a
marked increase in thrombin-like activity relative to
the uninjured optic nerves (P,0.0001) and corresponded in intensity to the activity evoked by approximately 9 3 1023 U of pure bovine thrombin. The
kinetics are shown in more detail in Fig. 1B. An
increase in thrombin-like activity levels could be
detected as early as 1.5 h after injury. On day 1
postinjury, thrombin levels were about 40-fold
higher than those in control nerves and then gradually declined, reaching baseline levels by day 4.
The sciatic nerve also showed an increase in
thrombin-like activity 1 day after the injury, but this
elevation was approximately 1/20th that of the optic
nerve (Fig. 1A) and could be simulated by 0.5 3 1023
U of pure bovine thrombin. Because these signals
were very low, we extended the incubation time with
Chromozym-TH from 1 to 10 h (Fig. 1A, insert). In
both cases, the thrombin-like activity in the sciatic
nerve on day 1 postinjury was approximately threefold higher than in the uninjured sciatic nerve
(P,0.001) and had returned to baseline by 4 days
after injury. Thus, both the sciatic and the optic
nerve showed an increase in thrombin-like activity 1
day after injury, but the increase in the optic nerve
was approximately 20-fold higher than in the sciatic
nerve.
Specific inhibition of thrombin-like activity in the
rat
To further verify the specificity of the measured
thrombin-like activity, we analyzed it in the presence
536
Vol. 13
March 1999
To find out whether the elevated thrombin-like
activity in the injured rat optic nerve on day 1
postinjury is associated with retrograde or anterograde degeneration, some of the nerves were divided
into a distal segment (between the injury site and the
optic chiasma) and a proximal segment (including
the site of injury and the part of the nerve between
the injury site and the optic disk). Conditioned
media from the pooled proximal segments and the
pooled distal segments were collected separately and
their thrombin activities were assayed. In all cases
examined, thrombin activity was found mainly in the
proximal segments (Fig. 3).
Increase of thrombin-like activity by prothrombin
activation
One possible source for the observed activity could
be the activation of prothrombin, known to be
Figure 3. Thrombin-like activity in different parts of the rat
optic nerve on day 1 after the injury. Nerves were subjected to
crush injury, the rats were killed 1 day later, and each optic
nerve was separated into a proximal segment (which included
the site of injury) and a distal segment. OD values at 405 nm
after 0.5 h of thrombin activity assay in conditioned media are
shown. The bar graph shows the mean 6sem of OD values
obtained from three different preparations.
The FASEB Journal
FRIEDMANN ET AL.
expressed in the nervous system (37). To compare
the amount of prothrombin (a reservoir for thrombin-like activity) in the rat optic or sciatic nerves
before and after injury, we measured prothrombin
levels in the various conditioned media. After the
addition of ecarin, which converts prothrombin to
thrombin, thrombin-like activity was assayed as before.
In the uninjured optic nerve (Fig. 4A), incubation
with 0.1 U of ecarin did not evoke a thrombin activity
signal. In the injured optic nerve, however, the high
signal observed 1 day after injury was further increased approximately threefold by ecarin, suggesting that the amount of thrombin-like activity in the
nerve after the injury had corresponded to about
30% of the potentially available activity. In the sciatic
nerve (Fig. 4B), incubation with ecarin increased the
signal of thrombin-like activity by more than 10-fold
in both injured and uninjured nerves. Thus, regardless of injury, it appears that less than 10% of the
potentially available thrombin-like activity in the
sciatic nerve exists in the active form.
Thrombin-like activity in fish optic nerve
The kinetics of thrombin-like activity was studied in a
similar manner in the regenerating fish optic nerve.
Figure 5A shows the thrombin-like activity measured
in fish optic nerve on different days after injury as
well as in uninjured nerves. Samples of conditioned
media were incubated with the chromogenic substrate Chromozym-TH and the OD values were measured after 4 h. Injury of the fish optic nerve resulted
in an increase in thrombin-like activity with kinetics
that differed from those observed in the rat. The
activity started to increase on day 1, reached a broad
peak that lasted from day 4 to day 7 after the injury,
and then declined. The observed increase was significant when compared to that in uninjured nerves
(P,0.05).
Specific inhibition of thrombin-like activity in fish
optic nerve
Thrombin-like activity in the fish optic nerve was
inhibited in a dose-dependent manner by the synthetic inhibitor thromstop (Fig. 5B) or hirudin (not
shown), demonstrating its specificity. Much larger
amounts of both inhibitors were needed to achieve
inhibition in the fish nerve than in either of the rat
nerves: for example, to achieve an inhibition of
about 60% in the fish nerve it was necessary to use
100 mM thromstop, a concentration 1000-fold higher
than that needed in the rat. The inhibition pattern
in the fish optic nerve was identical at several examined time points.
Figure 4. Prothrombin levels and thrombin-like activities in
rat optic (A) and sciatic (B) nerves. Nerves were subjected to
crush injury, the rats were killed 1 day later, and conditioned
media were collected from injured and uninjured nerves. For
conversion of prothrombin to thrombin, 0.1 U ecarin was
added prior to the thrombin activity assay. Representative
examples (mean 6sd) of OD values at 405 nm after 0.5 h of
the assay are shown.
Activities of gelatinases in rat optic and sciatic
nerves
To examine whether the observed up-regulation of
thrombin-like activity in the injured optic nerve of
the rat is unique to thrombin or merely part of a
general, nonspecific up-regulation of protease activity after injury (which would imply a general difference in protease regulation between regenerative
and nonregenerative nervous systems), we also measured the postinjury activities of gelatinases. As mentioned earlier, gelatinases are known to participate
in processes that are related to both regeneration
(27–29) and degeneration (30 –33) in the nervous
system. Conditioned media and nerve extracts from
rat optic and sciatic nerves were collected at specified times after injury, as well as from uninjured
nerves, and gel zymography was performed. Figure 6
shows representative gels of gelatinase activities in rat
optic (Fig. 6A) and sciatic (Fig 6B) nerves. In both
cases, gelatinase A (72 kDa) was found to be constitutively expressed and did not increase after injury.
In contrast, gelatinase B (92 kDa) could not be
detected in either of the nerves before injury, but
was up-regulated in both 1 day after injury, after
which its activity declined. In the rat optic nerve, this
up-regulation was detectable as early as 6 h after the
injury (Fig. 6A, lane 2). In some optic nerve preparations, an additional band below 200 kDa was visible
(Fig. 7A, lane 1), probably representing a complex
formed between a gelatinase and an inhibitor. Apart
from the existence of this complex, the kinetics of
the rat sciatic and optic nerves were similar. Identical
results were obtained when nerve extracts were used
instead of conditioned media (not shown).
GELATINASE AND THROMBIN-LIKE ACTIVITIES IN NERVOUS SYSTEMS
537
Figure 5. A) Thrombin-like activity in
fish optic nerve. Nerves were subjected to crush injury, the fish were
killed on the indicated days after
injury, and conditioned media were
collected. Activity-generated elevations in OD at 405 nm after 4 h of
thrombin activity assay are shown.
The bar graph shows the mean values 6sem of OD values obtained
from different preparations, whose
numbers (n) are indicated above
each bar. The increase in thrombinlike activity on days 4 and 7 after the
injury is significant (*P,0.05) compared to that obtained in uninjured
nerves. B) Thrombin-like activity in
the fish optic nerve in the presence
of the specific thrombin inhibitor
thromstop. Fish were killed on day 7 after injury, conditioned medium was collected, and thrombin activity assay was
performed in the presence of different amounts of thromstop. Mean OD values 6sd of a representative example are
shown. Calculation of the inhibition was based on OD elevation at 405 nm after 4 h of thrombin activity assay.
Gelatinase activity is potentially associated with
retrograde degeneration
Having found that thrombin-like activity in the rat
optic nerve is potentially associated with retrograde
degeneration (Fig. 3), we were interested in establishing which parts of the injured optic and sciatic
nerves are responsible for the observed changes in
gelatinase activity. Accordingly, both nerves were
again divided into a proximal segment, which included the site of injury, and a distal segment. Nerve
extracts were prepared from each segment separately. In both optic and sciatic nerves, induction of
gelatinase B was found to occur only in the proximal
segment (Fig. 7). Identical results were obtained
when conditioned media were used (not shown).
Gelatinase activities in the fish optic nerve
Figure 6. Gelatinase activities in rat optic (A) and sciatic (B)
nerve. Nerves were subjected to crush injury, the rats were
killed at the indicated times after the injury, and conditioned
media were collected. Gel zymography in 8% gels was performed with 10 mg of total protein. MMP-2, gelatinase A;
MMP-9, gelatinase B.
538
Vol. 13
March 1999
As with thrombin, we studied the kinetics of gelatinase activities in response to axonal injury in the
regenerating fish optic nerve. As before, conditioned
media and nerve extracts were collected at specified
times after the injury, as well as from uninjured
nerves, and gel zymography was performed. Figure 8
shows representative gels of gelatinase activities in
the fish optic nerve. Two major bands were found at
72 and 92 kDa. Since they were identical in size and
exhibited similar postinjury behavior to that of the
gelatinase activity observed in the rat, they were
assumed to be homologous to rat gelatinases A and
B, respectively. As in the rat, the 72 kDa gelatinase
was found to be constitutively expressed and unaffected by the injury, whereas the 92 kDa gelatinase
was up-regulated. Several additional phenomena,
not present in the rat, were observed in the fish. 1)
Up-regulation of gelatinase activity after injury remained high for much longer than in the rat; for
example, baseline levels of the 92 kDa gelatinase
were not reached even by day 7. 2) In uninjured
nerves, the ratio between the gelatinases at 72 and 92
kDa differed from that obtained in the rat, indicat-
The FASEB Journal
FRIEDMANN ET AL.
ing a greater abundance of the 92 kDa gelatinase in
the fish. 3) Additional bands at about 20, 35, 45, 55,
and 80 kDa were up-regulated upon injury. These
bands might be a result of up-regulation of additional gelatinases or of proteolysis of the known
gelatinases where the cleaved products were still
active. 4) Less protein was needed in the fish conditioned medium in order to achieve a band intensity
similar to that of the rat conditioned medium,
indicating that the activity in the former medium is
higher, either because gelatinases are more abundant or because the activities of the fish gelatinases
are higher than those of the rat. Identical results
were obtained when nerve extracts were used instead
of conditioned media (not shown).
Figure 8. Gelatinase activities in fish optic nerve. Nerves were
subjected to crush injury, the fish were killed on the indicated
days after injury, and conditioned media were collected. A)
Gel zymography in a 7–15% linear gradient gel was performed with 10 mg of total protein. B) Gel zymography in 8%
gel was performed with 2.5 mg of total protein.
DISCUSSION
Figure 7. Gelatinase activities in different parts of rat optic (A)
and sciatic (B) nerve. Nerves were subjected to crush injury,
the rats were killed on the indicated days after injury, and
nerve extracts were collected. Gel zymography in 8% gels was
performed with 30 mg of total protein. P, proximal segment;
D, distal segment; MMP-2, gelatinase A; MMP-9, gelatinase B.
The neuronal response to injury, as manifested by
two types of proteolytic activities, was compared in
the nonregenerative rat optic nerve and two regenerative nerves: the rat sciatic nerve and the fish optic
nerve. In the rat optic nerve, an extremely significant
increase in thrombin-like activity (approximately 40fold compared to that in the uninjured nerve) was
observed 1 day after injury. The increase was transient and started to decline 2 days after injury,
returning to baseline levels on about day 4. In the rat
sciatic nerve, a transient threefold increase in thrombin-like activity was seen 1 day after injury. Although
this increase was about 20-fold smaller than in the
optic nerve, it was still very significant. In both the
optic and the sciatic nerves of rat, the thrombin-like
activity was inhibited in a dose-dependent manner by
two specific inhibitors, thromstop and hirudin, confirming the specificity of the observed signal.
High levels of thrombin activity were previously
shown to inhibit neurite outgrowth (16) and to
induce apoptotic cell death in neurons and astrocytes (20 –22) in vitro. The high levels of thrombinlike activity in the rat optic nerve observed 1 day after
injury in our study suggest that the degenerative
processes seen after CNS injury might be related to
thrombin activity and that thrombin-like activity
might contribute to the nonregenerative nature of
the optic nerve environment. The marked increase
GELATINASE AND THROMBIN-LIKE ACTIVITIES IN NERVOUS SYSTEMS
539
in activity could also override the beneficial effects of
thrombin that are expected (from in vitro data) to
occur in the CNS, such as the stimulation of NGF
secretion by astrocytes (12). In contrast, a slightly
increased thrombin-like activity, such as that seen in
the rat sciatic nerve, might not only be below the
toxicity threshold but might even be beneficial for
the injured nerves. In the PNS, the injured nerve is
rapidly invaded by macrophages, which are crucial
for regeneration in the nervous system (1, 38 – 40).
Thrombin is a potent chemoattractant for human
monocytes (41) and stimulates the production of
monocyte chemotactic proteins (42– 44). In the
mammalian CNS, where macrophage recruitment is
impaired (38) because of an immune brain barrier
(45), these potential effects of thrombin could be
minimized or excluded. In addition, levels of thrombin-like activity in the sciatic nerve could be within
the range at which neurons are rescued from cell
death resulting from metabolic insults, which likely
to be present after nerve injury (20, 21).
Although it cannot be ruled out, the possibility
that the observed thrombin-like activity results from
a protein other than thrombin seems very unlikely,
an assumption supported both by the cleavage of the
specific thrombin substrate Chromozym-TH and the
inhibition by two specific thrombin inhibitors.
Thrombin activity results from the cleavage of prothrombin, which is known to be expressed in the
nervous system (37). The experiments with ecarin
showed that activation of the available prothrombin
would increase thrombin activity by more than 10fold in the injured sciatic nerve, but by only about
3-fold in the injured optic nerve. This suggests that
prothrombin activation is under tighter control in
the sciatic nerve than in the optic nerve. A prothrombin activator independent of the blood coagulation
cascade was recently described, and its involvement
after injury was suggested (46, 47). Although not
detected in uninjured brain, it was found to be active
in a cell line derived from mammalian CNS (46). In
addition, prothrombin activation has been shown to
occur on neuronal surfaces (48). Our data suggest
that different mechanisms of prothrombin activation
in the optic and sciatic nerves could account for the
high levels of thrombin-like activity observed after
injury in the rat optic nerve.
In a recent study of thrombin and prothrombin
elevation in rat sciatic nerve (49), thrombin-generated activity increased significantly from day 1 after
nerve crush, reaching a peak on day 3 and returning
to basal levels on day 6. Prothrombin could not be
detected in uninjured nerves, but the ratio of prothrombin-to-thrombin activity after injury was comparable to the ratio obtained in our study, and led
those authors to suggest a tight control of thrombin
activation in the sciatic nerve. The different kinetics
540
Vol. 13
March 1999
obtained in that study (reflecting a relatively delayed
increase in thrombin activity) and the inability to
detect prothrombin in the uninjured sciatic nerve
might be attributable to the methodology, as nerve
extracts and not conditioned media were used.
Differences in thrombin-like activities between the
rat optic and sciatic nerves could also result from
changes in the levels of thrombin inhibitors. A
known thrombin inhibitor in the nervous system is
the glial-derived protease nexin I (PN-I) (50), which
promotes neurite outgrowth (51, 52) and could
therefore reverse the negative effects of thrombinlike activity. In the sciatic nerve, PN-I is up-regulated
after injury, starting on about day 3 and peaking on
day 7 (49, 53). The behavior of PN-I shortly after
injury in the optic nerve has not been studied.
Nevertheless, our finding of high thrombin-like activity in the optic nerve 1 day after injury may reflect
the paucity or absence of thrombin inhibitors, such
as PN-I, at this time.
In the fish, the kinetics of thrombin-like activity in
the optic nerve differed from those in the rat. As in
the rat, the increase in activity started on day 1 but,
unlike in the rat, continued and reached a plateau
lasting from day 4 to day 7 after injury. As similar
peaking in the fish nerve was also observed in the
activities of the gelatinases, this elevation pattern
may represent a general response of this regenerative nerve to injury. The kinetics observed in the fish
optic nerve point to the establishment of a different
extracellular milieu, which is generated by the proteases and could account, at least in part, for the
regeneration-supportive properties of the neuronal
environment. As in the rat, thrombin-like activity in
the fish optic nerve was specifically inhibited, in a
dose-dependent manner, by the inhibitors thromstop and hirudin. Much larger amounts of these
inhibitors were needed in the fish than in the rat,
indicating the presence of other substances in the
fish conditioned medium that influence the interaction between the inhibitors and the enzyme.
Studies in our laboratory recently demonstrated
the presence of factor XIIIa in the nervous tissues of
rat and fish, as well as a correlation between the
postinjury appearance and activation of the enzyme
and the regenerative ability of the tissue (54).
Thrombin is crucial for the regulation of factor XIIIa
in vivo, and the differential activities of thrombin
might be responsible for the differences in the
observed behavior of factor XIIIa.
We found that the thrombin-like activity in the rat
optic nerve was associated mainly with the part of the
nerve adjacent to and including the site of injury,
i.e., the proximal segment. This suggests that the
observed injury-induced activity is associated with
retrograde degeneration and, accordingly, is evoked
either by changes emanating from the neuronal cell
The FASEB Journal
FRIEDMANN ET AL.
bodies or from a non-neuronal cell response at the
site of injury. As mentioned earlier, prothrombin
mRNA was shown to be expressed by cells of the
nervous system (37). In addition, thrombin has been
detected in brain and astroglial cell cultures (55).
Nevertheless, thrombin synthesis could also occur
elsewhere, for example, in platelets that adhere to
the site of injury. The cellular source responsible for
thrombin activity in the nervous system, and particularly after injury, is not yet known.
In contrast to the differences in thrombin-like
activity, we found that the activation pattern of
gelatinase A or B was similar in all three nerve types.
Gelatinase B was up-regulated after crush injury,
whereas gelatinase A was not affected. In the rat,
up-regulation of gelatinase B (which did not occur in
the uninjured nerve) peaked about 1 day after optic
or sciatic nerve injury and decreased on about day 4.
The fish showed a different proportion between the
two gelatinases in the uninjured nerve, as well as a
relatively delayed peak of the activity. As already
mentioned, this delayed peaking (which is not exclusive to gelatinases) may influence the regenerative
properties of the fish optic nerve.
Gelatinases are known to participate in processes
that are related to both regeneration (27–29) and
degeneration (30 –33) in the nervous system. This
study shows not only that gelatinases are present in
regenerative and nonregenerative white matter but
that, surprisingly, the induction patterns of gelatinases A and B in all three nerve types examined are
similar. One possible explanation is that rather than
being directly related to degeneration or regeneration, the mechanisms of activation or inhibition of
gelatinases in degenerating and regenerating systems might differ. Gelatinase activity can be inhibited by tissue inhibitors of matrix metalloproteases
(TIMPs). TIMP-1 was shown to be induced in the
sciatic nerve on days 1 and 4 postinjury, and immunostaining demonstrated its colocalization with
Schwann cells and macrophages (56). Another possibility is that not only total protease activity, but also
the ratio of ECM-degrading to ECM-producing proteases, might play a role in determining the permissiveness of the extracellular milieu to regeneration.
This is in line with our finding that the nonregenerative rat optic nerve differed significantly from the
regenerative rat sciatic nerve with respect to the
amounts of fibrin-producing thrombin, but not of
ECM-degrading gelatinases.
As with thrombin-like activity, gelatinase activities
in the rat optic and sciatic nerve were limited to the
proximal segment, indicating that the observed activities are associated not with anterograde degeneration, but with an active process occurring at the site
of injury or in the proximal part of the nerve. Our
findings in the sciatic nerve are in line with the
observations of La Fleur et al. (56). The cellular
source of the gelatinases in vivo remains to be
established. However, the inflammation-related cytokines interleukin (IL) -1a, IL-1b, and TNF-a are
potent inducers of gelatinases in cultured rat astrocytes (57), and activated microglia release gelatinase
B in vitro (58). Gelatinase activity has also been
reported in Schwann cells and glioma cells (59, 60).
In summary, a comparison of the general kinetics
of the induced thrombin-like and gelatinase activities in rat and fish showed that activity in the rat
peaks early and declines between days 1 and 4
postinjury, whereas in the fish the activity lasts for at
least a week. A comparison of activity levels in the
optic and sciatic nerves of the rat showed significant
differences in thrombin-like activity but similarity in
the activities of gelatinases. The induction pattern of
the 72 kDa and 92 kDa gelatinases was similar in fish
as well. Both the different kinetics and the differential activity levels might affect the nerve’s ability to
regenerate. This study further highlights the fact
that tissue repair and regeneration are closely linked
and that substances that generally participate in
wound healing establish an extracellular milieu that
can lead either to regeneration or its failure.
We thank Shirley Smith for editorial assistance and Haya
Avital for help with graphics. M.S. holds the Maurice and Ilse
Katz Professorial Chair in Neurobiology.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
GELATINASE AND THROMBIN-LIKE ACTIVITIES IN NERVOUS SYSTEMS
Lotan, M., and Schwartz, M. (1994) Cross talk between the
immune system and the nervous system in response to injury:
implications for regeneration. FASEB J. 8, 1026 –1033
Weinberg, E. L., and Spencer, P. S. (1979) Studies on the
control of myelinogenesis. 3. Signalling of oligodendrocyte
myelination by regenerating peripheral axons. Brain Res. 162,
273–279
Richardson, P. M., McGuinness, U. M., and Aguayo, A. J. (1980)
Axons from CNS neurons regenerate into PNS grafts. Nature
(London) 284, 264 –265
Faber Elman, A., Miskin, R., and Schwartz, M. (1995) Components of the plasminogen activator system in astrocytes are
modulated by tumor necrosis factor-alpha and interleukin-1
beta through similar signal transduction pathways. J. Neurochem. 65, 1524 –1535
Faber Elman, A., Solomon, A., Abraham, J. A., Marikovsky, M.,
and Schwartz, M. (1996) Involvement of wound-associated factors in rat brain astrocyte migratory response to axonal injury: in
vitro simulation. J. Clin. Invest. 97, 162–171
Raghow, R. (1994) The role of extracellular matrix in postinflammatory wound healing and fibrosis. FASEB J. 8, 823– 831
Coughlin, S. R., Vu, T. K., Hung, D. T., and Wheaton, V. I.
(1992) Characterization of a functional thrombin receptor.
Issues and opportunities. J. Clin. Invest. 89, 351–355
Monard, D. (1988) Cell-derived proteases and protease inhibitors as regulators of neurite outgrowth. Trends Neurosci. 11,
541–544
Halkier, T. (1991) Mechanisms in Blood Coagulation, Fibrinolysis and the Complement System. Cambridge University Press,
Cambridge
Cavanaugh, K. P., Gurwitz, D., Cunningham, D. D., and Bradshaw, R. A. (1990) Reciprocal modulation of astrocyte stellation
by thrombin and protease nexin-1. J. Neurochem. 54, 1735–
1743
541
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
542
Perraud, F., Besnard, F., Sensenbrenner, M., and Labourdette,
G. (1987) Thrombin is a potent mitogen for rat astroblasts but
not for oligodendroblasts and neuroblasts in primary culture.
Int. J. Dev. Neurosci. 5, 181–188
Neveu, I., Jehan, F., Jandrot Perrus, M., Wion, D., and Brachet,
P. (1993) Enhancement of the synthesis and secretion of nerve
growth factor in primary cultures of glial cells by proteases: a
possible involvement of thrombin. J. Neurochem. 60, 858 – 867
Lipton, S. A., Wagner, J. A., Madison, R. D., and D’Amore, P. A.
(1988) Acidic fibroblast growth factor enhances regeneration of
processes by postnatal mammalian retinal ganglion cells in
culture. Proc. Natl. Acad. Sci. USA 85, 2388 –2392
Lobb, R. R. (1988) Thrombin inactivates acidic fibroblast
growth factor but not basic fibroblast growth factor. Biochemistry 27, 2572–2578
Loret, C., Sensenbrenner, M., and Labourdette, G. (1989)
Differential phenotypic expression induced in cultured rat
astroblasts by acidic fibroblast growth factor, epidermal growth
factor, and thrombin. J. Biol. Chem. 264, 8319 – 8327
Gurwitz, D., and Cunningham, D. D. (1988) Thrombin modulates and reverses neuroblastoma neurite outgrowth. Proc. Natl.
Acad. Sci. USA 85, 3440 –3444
Suidan, H. S., Stone, S. R., Hemmings, B. A., and Monard, D.
(1992) Thrombin causes neurite retraction in neuronal cells
through activation of cell surface receptors. Neuron 8, 363–375
Debeir, T., Benavides, J., and Vige, X. (1996) Dual effects of
thrombin and a 14-amino acid peptide agonist of the thrombin
receptor on septal cholinergic neurons. Brain Res 708, 159 –166
Donovan, F. M., Pike, C. J., Cotman, C. W., and Cunningham,
D. D. (1997) Thrombin induces apoptosis in cultured neurons
and astrocytes via a pathway requiring tyrosine kinase and RhoA
activities. J. Neurosci. 17, 5316 –5326
Vaughan, P. J., Pike, C. J., Cotman, C. W., and Cunningham,
D. D. (1995) Thrombin receptor activation protects neurons
and astrocytes from cell death produced by environmental
insults. J. Neurosci. 15, 5389 –5401
Donovan, F. M., and Cunningham, D. D. (1998) Signaling
pathways involved in thrombin-induced cell protection. J. Biol.
Chem. 273, 12746 –12752
Smirnova, I. V., Zhang, S. X., Citron, B. A., Arnold, P. M., and
Festoff, B. W. (1998) Thrombin is an extracellular signal that
activates intracellular death proteases pathways inducing apoptosis in model motor neurons. J. Neurobiol. 36, 64 – 80
Woessner, J. F., Jr. (1991) Matrix metalloproteinases and their
inhibitors in connective tissue remodeling. FASEB J. 5, 2145–
2154
Krane, S. M. (1994) Clinical importance of metalloproteinases
and their inhibitors. Ann. N.Y. Acad. Sci. 732, 1–10
Stamatoglou, S. C., Hughes, R. C., and Lindahl, U. (1987) Rat
hepatocytes in serum-free primary culture elaborate an extensive extracellular matrix containing fibrin and fibronectin.
J. Cell Biol. 105, 2417–2425
Williams, L. R., Longo, F. M., Powell, H. C., Lundborg, G., and
Varon, S. (1983) Spatial-temporal progress of peripheral nerve
regeneration within a silicone chamber: parameters for a bioassay. J. Comp. Neurol. 218, 460 – 470
Knoops, B., Hubert, I., Hauw, J. J., and van den Bosch de
Aguilar, P. (1991) Axonal growth and glial migration from
cocultured hippocampal and septal slices into fibrin-fibronectin-containing matrix of peripheral regeneration chambers: a
light and electron microscope study. Brain Res. 540, 183–194
Pittman, R. N., and Buettner, H. M. (1989) Degradation of
extracellular matrix by neuronal proteases. Dev. Neurosci. 11,
361–375
Muir, D. (1994) Metalloproteinase-dependent neurite outgrowth within a synthetic extracellular matrix is induced by
nerve growth factor. Exp. Cell. Res. 210, 243–252
Rosenberg, G. A., Dencoff, J. E., McGuire, P. G., Liotta, L. A.,
and Stetler Stevenson, W. G. (1994) Injury-induced 92-kilodalton gelatinase and urokinase expression in rat brain. Lab.
Invest. 71, 417– 422
Rosenberg, G. A., Estrada, E. Y., Dencoff, J. E., and Stetler
Stevenson, W. G. (1995) Tumor necrosis factor-alpha-induced
gelatinase B causes delayed opening of the blood– brain barrier:
an expanded therapeutic window. Brain Res. 703, 151–155
Gijbels, K., Masure, S., Carton, H., and Opdenakker, G. (1992)
Gelatinase in the cerebrospinal fluid of patients with multiple
Vol. 13
March 1999
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
sclerosis and other inflammatory neurological disorders. J. Neuroimmunol. 41, 29 –34
Gijbels, K., Proost, P., Masure, S., Carton, H., Billiau, A., and
Opdenakker, G. (1993) Gelatinase B is present in the cerebrospinal fluid during experimental autoimmune encephalomyelitis and cleaves myelin basic protein. J. Neurosci. Res. 36,
432– 440
Duvdevani, R., Rosner, M., Belkin, M., Sautter, J., Sabel, B. A.,
and Schwartz, M. (1990) Graded crush of the rat optic nerve as
a brain injury model: Combining electrophysiological and behavioral outcome. Restor. Neurol. Neurosci. 2, 31–38
Lottenberg, R., Hall, J. A., Blinder, M., Binder, E. P., and
Jackson, C. M. (1983) The action of thrombin on peptide
p-nitroanilide substrates. Substrate selectivity and examination
of hydrolysis under different reaction conditions. Biochim.
Biophys. Acta 742, 539 –557
Gallagher, S. R. (1997) One dimensional gel electrophoresis of
proteins. In: Current Protocols in Molecular Biology (Ausubel,
F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G.,
Smith, J. A., and Struhl, K., eds) Unit 10.12, Greene Publishing
and Wiley-Interscience, New York
Dihanich, M., Kaser, M., Reinhard, E., Cunningham, D., and
Monard, D. (1991) Prothrombin mRNA is expressed by cells of
the nervous system. Neuron 6, 575–581
Perry, V. H., Brown, M. C., and Gordon, S. (1987) The macrophage response to central and peripheral nerve injury. A
possible role for macrophages in regeneration. J. Exp. Med. 165,
1218 –1223
Rapalino, O., Lazarov-Spiegler, O., Agranov, E., Velan, G. J.,
Fraidakis, M., Yoles, E., Solomon, A., Gepstein, R., Katz, A.,
Belkin, M., Hadani, M., and Schwartz, M. (1998) Implantation
of stimulated homologous macrophages results in partial recovery of paraplegic rats. Nature Med. 4, 814 – 821
Lazarov Spiegler, O., Solomon, A. S., Zeev Brann, A. B.,
Hirschberg, D. L., Lavie, V., and Schwartz, M. (1996) Transplantation of activated macrophages overcomes central nervous
system regrowth failure. FASEB J. 10, 1296 –1302
Bar Shavit, R., Kahn, A., Wilner, G. D., and Fenton, J. W. D.
(1983) Monocyte chemotaxis: stimulation by specific exosite
region in thrombin. Science 220, 728 –731
Sower, L. E., Froelich, C. J., Carney, D. H., Fenton, J. W. N., and
Klimpel, G. R. (1995) Thrombin induces IL-6 production in
fibroblasts and epithelial cells. Evidence for the involvement of
the seven-transmembrane domain (STD) receptor for alphathrombin. J. Immunol. 155, 895–901
Shankar, R., de la Motte, C. A., and DiCorleto, P. E. (1992)
Thrombin stimulates PDGF production and monocyte adhesion
through distinct intracellular pathways in human endothelial
cells. Am. J. Physiol. 262, C199 –C206
Grandaliano, G., Valente, A. J., and Abboud, H. E. (1994) A
novel biologic activity of thrombin: stimulation of monocyte
chemotactic protein production. J. Exp. Med. 179, 1737–1741
Hirschberg, D. L., and Schwartz, M. (1995) Macrophage recruitment to acutely injured central nervous system is inhibited by a
resident factor: a basis for an immune-brain barrier. J. Neuroimmunol. 61, 89 –96
Sekiya, F., Usui, H., Inoue, K., Fukudome, K., and Morita, T.
(1994) Activation of prothrombin by a novel membrane-associated protease. An alternative pathway for thrombin generation
independent of the coagulation cascade. J. Biol. Chem. 269,
32441–32445
Shikamoto, Y., Shibusawa, S., Okuyama, I., and Morita, T.
(1997) Characterization of membrane-associated prothrombin
activator in normal and injured murine tissues. FEBS Lett. 412,
526 –530
Grabham, P. W., Monard, D., Gallimore, P. H., and Grand,
R. J. A. (1991) Modulation of human neurite outgrowth by
serine proteases: a comparison of the interaction of thrombin
and prothrombin with glia-derived nexin. Eur. J. Neurosci. 3,
663– 668
Smirnova, I. V., Ma, J. Y., Citron, B. A., Ratzlaff, K. T., Gregory,
E. J., Akaaboune, M., and Festoff, B. W. (1996) Neural thrombin
and protease nexin I kinetics after murine peripheral nerve
injury. J. Neurochem. 67, 2188 –2199
Gloor, S., Odink, K., Guenther, J., Nick, H., and Monard, D.
(1986) A glia-derived neurite promoting factor with protease
The FASEB Journal
FRIEDMANN ET AL.
51.
52.
53.
54.
55.
56.
inhibitory activity belongs to the protease nexins. Cell 47,
687– 693
Monard, D., Niday, E., Limat, A., and Solomon, F. (1983)
Inhibition of protease activity can lead to neurite extension in
neuroblastoma cells. Prog. Brain Res. 58, 359 –364
Zurn, A. D., Nick, H., and Monard, D. (1988) A glia-derived
nexin promotes neurite outgrowth in cultured chick sympathetic neurons. Dev. Neurosci. 10, 17–24
Meier, R., Spreyer, P., Ortmann, R., Harel, A., and Monard, D.
(1989) Induction of glia-derived nexin after lesion of a peripheral nerve. Nature (London) 342, 548 –550
Monsonego, A., Mizrahi, T., Moalem, G., Bardos, H., Adany, R.,
and Schwartz, M. (1998) Factor XIIIa as a nerve-associated
transglutaminase. FASEB J. 12, 1163–1171
Deschepper, C. F., Bigornia, V., Berens, M. E., and Lapointe,
M. C. (1991) Production of thrombin and antithrombin III by
brain and astroglial cell cultures. Brain Res. Mol. Brain. Res. 11,
355–358
La Fleur, M., Underwood, J. L., Rappolee, D. A., and Werb, Z.
(1996) Basement membrane and repair of injury to peripheral
nerve: defining a potential role for macrophages, matrix metal-
57.
58.
59.
60.
GELATINASE AND THROMBIN-LIKE ACTIVITIES IN NERVOUS SYSTEMS
loproteinases, and tissue inhibitor of metalloproteinases-1. J.
Exp. Med. 184, 2311–2326
Gottschall, P. E., and Deb, S. (1996) Regulation of matrix
metalloproteinase expressions in astrocytes, microglia and neurons. Neuroimmunomodulation 3, 69 –75
Gottschall, P. E., Yu, X., and Bing, B. (1995) Increased production of gelatinase B (matrix metalloproteinase-9) and interleukin-6 by activated rat microglia in culture. J. Neurosci. Res 42,
335–342
Apodaca, G., Rutka, J. T., Bouhana, K., Berens, M. E., Giblin,
J. R., Rosenblum, M. L., McKerrow, J. H., and Banda, M. J.
(1990) Expression of metalloproteinases and metalloproteinase
inhibitors by fetal astrocytes and glioma cells. Cancer Res. 50,
2322–2329
Muir, D. (1995) Differences in proliferation and invasion by
normal, transformed and NF1 Schwann cell cultures are influenced by matrix metalloproteinase expression. Clin. Exp. Metastasis 13, 303–314
Received for publication July 27, 1998.
Revised for publication October 29, 1998.
543