Báo cáo khoa học: Thrombin-mediated impairment of fibroblast growth factor-2 activity doc

Thrombin-mediated impairment of fibroblast growth
factor-2 activity
Pierangela Totta
1,
*, Raimondo De Cristofaro
2,
*, Claudia Giampietri
3
, Maria S. Aguzzi
1
,
Debora Faraone
1
, Maurizio C. Capogrossi
1
and Antonio Facchiano
1
1 Laboratorio di Patologia Vascolare, IDI-IRCCS, Istituto Dermopatico dell’Immacolata-Istituto di Ricovero e Cura a Carattere Scientifico,
Rome, Italy
2 Institute of Internal Medicine, Haemostasis Research Center, Catholic University School of Medicine, Rome, Italy
3 Department of Histology and Medical Embriology, University of Rome ‘Sapienza’, Italy
Fibroblast growth factor (FGF)-2 belongs to the
23-member family of FGFs [1]. It is known as one of
the most potent angiogenic factors controlling embry-
onic development [2], tissue remodeling [3], stem cell
physiology [4] and tumor growth [5]. FGF-2 activity is
finely modulated at several levels, and recent evidence
shows that different FGF-2 concentrations may exert
opposing effects [6]. Studies have shown that the
interaction of several molecules with this growth factor
[7–9] or with its receptors [10,11] participates in the

control of FGF-2 activity. A few studies have
examined FGFs proteolytic degradation [12–15]. For
example, FGF-2 degradation by the zinc-endopro-
tease neprilysin has been recently demonstrated. This
Keywords
cell proliferation; digestion; fibroblast growth
factor-2; maturation; thrombin
Correspondence
A. Facchiano, Laboratorio di Patologia
Vascolare, Istituto Dermopatico
dell’Immacolata, IDI-IRCCS, Via Monti di
Creta 104, 00167 Rome, Italy
Fax: +39 06 6646 2430
Tel: +39 06 6646 2431
E-mail:
*These authors contributed equally to this
work
(Received 6 October 2008, revised 19
March 2009, accepted 6 April 2009)
doi:10.1111/j.1742-4658.2009.07042.x
Thrombin generation increases in several pathological conditions, including
cancer, thromboembolism, diabetes and myeloproliferative syndromes.
During tumor development, thrombin levels increase along with several
other molecules, including cytokines and angiogenic factors. Under such
conditions, it is reasonable to predict that thrombin may recognize new
low-affinity substrates that usually are not recognized under low-expression
levels conditions. In the present study, we hypothesized that fibroblast
growth factor (FGF)-2 may be cleaved by thrombin and that such action
may lead to an impairment of its biological activity. The evidence collected
in the present study indicates that FGF-2-induced proliferation and chemo-

taxis ⁄ invasion of SK-MEL-110 human melanoma cells were significantly
reduced when FGF-2 was pre-incubated with active thrombin. The inhibi-
tion of proliferation was not influenced by heparin. Phe-Pro-Arg-chlorom-
ethyl ketone, a specific inhibitor of the enzymatic activity of thrombin,
abolished the thrombin-induced observed effects. Accordingly, both
FGF-2-binding to cell membranes as well as FGF-2-induced extracellular
signal-regulated kinase phosphorylation were decreased in the presence of
thrombin. Finally, HPLC analyses demonstrated that FGF-2 is cleaved by
thrombin at the peptide bond between residues Arg42 and Ile43 of the
mature human FGF-2 sequence. The apparent k
cat
⁄ K
m
of FGF-2 hydroly-
sis was 1.1 · 10
4
m
)1
Æs
)1
, which is comparable to other known low-affinity
thrombin substrates. Taken together, these results demonstrate that throm-
bin digests FGF-2 at the site Arg42-Ile43 and impairs FGF-2 activity
in vitro, indicating that FGF-2 is a novel thrombin substrate.
Abbreviations
ERK, extracellular signal-regulated kinase; FGF, fibroblast growth factor; HMW, high molecular weight; HUVEC, human umbilical vein
endothelial cell line; LMW, low molecular weight; PAR, protease-activated receptor; PPACK, Phe-Pro-Arg-chloromethyl ketone; TRAP,
thrombin receptor-activating peptide.
FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS 3277
metalloprotease was found to cleave the Leu135-

Gly136 peptide bond of FGF-2, severely inhibiting its
angiogenic activity, as demonstrated in a murine cor-
neal pocket angiogenesis model [12]. In addition, more
recently, high molecular weight (HMW) FGF-2 was
shown to be cleaved by thrombin, and its degradation
product stimulated endothelial cell migration and pro-
liferation in a similar manner to low molecular weight
(LMW) FGF-2 [13]. Moreover, fibrinogen and fibrin,
by a direct interaction, are known to protect FGF-2
from in vitro proteolytic degradation induced by tryp-
sin and chymotrypsin [16,17]. Finally, different FGF-2
fragments inhibit FGF-2 [18,19], further suggesting
that proteolytic processing of FGF-2 may represent an
endogenous way to modulate its activity.
Thrombin is a serine protease generated from its
zymogen precursor prothrombin after endothelial cell
damage and induction of the coagulation cascade [20].
Its activation is known to be increased in thrombo-
embolism [21], diabetes [22] and cancer [23]. Thrombin
pro-coagulant activity converts fibrinogen to fibrin
monomer, which then polymerizes to form the fibrous
matrix of blood clots [24]. Moreover, thrombin cleaves
protease-activated receptor (PAR)-1 and PAR-4, which
are expressed on human platelet membranes, activating
their hemostatic properties [25]. In addition to these
clot-promoting activities, thrombin down-regulates its
own generation through activation of the protein C
pathway. Activated protein C inactivates cofactors Va
and VIIIa, thereby blunting further thrombin genera-
tion [26]. Thrombin also participates directly in its

final inhibition and clearance from the circulation by
specifically recognizing the serine protease inhibitors
(serpins) antithrombin and heparin cofactor II [27].
Thrombin interaction with PAR-1 and PAR-4 does
not activate only hemostatic functions. Indeed, these
receptors are expressed on the membrane of different
cell types, including fibroblast [27], endothelial [28] and
cancer cells [29], and their activation enhances cytokine
release, cell permeability and cell growth. Furthermore,
thrombin recognizes several other membrane receptors
and substrates, such as platelet glycoprotein Ib and
glycoprotein V [30], and additional noncanonical
thrombin substrates, even in the intracellular compart-
ment, have been identified [31,32].
Taken together, these data indicate that thrombin
recognizes a complex substrates network with several
biological functions, in addition to the classical clot-
promoting effects, and prompted us to investigate
additional substrates not directly involved in blood
coagulation. In the present study, we show, for the
first time, that thrombin digests the 18 kDa LMW
isoform of human FGF-2, modulating its biological
activities such as in vitro cell proliferation and chemo-
taxis induction. In addition, we also identified the
cleavage site on FGF-2.
Results
FGF-2-induced proliferation of SK-MEL-110 is
inhibited by thrombin
Human metastatic melanoma cell line SK-MEL-110
was chosen as a model to test the FGF-2 mitogenic and

chemotactic activity in vitro. Figure 1A shows that 1 h
of thrombin pre-incubation significantly reduces FGF-
2-induced cell growth as a function of concentration in
48 h proliferation assays. The inhibitory action reached
a plateau at 0.1 nm thrombin; therefore, all the experi-
ments were carried out at this thrombin concentration,
which corresponds to 0.01 UÆmL
)1
. Figure 1A shows
that thrombin activity was blocked with the irreversible
selective thrombin inhibitor Phe-Pro-Arg-chloromethyl
ketone (PPACK). As a control, thrombin and PPACK,
alone or mixed, do not affect cell proliferation under
these experimental conditions (Fig. 1A, inset).
Time-course proliferation experiments were then car-
ried out. Pre-treating FGF-2 with thrombin for 1 h was
sufficient to completely inhibit the mitogenic effect of
FGF-2 at 48 and 72 h of proliferation (Fig. 1B). The
experiments shown in Fig. 1 were performed by pre-
incubating FGF-2 with active thrombin; next, before
exposing cells to this mixture, PPACK was added to
block the enzyme. This protocol was chosen to exclude
the possibility that thrombin enzymatic action may
interfere with cell growth by cleaving and activating
PARs receptors, which are known to be expressed in
SK-MEL-110 cells (data not shown) and in other mela-
noma cells [29]. To further rule out this possibility, we
exposed SK-MEL-110 cells to the action of specific
PARs agonists, namely thrombin receptor-activating
peptide (TRAP)-1 and TRAP-4, to show that specifi-

cally activating thrombin-receptors does not itself
determine any inhibition of cell-proliferation. Fig-
ure 2A shows that the specific agonists TRAP-1 alone
and TRAP-4 alone induced some proliferation of mela-
noma cells in the absence of other mitogenic stimuli,
whereas they did not affect the FGF-2-induced prolifer-
ation (Fig. 2B). These data allowed us to exclude the
possibility that thrombin agonism may per se inhibit
SK-MEL-110 proliferation in vitro. Therefore, to mimic
more closely the conditions occurring under in vivo
conditions, cells were directly exposed to the mixture
containing FGF-2 and active thrombin. Figure 3 shows
that, under these conditions, SK-MEL-110 proliferate
significantly less than cells exposed to FGF-2 alone and
FGF-2 is a thrombin substrate P. Totta et al.
3278 FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS
that such inhibition was absent in the presence of inac-
tive thrombin (i.e. thrombin pre-incubated with
PPACK), further confirming that the enzymatic activity
of thrombin inhibits the mitogenic action of FGF-2.
To evaluate the influence of heparan sulfate proteogly-
cans on thrombin susceptibility of FGF-2, the effect of
heparin was investigated in proliferation assays. Hepa-
rin alone (at 10 and 50 nm) did not influence the spon-
taneous proliferation of SK-MEL-110 (Fig. 4A). Thus,
the effect of heparin (at 50 nm) was tested in the pres-
ence of FGF-2 and thrombin. Figure 4B shows that
heparin reduces the mitogenic activity of FGF-2; how-
ever, the inhibitory effect induced by thrombin was
maintained both in the absence and in the presence of

heparin, indicating that, at these doses, heparin does
not influence the observed thrombin–FGF-2 interplay.
As a specificity control, similar experiments were car-
ried out on a different cellular model. Figure 5 shows
that, similar to melanoma cells, the mitogenic effect of
FGF-2 was inhibited by thrombin (0.1 nm) in the
primary human umbilical vein endothelial cell line
(HUVEC). Taken together, these findings indicate that
the mitogenic activity of FGF-2 is controlled by throm-
bin and that FGF-2 may be a thrombin substrate.
FGF-2-induced SK-MEL-110 invasion/migration
on different matrices
We then investigated whether thrombin modulates the
ability of FGF-2 to induce cell invasion through differ-
ent matrices. Invasion assays were carried out in vitro,
in modified Boyden chambers, on filters coated either
with vitronectin, collagen IV or fibronectin. Migration
0 h
48 h
25
50
75
100

A

B
Cell number (%)
+ + + – – –
FGF-2

Thrombin (n
M
) 0.1
PPACK
+
0.1 –
–
–
– –
–
–
–
–
0
25
50
75
100
Cell number (%)
–
FGF-2
0.001
Thrombin (n
M)
0.01 1–
+++++
0.1–
–
PPACK
+++++

*
*
**
**
Cell number (%)
0
100
300
500
700
900
1100
1300
0
24 48 72
Time
(
h
)

*
**
BSA
BSA + PPACK
Thrombin
Thrombin + PPACK
FGF2 + PPACK
FGF2 + Thrombin + PPACK
Fig. 1. Dose–response and time-course proliferation assays. (A)
Dose–response proliferation assay: SK-MEL-110 (4 · 10

4
) cells
were seeded in six-well plates and grown for 24 h at 37 °C, 5%
CO
2
in complete medium. Medium was then replaced and cells
were starved overnight with incomplete medium. Subsequently,
cells were stimulated for 48 h with medium 0.1% BSA, FGF-2
(10 ngÆmL
)1
, 0.6 nM) or thrombin (0.1 nM) pre-incubated for 1 h at
37 °C. Other cells were stimulated with medium 0.1% BSA con-
taining FGF-2 alone (10 ngÆmL
)1
, 0.6 nM), thrombin alone (0.1 nM)
or FGF-2 (10 ngÆmL
)1
, 0.6 nM) with different thrombin concentra-
tions (0.001, 0.01, 0.1 and 1 n
M), pre-incubated for 1 h at 37 °C
and then supplemented with PPACK (50 n
M) to block the enzymatic
activity of thrombin. The mitogenic effect of FGF-2 is significantly
reduced in the presence of different concentrations of thrombin
(*P < 0.005; **P < 0.001). Neither thrombin alone nor thrombin
with PPCAK nor PPACK alone influenced SK-MEL-110 growth with
respect to control (A, inset). Data are expressed as a percentage of
cell number versus cells treated with FGF-2 alone (100% corre-
sponds to 1.9 · 10
5

cells). Data reported are the mean ± SEM of
five independent experiments carried out in duplicate. (B) Time-
course proliferation assay: SK-MEL-110 (4 · 10
4
) cells were seeded
and grown for 24 h in complete medium. Medium was then
replaced and cells were starved overnight with incomplete med-
ium. Subsequently, cells were stimulated for 24, 48 and 72 h with
the indicated stimuli, with doses as described in (A). FGF-2-induced
SK-MEL-110 proliferation is inhibited when FGF-2 is pre-incubated
with thrombin alone for 1 h and then thrombin is blocked with
PPACK (*P < 0.0005; **P < 0.05). Both thrombin alone and throm-
bin + PPACK do not influence SK-MEL-110 cell proliferation com-
pared to BSA and BSA + PPACK. Data are expressed as a
percentage of cell number versus cell number at t
0
(100% corre-
sponds to 1.9 · 10
5
cells). Data reported are the mean ± SEM of
five independent experiments carried out in duplicate.
P. Totta et al. FGF-2 is a thrombin substrate
FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS 3279
assays were also carried out using gelatin-coated filters.
The invasion ⁄ chemoattractant properties of FGF-2
were markedly and significantly reduced in the presence
of thrombin (0.1 nm, corresponding to 0.01 UÆmL
)1
)
on all tested matrices (Fig. 6A–D), with different

potency. We observed approximately 50% inhibition
on vitronectin (Fig. 6A), approximately 30% inhibition
on collagen IV and fibronectin (Fig. 6B,C) and approx-
imately 40% inhibition on gelatin (Fig. 6D). These data
indicate that thrombin strongly inhibits the chemotactic
and invasion properties of FGF-2.
FGF-2-dependent extracellular signal-regulated
kinase (ERK)/mitogen-activated protein kinase
phosphorylation is reduced by thrombin
Mitogen-activated protein kinases and ERK1 ⁄ 2 are
important transducers of mitogenic and differentiation
signals induced by FGF-2 and are often altered in mel-
anoma progression [33]. We therefore investigated
whether FGF-2 pre-incubation with thrombin alters the
ability to activate ERK1 ⁄ 2 in our experimental model.
Figure 7A shows that ERK1 ⁄ 2 phosporylation
induced by FGF-2 is reduced when FGF-2 is incu-
bated with thrombin. The lower band corresponding
to the 42 kDa form (i.e. ERK2) was found to be
reduced, as also revealed in densitometry analysis
(Fig. 7B), suggesting that thrombin incubation impairs
the ability of FGF-2 to signal toward one of the key
0 h
48 h
–
0
50
100
150
200

AB
Cell number (%)
0
50
100
150
200
Cell number (%)
FGF-2
TRAP-1
TRAP-4
–
+
–
–
–
–
–
–
–
+
–
*
+
–
+
–
–
–
+

–
–
+
+
–
–
–
–
FGF-2
TRAP-1
TRAP-4
0 h
48 h
Fig. 2. TRAP-treatment of SK-MEL-110 cells. SK-MEL-110 (4 · 10
4
) cells were seeded and grown for 24 h in complete medium as
described in Fig. 1. Medium was then replaced and cells were starved overnight with incomplete medium. Then, cells were directly exposed
(48 h) to FGF-2 (10 ngÆmL
)1
) in the presence or the absence of 5.7 lM TRAP-1 or TRAP-4. (A) TRAP-1 and TRAP-4 induced some prolifera-
tion as compared to control (*P < 0.05). The mitogenic effect of FGF-2 was not influenced by TRAP-1 or TRAP-4 (B). Data are expressed as
a percentage of cell number versus cells treated with FGF-2 (100% corresponds to 1.9 · 10
4
cells). The data reported are the mean ± SEM
of five independent experiments carried out in duplicate.
0
20
40
60
80

100
Cell number (%)
FGF-2
Thrombin
PPACK
0 h
48 h
–
+
++
+
+
–
+
+
+
–
+–
+
–
+
–
–
–
–
–
*
Fig. 3. Influence of PPACK-thrombin in SK-MEL-110 proliferation.
SK-MEL-110 (4 · 10
4

) cells were seeded and grown for 24 h in
complete medium. Medium was then replaced and cells were
starved overnight with incomplete medium. Then, cells were
directly exposed for 48 h to FGF-2 (10 ngÆmL
)1
) in the presence or
absence of thrombin (0.1 n
M) pre-incubated or not with PPACK
(50 n
M) or PPACK alone (50 nM). As a control, cells were stimulated
with thrombin alone (0.1 n
M) or PPACK alone (50 nM). FGF-2-
induced proliferation was measured after 48 h stimulation with
thrombin in the presence or in the absence of PPACK. PPACK
blocks thrombin enzymatic activity and reverts the thrombin anti-
mitogenic effect (*P < 0.01), suggesting that FGF-2 is degraded by
thrombin protease function. Data are expressed as a percentage of
cell number versus FGF-2 (100% corresponds to 1.9 · 10
5
cells).
The data reported are the mean ± SEM of five independent experi-
ments carried out in duplicate.
FGF-2 is a thrombin substrate P. Totta et al.
3280 FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS
intracellular signal pathways mediating the biological
activity of FGF-2 [34–36].
Proteolytic degradation of FGF-2 induced by
thrombin
According to data reported above, we hypothesized
that FGF-2 may be directly cleaved by thrombin. Its

degradation was then investigated by HPLC analysis
(Fig. 8A) and by Western blotting (see, Fig. S1). Fig-
ure 8A shows the elution chromatogram of FGF-2
alone and FGF-2 incubated with thrombin for 30 min.
Incubation of FGF-2 with thrombin lowered the peak
corresponding to the full length FGF-2, eluting at
23.5 min, and induced the appearance of the peak
eluting at approximately 13.6 min, corresponding to a
proteolytic fragment of FGF-2. Area values corre-
sponding to the full length FGF-2 were then fitted
according to Eqn (1) (see Experimental procedures),
which allows calculation of the kinetic rate constant of
the peak’s area decay, equal to 4.8 ± 0.3 · 10
2
min
)1
.
This value, under our experimental conditions, reflects
an apparent k
cat
⁄ K
m
of FGF-2 hydrolysis equal to
1.1 · 10
4
m
)1
Æs
)1
, which is comparable to the value of

other low-affinity thrombin substrates, such as zymo-
gen protein C or thrombin-activatable fibrinolysis
inhibitor [37,38]. As shown in the (Fig. S1), FGF-2
incubated with PPACK appears as a main unique
band (arrowhead, 18 kDa), whereas one immunoreac-
tive additional band appears in the FGF-2 incubated
with thrombin for 1 h (1 : 1 molar ratio) with an
apparent molecular weight of 15 kDa. FGF-2 incu-
bated with inactive thrombin (i.e. thrombin pre-incu-
bated with PPACK) shows no additional bands
compared to basal conditions. We then confirmed the
observed biochemical degradation of FGF-2 induced
by thrombin with a biological assay. Structural degra-
dation and loss of function were then investigated as a
0 h
24 h
FGF-2
Thrombin
*
0
25
50
75
100
Cell number (%)
–
–
+
+
+

+–
––
–
Fig. 5. HUVEC proliferation assay. HUVEC (8 · 10
4
) were seeded
and grown for 24 h in complete medium. Medium was then
replaced and cells were starved overnight with incomplete med-
ium. HUVEC growth was examined after 24 h of stimulation with
FGF-2 (10 ngÆmL
)1
) in the presence or in the absence of thrombin
(0.1 n
M). The mitogenic effect of FGF-2 is significantly reduced in
the presence of thrombin and thrombin alone has no effect on
HUVEC proliferation. (*P < 0.01). Data are expressed as a percent-
age of cell number versus FGF-2 (100% corresponds to
12.7 · 10
4
cells) and are the mean ± SEM of five independent
experiments carried out in duplicate.
0
25
50
75
100
Cell number (%)
–Thrombin
+Thrombin
FGF-2

Heparin (n
g
·mL
–1
)
–
–
–
+
50
+
*
*
48 h
0 h
10
50
–
0
25
50
75
100A
B
FGF-2
Heparin (ng·mL
–1
)
–
–

–
– –
Fig. 4. Heparin does not influence the effect of thrombin on FGF-2.
SK-MEL-110 (4 · 10
4
) cells were seeded and grown for 24 h in
complete medium. Medium was then replaced and cells were
starved overnight with incomplete medium. Cells were stimulated
for 48 h with two different heparin concentrations (10 and
50 ngÆmL
)1
). At either dose, heparin alone does not influence spon-
taneous SK-MEL-110 growth (A); FGF-2-induced proliferation is par-
tially affected by 50 n
M heparin; however, heparin does not
influence the inhibitory effect of thrombin. Indeed, thrombin pre-
incubation inhibits FGF-2 activity similarly both in the absence and
in the presence of heparin (B) (*P < 0.05). Data are expressed as a
percentage of cell number versus FGF-2 (100% corresponds to
1.9 · 10
5
cells) and are the mean ± SEM of five independent
experiments carried out in duplicate.
P. Totta et al. FGF-2 is a thrombin substrate
FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS 3281
function of time of exposure to thrombin. Figure 8B
shows that increasing times of FGF-2 ⁄ thrombin incu-
bation strongly reduced both the integrity of FGF-2
(expressed as a percentage of peak area on the chro-
matogram) as well as its mitogenic action (expressed as

a percentage of proliferation). After 60 min of incuba-
tion, both the integrity and function of FGF-2 were
markedly reduced to an approximately residual 20%
of that at t
0
, with an almost complete loss of signal
after 120 min of incubation. The strong loss of func-
tion and loss of integrity observed after 60 min of
incubation agree with the almost complete inhibition
of mitogenic activity of FGF (Fig. 1B) carried out
under similar experimental conditions (1 h of incuba-
tion of FGF-2 and thrombin).
FGF-2 binding to cells was then investigated in a
cytofluorimetric assay. Figure 8C shows that FGF-2
binding to cells (detected via a primary antibody
recognizing FGF-2 and a secondary fluorescent anti-
body) is lowered to control levels when FGF-2 is pre-
treated with thrombin for 1 h. These data indicate that
FGF-2 pre-treated with thrombin diminishes binding
to cell membranes, explaining, at least in part, the
observed impairment of biological activity.
Cleavage site determination
Additional investigations were then carried out to iden-
tify the FGF-2 site cleaved by thrombin. The fragment
eluting at 13.6 min (Fig. 8A) was sequenced and
revealed a N-terminal sequence of I-H-P-D-G-R-V-D,
corresponding to the fragment Ile43 to Asp50 of
mature FGF-2. As expected, the N-terminal sequence
of undigested FGF-2 was sequenced as M-A-A-G-S-I-
T, corresponding to the reported N-terminal sequence

of mature human FGF-2 (accession number P09038).
These experiments show that thrombin cleaves FGF-2
at the peptide bond between Arg42 and Ile43, releasing
the N-terminal segment from Met1 to Arg42 and the
remaining C-terminal fragment from the intact mole-
cule. We then investigated whether FGF-2 shares
sequence homology with known thrombin-recognized
cleavage sites. The sequence alignment reported in
Table 1 shows structural similarities of several known
thrombin cleavage sites to that of the site found on
FGF-2 in the present study. Beside the invariant argi-
nine (R) residue at the P1 position, common residues
are present at any position, specifically the aromatic
residues phenylalanine (F), tyrosine (Y) and tryptophan
VitronectinAB
CD
0
20
40
60
80
100
Cells/field (%)
FGF-2
Thrombin
Collagen IV
0
20
40
60

80
100
Gelatin
0
20
40
60
80
100
Fibronectin
0
20
40
60
80
100
Cell/field (%)
FGF-2
Thrombin –
–
+
+
+
+–
––
–
+
+
+
+–

–
–
–
+
+
+
+–
–
–
–
+
+
+
+–
–
Fig. 6. FGF-2-induced SK-MEL-110 inva-
sion ⁄ migration in modified Boyden cham-
bers. The invasion ⁄ migration properties of
FGF-2 is reduced in the presence of throm-
bin on different protein matrices, namely
vitronectin (A), collagen IV (B), fibronectin
(C) and gelatin (D). Data are expressed as a
percentage of cell number ⁄ field versus cells
exposed to FGF-2, (100% corresponds to 49
cells per field for vitronectin, 22 cells per
field for collagen IV, 30 cells per field for
fibronectin and 41 cells per field for gelatin).
The data reported are the mean ± SEM of
three independent experiments carried out
in duplicate.

FGF-2 is a thrombin substrate P. Totta et al.
3282 FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS
(W) at P3; the leucine (L) and isoleucine (I) residues at
P2; and proline (P) residues at the P3¢ position.
Discussion
FGF-2 is a multi-function factor with a key role in cell
proliferation and tissue differentiation. It is mainly
bound to low-affinity receptors (heparan sulfates) and
stored on the membranes and the extracellular matrix;
low levels of FGF-2 are also found circulating in the
blood [39]. Thrombin cleaves the coagulation cascade
substrates and binds and cleaves PARs receptors
[25,29], as well as other membrane-bound substrates
such as platelet glycoprotein V [30]. It circulates either
in inactive form (prothrombin) or at low concentration
in active form, subsequent to vascular damage and
activation of the coagulation cascade. In particular,
within 45 s to 5 min after venipuncture, active
thrombin appears ($3ngÆmL
)1
, 0.08 nm) and, after
2–10 min, further thrombin generation occurs, result-
ing in clotting after 15–27 min at a thrombin concen-
tration of 40–50 ngÆmL
)1
(1–2.4 nm) [40]. In the
present study, we hypothesized that FGF-2 may be
cleaved by thrombin when either thrombin or FGF-2
expression levels strongly increase. This hypothesis
stems from the observation that thrombin levels

increase and coagulation is activated in several physio-
pathological conditions [41,42]. During cancer, levels
of thrombin, FGF-2 and several other factors increase
[43,44]; therefore, we hypothesized that, under such
conditions, low-affinity substrates, which usually are
only marginally hydrolyzed, may be recognized and
digested as a part of a feedback control mechanism.
FGF-2 is present in four isoforms: three HMW
forms (22, 22.5 and 24 kDa), predominantly localized
in the nucleus, and one LMW form (18 kDa), mainly
present in the cytoplasm and on the membranes. A
recent study showed that thrombin is able to cleave
HMW FGF-2, generating a fragment somewhat simi-
lar to the LMW FGF-2. Indeed, similar to LMW, the
cleaved form stimulates endothelial cell migration and
proliferation [13]. By contrast, in the present study, we
assessed whether the real LMW FGF-2 is recognized
by thrombin; indeed, the real LMW FGF-2 is the cir-
culating form and therefore the form naturally exposed
to thrombin [39]. Thrombin at a dose of 0.1 nm (i.e.
the level reached in vivo after vascular damage [45])
strongly reduced the biological functions of FGF-2
and such inhibitory effects were abolished by blocking
the enzymatic activity of thrombin with its selective
inhibitor PPACK, indicating that the enzymatic action
of trhombin was involved. The presence of heparin did
not change the observed effects. Thrombin also
decreased binding of FGF-2 to cell membranes, as well
as FGF-2-dependent ERK2 phosphorylation, one of
the main signal-pathways mediating FGF-2 activity.

Finally, HPLC analysis indicated a thrombin-depen-
dent digestion, with kinetics matching the functional
impairment of FGF-2. The inhibitory effects of throm-
bin on FGF-2, as observed in the present study on the
human metastatic melanoma cell line SK-MEL-110,
were also confirmed in additional proliferation assays
carried out using human endothelial cells (HUVEC).
The k
cat
⁄ K
m
value of FGF-2 hydrolysis revealed a
relatively low specificity for thrombin; however, the
high concentrations of active thrombin attained in vivo
during the activation of coagulation [45] may be suffi-
cient to sustain significant FGF-2 cleavage. Although
conclusive in vivo evidence is lacking, these data sug-
gest that thrombin-dependent FGF-2 cleavage may
occur under physiological ⁄ pathological conditions with
enhanced thrombin generation, such as in athero-
thrombosis, or with elevated levels of FGF-2 and
thrombin, such as in cancer. The cleavage site identi-
fied between Arg42 and Ile43 of the mature FGF-2
1.5
p-ERK1/2
A
B
Total ERK1/2
FGF-2
Thrombin –

+
+
–
+
+
–
–
0
0.5
1.0
p-ERK2 / total ERK2
Fig. 7. FGF-2-induced SK-MEL-110 ERK ⁄ MAPK phosphorylation in
the presence of active thrombin. SK-MEL-110 (4 · 10
4
) cells were
seeded and grown for 24 h in complete medium. Medium was
then replaced and cells were starved overnight with incomplete
medium. These cells were stimulated for 10 min with FGF-2
(10 ngÆmL
)1
), thrombin (0.1 nM) or FGF-2 with thrombin pre-incu-
bated for 1 h at 37 °C and then added PPACK (50 n
M) to block
thrombin enzymatic activities. Next, cells were lysated and
Western blotting analysis was performed to evaluate ERK1 ⁄ 2
phosporylation. (A) Active thrombin in the presence of FGF-2 (lane
FGF-2 + thrombin) significantly reduces ERK-2 (42 kDa) phosphory-
lation (one representative experiment). (B) Densitometry analysis of
three separate experiments.
P. Totta et al. FGF-2 is a thrombin substrate

FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS 3283
sequence is close to the Asp57 residue, initiating the
FREG fragment of FGF-2, which was recently demon-
strated to strongly modulate FGF-2 activity [19]. We
hypothesize that proteolytic degradation as well as the
release of active fragments may represent, at least in
part, a feedback regulatory mechanism for modulating
the angiogenic properties of FGF-2 when FGF-2 levels
increase.
The reported proteolytic effect of thrombin on
human matrix proteins, such as vitronectin, fibronectin
Fig. 8. Thrombin-induced FGF-2 structural and functional impair-
ment. (A) FGF-2 degradation investigated by HPLC. The lower chro-
matogram reports the elution profile of FGF-2 alone; the upper
chromatogram reports the elution profile of FGF-2 incubated with
thrombin for 30 min. In the presence of thrombin, the peak eluting
at 23.5 min is reduced and an additional peak appears after
13.6 min of elution, corresponding to the degradation fragment. (B)
The structural impairment parallels the functional impairment of
FGF-2 in the presence of different time points of thrombin incuba-
tion. Structural impairment is expressed as time-dependent
decrease of the peak eluting at 23.5 min, whereas functional
impairment is expressed as the time-dependent loss of a mitogenic
effect on SK-MEL-110 cells. Structural and functional impairment
are approximately 50% at 30 min, approximately 80% at 60 min
and almost complete after 120 min of incubation. (C) FGF-2 binding
to SK-Mel-110 cells. SK-MEL-110 (2 · 10
5
) cells were seeded and
grown for 24 h in complete medium. Medium was then replaced

and cells were starved overnight with incomplete medium. These
cells were exposed to FGF-2 (50 ngÆmL
)1
) or to thrombin (0.1 nM)
or to FGF-2 previously incubated with thrombin for 1 h at 37 °C
(50 ngÆmL
)1
and 0.1 nM respectively) FGF-2 binding to cells,
detected with a primary antibody and a secondary fluorescent anti-
body, was then measured by flow cytometry. The peak on the
right, corresponding to the peak at higher fluorescence, is
increased and enlarged toward the right-hand side (i.e. higher fluo-
rescence) in cells exposed to FGF-alone compared to control cells,
whereas it is lowered to the control level in cells exposed to FGF-2
pre-incubated with thrombin. One representative graph from three
experiments is reported.
80
100
80
100
FGF-2
peak area (%)
60
40
60
40
Cell number (%)
Time (min)
Absorbance 214 nm
5 1015202530

Time (min)
30
20
0
20
0
60 90 120
FGF-2
FGF-2 + Thrombin
No FGF-2
Counts
FL1 log
10
0
10
1
10
2
0
A
B
C
Table 1. P3–P3¢ sequence of the most common thrombin substrates. The structural similarities of FGF-2 to other known thrombin sub-
strates are indicated in bold.
Substrate
Position
Residue no.
Swiss-Prot
database entryP3 P2 P1 P1¢ P2¢ P3¢
FGF-2 FLRI HP 40–45 P09038

Fibrinogen S A R G H R 12–17 P02675
PAR-4 A P R GYP 45–50 Q96RI0
Factor V (1) G IRS F R 737–742 Q15430
Factor V (2) YL RS N N 1571–1576 Q15430
Factor VIII Q IRS V A 389–394 P00451
Platelet glycoproteinV G P R GPP 474–479 P40197
Carboxypeptidase B2 (TAFI) S P RA S A 112–117 Q96IY4
Vitronectin W G R T S A 302–307 P04004
FGF-2 is a thrombin substrate P. Totta et al.
3284 FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS
and murine collagen IV [46,47], might indicate dual
effects on cell migration ⁄ invasion. Indeed, on the one
hand, thrombin digests FGF-2, reducing its angiogenic
and invasive action; on the other hand, thrombin
degrades matrix proteins, facilitating cell invasion. The
net effect measured in vitro in the present study was a
strong reduction of FGF-2-induced cell migra-
tion ⁄ invasion. Preliminary results indicate that, at simi-
lar concentrations, thrombin recognizes and degrades
at least one other angiogenic growth factor, namely
platelet-derived growth factor-BB, suggesting that
increased thrombin concentrations may modulate
angiogenesis and cell chemotaxis ⁄ invasion at different
levels. A number of studies indicate that coagulation
proteases play significant roles in cancer biology [28].
For example, melanoma is a highly metastatic cancer,
and evidence exists indicating that thrombin contrib-
utes to this aggressive pattern. Furthermore, previous
studies show that the assembly and regulation of the
prothrombinase complex on the murine melanoma cell

line B16F10 is accelerated with enhanced thrombin
formation [48]. These conditions, along with PAR-1
expression in melanoma cell lines [29], likely play a rel-
evant role in the metastatic potential of these cancer
cells. The additional effects of thrombin on growth
factors and matrix proteins reported in the present
study may further contribute to these effects, indicat-
ing that pathological conditions with increased
thrombin levels may lead to proteolytic matura-
tion ⁄ degradation of both growth factors and extracel-
lular matrix proteins, affecting cell mitogenic ⁄ invasion
features at different levels. Cancer can activate the
coagulation cascade, as also demonstrated by the
enhanced rate of thromboembolic complications and
the beneficial effect of anticoagulant therapies in the
prevention of these disorders in cancer patients [49].
However, it is less well known whether activation of
the coagulation system may also support or inhibit
tumor progression. The findings reported in the pres-
ent study outline the functional link between thrombin
activity and the mitogenic ⁄ invasive properties of mela-
noma cells. This observation may contribute to
explaining why paradoxical pro-apoptotic effects of a
high thrombin concentration on different cell types
recently were reported [49].
Two reports available in the literature indicate that
HMW FGF-2 and FGF-1 are cleaved by thrombin,
whereas they also report that LMW FGF-2 is not
cleaved by thrombin [13,50]. In this respect, it should
be highlighted that the experimental conditions (i.e.

the amount of thrombin and FGF-2, as well as the
incubation times) are markedly different. Indeed, we
used 0.1 nm = 0.01 UÆmL
)1
thrombin, whereas both
Lobb [50] and Yu et al. [13] used at least 100-fold
more thrombin. The source of thrombin was also dif-
ferent; both Lobb [50] and Yu et al. [13] used commer-
cial human thrombin, whereas we used in-house
highly-purified human thrombin [51]. Lobb [50] used
in-house purified bovine brain-derived FGF-2, whereas
both Yu et al. [13] and ourselves used commercial
human recombinant FGF-2. The incubation times
were also markedly different: Lobb [50] reports 6–20 h
of incubation of brain-derived bovine FGF-2 with
thrombin, whereas Yu et al. [13] used human recombi-
nant FGF-2 at unspecified doses incubated with
thrombin for 5–15 min. In the present study, specified
doses of human recombinant FGF-2 were incubated
for significantly longer times (60–120 min). The experi-
mental conditions employed in the present study were
designed to mimic in vitro, as much as possible, the
conditions occurring under pathological states that
show increased thrombin levels and increased FGF-2
levels, and were confirmed, in different cell models and
in the presence of heparin, to support the possible
in vivo relevance of the observed effects.
In conclusion, in the present study, we show that
thrombin is able to cleave human LMW FGF-2 in vitro,
further unravelling the complex linkage between coagu-

lation activation and cancer progression.
Experimental procedures
Cell culture and proliferation
Human metastatic melanoma cell line SK-MEL-110 [52]
(4 · 10
4
) cells were seeded in six-well plates and grown for
24 h at 37 °C, 5% CO
2
in DMEM supplemented with 1%
penicillin–streptomycin, 1% l-glutamine and 10% charcoal
stripped fetal bovine serum (Gibco, Carlsbad, CA, USA).
Medium was then replaced and cells were starved overnight
with DMEM supplemented with 1% penicillin–streptomy-
cin and 1% l-glutamine. Subsequently, dose–response and
time-course proliferation assays were performed using dif-
ferent stimuli and different time points. For dose–response
and for time-course assays (Fig. 1), human recombinant
FGF-2 (10 ngÆmL
)1
, 0.6 nm) (Pierce Endogen, Rockford,
IL, USA) and human a-thrombin (thrombin) [51] (0.001,
0.01, 0.1 and 1 nm or 0.1 nm) were pre-incubated for 1 h at
37 °C; enzymatic activity was then blocked with PPACK
(50 nm) (Calbiochem, BIOMOL International LP, Exeter,
UK) and cells were stimulated with these mixtures for 48
or 24 h and 48 and 72 h, respectively.
In other experiments (Figs 2–4), SK-MEL-110 cells were
directly exposed for 48 h to FGF-2 (10 ngÆmL
)1

, 0.6 nm)in
the presence or the absence of 5.7 lm TRAP-1 or TRAP-4
[53] (PRIMM Srl, Milan, Italy), or different mixtures of
P. Totta et al. FGF-2 is a thrombin substrate
FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS 3285
PPACK (50 nm) or heparin 25 000 IU per 5 mL (10 and
50 ngÆmL
)1
under the assumption that 170 IU corresponds
to 1 mg) (Epsoclar from Mayne Pharma Srl, Napoli, Italia)
[54] and ⁄ or thrombin (0.1 nm, 0.01 UÆmL
)1
). Finally, in the
experiments depicted in Fig. 8B, FGF-2 (10 ngÆmL
)1
,
0.6 nm) was pre-incubated at different time points with
thrombin (0.1 nm, 0.01 UÆmL
)1
); the enzymatic reaction
was then blocked with PPACK (50 nm) and SK-MEL-110
proliferation was stimulated for 48 h. Stimuli were resus-
pended in all cases in DMEM with 0.1% BSA (Sigma-
Aldrich, St Louis, MO, USA).
Primary HUVEC (Cambrex, Walkersville, MD, USA)
were also used. Cells were grown at 37 °C, 5% CO
2
. HUVEC
(8 · 10
4

) were seeded in six-well plates and grown for 24 h in
endothelial cell basal medium 2 (EBM-2; Cambrex) supple-
mented with endothelial growth medium 2 (Cambrex), 1%
penicillin–streptomycin and 1% l-glutamine. Medium was
then replaced and cells were starved overnight with EBM-2
supplemented with 1% penicillin–streptomycin and 1%
l-glutamine. Subsequently, HUVECs were stimulated with
10 ngÆmL
)1
FGF-2 resuspended in EBM-2 with 0.1% BSA in
the presence or the absence of thrombin 0.1 nm, for 24 h.
Time-course experiments were also carried out.
After treatment, cells were harvested with trypsin
(Gibco), stained with trypan blue solution (Sigma-Aldrich)
and counted in a hemocytometer (improved Neubauer
chamber) in quadruplicate.
Cell invasion assay
The SK-MEL-110 invasion assay was carried out in mod-
ified Boyden chambers as previously described [55].
Briefly, 8 lm pore-size polycarbonate filters (Costar, Cam-
bridge, MA, USA) were coated with human plasma vitro-
nectin (Calbiochem), murine collagen type IV (Becton
Dickinson, Bedford, MA, USA), human plasma fibronec-
tin (Invitrogen, Carlsbad, CA, USA) or type A gelatin
from porcine skin (10 lgÆmL
)1
) (Sigma-Aldrich).
SK-MEL-110 cells were grown in DMEM supplemented
with 1% penicillin–streptomycin, 1% l-glutamine and
10% fetal bovine serum and then replaced with DMEM

supplemented with 1% penicillin–streptomycin, 1% l-glu-
tamine overnight. Cells were then harvested by trypsiniza-
tion, resuspended in DMEM supplemented with 1%
penicillin–streptomycin, 1% l-glutamine and 0.1% BSA,
and 800 lL were added to the upper portion of the Boy-
den chambers containing 1 · 10
6
cellsÆmL
)1
; the lower
portion of the chambers contained either DMEM supple-
mented with 0.1% BSA or 10 ngÆmL
)1
(0.6 nm) FGF-2
with 0.1% BSA, alone or mixed with 0.1 nm thrombin.
Invasion was allowed to proceed for 4 h at 37 °Cin5%
CO
2
; then cells were fixed in 95% ethanol and stained
with 0.5% toluidine blue in NaCl ⁄ Pi (Gibco), pH 7.4, for
10 min. Migrated cells were counted by evaluation of 15
fields at · 400 magnification.
Phosphorylation and degradation analyses by
Western blotting
ERK1 ⁄ 2 phosphorylation analysis (Fig. 6): SK-MEL-110
cells were grown in DMEM supplemented with 1% penicil-
lin–streptomycin, 1% l-glutamine and 10% fetal bovine
serum, replaced with DMEM supplemented with 1% peni-
cillin–streptomycin, 1% l-glutamine overnight. FGF-2
(10 ngÆmL

)1
, 0.6 nm), human a-thrombin (thrombin)
(0.1 nm, 0.01 UÆmL
)1
) and FGF-2 with human a-thrombin
(thrombin) (0.1 nm, 0.01 UÆmL
)1
) were pre-incubated for
1 h at 37 °C and then PPACK (50 nm) was added to block,
where necessary, thrombin enzymatic activities. Cells were
treated with these mixtures for 10 min, rinsed with ice-cold
NaCl ⁄ Pi and lysed for 15 min with 1% triton, 10% glyc-
erol, 100 mm NaCl, 5 mm EDTA, 20 mm Hepes (pH 7.4),
10 mm NaF, 2 mm phenylmethanesulfonyl fluoride, 10 lm
NaVO3 and 1% protease inhibitor cocktail (Sigma-
Aldrich). Lysates were then boiled for 7 min. After determi-
nation of protein concentration (Bio-Rad Laboratories,
Hercules, CA, USA), 30 lg of total proteins were resolved
in SDS ⁄ PAGE with NuPAGE
Ò
pre-cast gels for protein
electrophoresis (10%) and NuPAGE
Ò
SDS running buffer
(Mops buffer) (Invitrogen), transferred to nitrocellulose
membrane and blocked with 5% milk in T-NaCl ⁄ Pi (0.1%
Tween 20 in NaCl ⁄ Pi, pH 7.4). After three washes with
T-NaCl ⁄ Pi, membranes were incubated with monoclonal
anti-pERK or anti-ERK sera (Cell Signaling Technology,
Beverly, MA, USA) for 1 h. For detection, secondary anti-

mouse serum (1 : 5000) (Pierce Endogen) was used followed
by chemiluminescence (ECL; Amersham, Little Chalfont,
UK) and autoradiography.
FGF-2 degradation analysis (see, Fig. S1)
FGF-2 was dissolved in a 50 mm Tris–HCl containing
150 mm NaCl, pH 8, and incubated for 1 h alone, with
PPACK, with thrombin, or with inactivated-thrombin
(thrombin pre-incubated with PPACK, PPACK-thrombin)
(1 : 1 molar ratio, 100 lgÆmL
)1
)at37°C. After incubation,
active-thrombin was blocked by PPACK (1 : 100 molar
ratio). FGF-2 (500 ng) was then resolved in SDS ⁄ PAGE
with NuPAGE
Ò
pre-cast gels for protein electrophoresis
(12%) and NuPAGE
Ò
SDS running buffer (Mes buffer;
Invitrogen), transferred to nitrocellulose membrane and
blocked with 5% milk in T-NaCl ⁄ Pi. After three washes
with T-NaCl ⁄ Pi, membranes were incubated with poly-
clonal anti-FGF-2 (0.3 lgÆmL
)1
) serum (Oncogene
Research Products, Darmstad, Germany) for 1 h. For
detection, secondary anti-goat serum (1 : 5000) (Pierce
Endogen) was used followed by chemiluminescence (ECL;
Amersham) and autoradiography.
Bands were quantified using a calibrated imaging densi-

tometer (GS 710; Bio-Rad Laboratories) and analyzed
using quantity one software (Bio-Rad Laboratories).
FGF-2 is a thrombin substrate P. Totta et al.
3286 FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS
Thrombin-induced digestion
Thrombin-induced FGF-2 digestion monitored by
HPLC analysis
FGF-2 hydrolysis was performed in 10 mm Tris–HCl,
150 mm NaCl, pH 7.5, at 37 °C containing 75 nm thrombin
and FGF-2 at 10 lgÆmL
)1
(0.6 lm). The proteolytic reac-
tion was stopped at various times with PPACK (10 lm)
and the reaction products were analyzed by HPLC on a
250 · 4.6 mm reverse-phase 304 column (Bio-Rad Labora-
tories). The chromatographic run was performed with the
conditions: 5–60% acetonitrile in 0.1% trifluoroacetic acid
for 40 min at a flow rate of 1 mLÆmin
)1
. The peaks were
detected routinely at 214 nm. Under these conditions, the
observed velocity of the reaction did not reach saturation,
such that pseudo-first order conditions were met (concen-
tration < K
m
of the reaction). Hence, the FGF-2 peak area
remaining after cleavage by thrombin was fitted to the
equation:
Pt (%) ¼ 100  exp ðÀk
obs

tÞð1Þ
where k
obs
is the pseudo-first order rate of its hydrolysis,
equal to e
0
k
cat
⁄ K
m
(where e
0
is the concentration of
thrombin).
Sequencing of FGF-2 fragment
Fractions pertaining to the hydrolysis peaks were pooled
and dried under vacuum. Their molecular identity was
checked by N-terminal sequence performed by the Edman
reaction on an automatic analyzer (Applied Biosystems,
Foster City, CA, USA) according to standard procedures.
FGF-2 binding
Binding of FGF-2 alone or FGF-2 incubated with thrombin
was carried out on live cells and detected by flow cytometry.
SK-MEL-110 cells were exposed to FGF-2 alone
(50 ngÆmL
)1
) or thrombin alone (0.1 nm) or FGF-2 previ-
ously incubated with thrombin for 1 h at 37 °C (50 ngÆmL
)1
and 0.1 nm, respectively). Binding was allowed to proceed

for 20 min at 37 °C. Then, cells were washed three times,
detached, and washed with NaCl ⁄ Pi–1% BSA. Primary goat
anti-(human FGF-2) serum (SC-79G; Santa Cruz Biotech-
nology, Santa Cruz, CA, USA) was added for 60 min at
room temperature, and then secondary donkey anti-(goat
Alexafluor 488) (Invitrogen, Molecular Probes) serum was
added for 30 min at room temperature. After washings, cells
were analyzed at a Beckman Coulter CYAN ADP 9-colours
flow cytometer (Beckman Coulter, Fullerton, CA, USA).
Cells were gated using forward versus side scatter to exclude
dead cells and debris. Fluorescence of 10 000 cells per sample
was acquired in logarithmic mode to quantify the binding
of the relevant molecules. summit 4.3 software (Beckman
Coulter) was used for data elaboration.
Statistical analysis
All experiments were performed at least three times in
duplicate. A Student’s t-test was carried out to evaluate
statistical significance at P < 0.05.
Acknowledgements
The present study was supported by the Italian Min-
istry of University and Research (‘PRIN-2005’) to
R.D.C.; Progetto Oncoproteomica Italia-USA (no.
527B ⁄ 2A ⁄ 5) to A.F.; and Ministry of Health
MS-RO2006, Prog. Ord. to A.F. We thank the Docto-
rate School of Scienze e Tecnologie Cellulari at
Sapienza University in Rome for the support.
References
1 Dailey L, Ambrosetti D, Mansukhani A & Basilico C
(2005) Mechanisms underlying differential responses to
FGF signaling. Cytokine Growth Factor Rev 16, 233–

247.
2 Martin C, Bueno D, Alonso MI, Moro JA, Callejo S,
Parada C, Martin P, Carnicero E & Gato A (2006)
FGF2 plays a key role in embryonic cerebrospinal fluid
trophic properties over chick embryo neuroepithelial
stem cells. Dev Biol 297, 402–416.
3 Marie PJ, Coffin JD & Hurley MM (2005) FGF and
FGFR signaling in chondrodysplasias and craniosynos-
tosis. J Cell Biochem 96, 888–896.
4 Dvorak P, Dvorakova D & Hampl A (2006) Fibroblast
growth factor signaling in embryonic and cancer stem
cells. FEBS Lett 580, 2869–2874.
5 Kankuri E, Cholujova D, Comajova M, Vaheri A
& Bizik J (2005) Induction of hepatocyte growth
factor ⁄ scatter factor by fibroblast clustering directly
promotes tumor cell invasiveness. Cancer Res 65,
9914–9922.
6 Garcia-Maya M, Anderson AA, Kendal CE, Kenny
AV, Edwards-Ingram LC, Holladay A & Saffell JL
(2006) Ligand concentration is a driver of divergent sig-
naling and pleiotropic cellular responses to FGF. J Cell
Physiol 206, 386–393.
7 De Marchis F, Ribatti D, Giampietri C, Lentini A,
Faraone D, Scoccianti M, Caporossi MC & Facchiano
A (2002) A platelet-derived growth factor inhibits basic
fibroblast growth factor angiogenic properties in vitro
and in vivo through its alpha receptor. Blood 99, 2045–
2053.
8 Mathew S, Arandjelovic S, Beyer WF, Gonias SL &
Pizzo SV (2003) Characterization of the interaction

between alpha2-macroglobulin and fibroblast growth
factor-2: the role of hydrophobic interactions. Biochem
J 374, 123–129.
P. Totta et al. FGF-2 is a thrombin substrate
FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS 3287
9 Abuharbeid S, Czubayko F & Aigner A (2006) The
fibroblast growth factor-binding protein FGF-BP. Int J
Biochem Cell Biol 38, 1463–1468.
10 Faraone D, Aguzzi MS, Ragone G, Russo K, Capog-
rossi MC & Facchiano A (2006) Heterodimerization of
FGF-receptor 1 and PDGF-receptor-alpha: a novel
mechanism underlying the inhibitory effect of PDGF-
BB on FGF-2 in human cells. Blood 107, 1896–1902.
11 Sanchez-Heras E, Howell FV, Williams G & Doherty P
(2006) The FGF receptor acid box is essential for inter-
actions with N-cadherin and all of the major isoforms
of NCAM. J Biol Chem 281, 35208–35216.
12 Goodman OB Jr, Febbraio M, Simantov R, Zheng R,
Shen R, Silverstein RL & Nanus DM (2006) Neprilysin
inhibits angiogenesis via proteolysis of fibroblast growth
factor-2. J Biol Chem 281 , 33597–33605.
13 Yu PJ, Ferrari G, Pirelli L, Galloway AC, Mignatti P
& Pintucci G (2008) Thrombin cleaves the high molecu-
lar weight forms of basic fibroblast growth factor
(FGF-2): a novel mechanism for the control of FGF-2
and thrombin activity. Oncogene 27, 2594–2601.
14 Stallmach R & Gloor SM (2008) Neurobin ⁄
TMPRSS11c, a novel type II transmembrane serine
protease that cleaves fibroblast growth factor-2 in vitro.
Biochem J 412 , 81–91.

15 Meddahi A, Lemdjabar H, Caruelle JP, Barritault D &
Hornebeck W (1995) Inhibition by dextran derivatives
of FGF-2 plasmin-mediated degradation. Biochimie 77,
703–706.
16 Sahni A, Khorana AA, Baggs RB, Peng H & Francis
CW (2006) FGF-2 binding to fibrin (ogen) is required
for augmented angiogenesis. Blood 107, 126–131.
17 Sahni A, Baker CA, Sporn LA & Francis CW (2000)
Fibrinogen and fibrin protect fibroblast growth factor-2
from proteolytic degradation. Thromb Haemost 83,
736–741.
18 Ding L, Donate F, Parry GC, Guan X, Maher P &
Levin EG (2002) Inhibition of cell migration and
angiogenesis by the amino-terminal fragment of 24kD
basic fibroblast growth factor. J Biol Chem 277,
31056–31061.
19 Facchiano A, Russo K, Facchiano AM, De Marchis F,
Facchiano F, Ribatti D, Aguzzi MS & Caporossi MC
(2003) Identification of a novel domain of fibroblast
growth factor 2 controlling its angiogenic properties.
J Biol Chem 278, 8751–8760.
20 Mann KG, Butenas S & Brummel K (2003) The
dynamics of thrombin formation. Arterioscler Thromb
Vasc Biol 23, 17–25.
21 Hron G, Kollars M, Binder BR, Eichinger S & Kyrle
PA (2006) Identification of patients at low risk for
recurrent venous thromboembolism by measuring
thrombin generation. JAMA 296, 397–402.
22 Stegenga ME, van der Crabben SN, Levi M, de Vos
AF, Tanck MW, Sauerwein HP & van der Poll T

(2006) Hyperglycemia stimulates coagulation, whereas
hyperinsulinemia impairs fibrinolysis in healthy humans.
Diabetes 55, 1807–1812.
23 Nierodzik ML & Karpatkin S (2006) Thrombin induces
tumor growth, metastasis, and angiogenesis: evidence
for a thrombin-regulated dormant tumor phenotype.
Cancer Cell 10, 355–362.
24 Davie EW, Fujikawa K & Kisiel W (1991) The coagula-
tion cascade: initiation, maintenance, and regulation.
Biochemistry 30, 10363–10370.
25 Kahn ML, Nakanishi-Matsui M, Shapiro MJ, Ishihara
H & Coughlin SR (1999) Protease-activated receptors 1
and 4 mediate activation of human platelets by throm-
bin. J Clin Invest 103, 879–887.
26 Huntington JA (2005) Molecular recognition mecha-
nisms of thrombin. J Thromb Haemost 3
, 1861–1872.
27 Wang L, Luo J & He S (2007) Induction of MMP-9
release from human dermal fibroblasts by thrombin:
involvement of JAK ⁄ STAT3 signaling pathway in
MMP-9 release. BMC Cell Bio 8, 14.
28 McLaughlin JN, Patterson MM & Malik AB (2007)
Protease-activated receptor-3 (PAR3) regulates PAR1
signaling by receptor dimerization. Proc Natl Acad Sci
USA 104, 5662–5667.
29 Tellez C & Bar-Eli M (2003) Role and regulation of the
thrombin receptor (PAR-1) in human melanoma. Onco-
gene 22, 3130–3137.
30 Ramakrishnan V, DeGuzman F, Bao M, Hall SW,
Leung LL & Phillips DR (2001) A thrombin receptor

function for platelet glycoprotein Ib-IX unmasked by
cleavage of glycoprotein V. Proc Natl Acad Sci USA
98, 1823–1828.
31 Arai T, Guo JP & McGeer PL (2005) Proteolysis of
non-phosphorylated and phosphorylated tau by
thrombin. J Biol Cell 280, 5145–5153.
32 Crawley JT, Lam JK, Rance JB, Mollica LR,
O’Donnell JS & Lane DA (2005) Proteolytic
inactivation of ADAMTS13 by thrombin and plasmin.
Blood 105, 1085–1093.
33 Smalley KS (2003) A pivotal role for ERK in the
oncogenic behaviour of malignant melanoma? Int J
Cancer 104, 527–532.
34 Reiland J, Kempf D, Roy M, Denkins Y & Marchetti
D (2006) FGF2 binding, signaling, and angiogenesis are
modulated by heparanase in metastatic melanoma cells.
Neoplasia 8, 596–606.
35 Ma C, Bower KA, Chen G, Shi X, Ke ZJ & Luo J
(2008) Interaction between ERK and GSK 3beta
mediates basic FGF-induced apoptosis in SK-N-MC
neuroblastoma cells. J Biol Chem 283, 9248–
9256.
36 Gentilini D, Busacca M, Di Francesco S, Vignali M,
Vigano
`
P & Di Blasio AM (2007) PI3K ⁄ Akt and
ERK1 ⁄ 2 signalling pathways are involved in
endometrial cell migration induced by 17beta-estradiol
FGF-2 is a thrombin substrate P. Totta et al.
3288 FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS

and growth factors 1. Mol Hum Reprod 13, 317–
322.
37 De Cristofaro R (1996) Effect of thrombomodulin on
the molecular recognition and early catalytic events in
thrombin-protein C interaction. Thromb Haemost 76,
556–560.
38 Schneider M, Nagashima M, Knappe S, Zhao L, Morser
J & Nesheim M (2002) Amino acid residues in the P6-P’3
region of thrombin-activable fibrinolysis inhibitor
(TAFI) do not determine the thrombomodulin depen-
dence of TAFI activation. J Biol Chem 277, 9944–9951.
39 Kaminska J, Kowalska M, Kotowicz B, Fuksiewicz M,
Glogowski M, Wojcik E, Chechlinska M & Steffen J.
(2006) Pretreatment serum levels of cytokines and
cytokine receptors in patients with non-small cell lung
cancer, and correlations with clinicopathological
features and prognosis. M-CSF-an independent
prognostic factor. Oncology 70, 115–125.
40 Shuman MA & Majerus PW (1976) The measurement
of thrombin in clotting blood by radioimmunoassay.
J Clin Invest 58, 1249–1258.
41 De Cristofaro R, Rocca B, Vitacolonna E, Falco A,
Marchesani P, Ciabattoni G, Landolfi R, Patrono C &
Davi G (2003) Lipid and protein oxidation contribute
to a prothrombotic state in patients with type 2 diabetes
mellitus. J Thromb Haemost 1, 250–256.
42 Esmon CT (2006) Inflammation and the activated pro-
tein C anticoagulant pathway. Semin Thromb Hemost
32, 49–60.
43 Schiller H, Bartscht T, Arlt A, Zahn MO, Seifert A,

Bruhn T, Bruhn HD & Gieseler F (2002) Thrombin as
a survival factor for cancer cells: thrombin activation in
malignant effusions in vivo and inhibition of idarubicin-
induced cell death in vitro. Int J Clin Pharmacol Ther
40, 329–335.
44 Gieseler F, Lu
¨
hr I, Kunze T, Mundhenke C, Maass N,
Erhart T, Denker M, Beckmann D, Tiemann M, Schulte
C et al. (2007) Activated coagulation factors in human
malignant effusions and their contribution to cancer cell
metastasis and therapy. Thromb Haemost 97, 1023–1030.
45 Borrelli V, Sterpetti AV, Coluccia P, Randone B,
Cavallaro A, Santoro D’Angelo L & Cucina A (2001)
Bimodal concentration-dependent effect of thrombin on
endothelial cell proliferation and growth factor release
in culture. J Surg Res 100, 154–160.
46 Gechtman Z, Belleli A, Lechpammer S & Shaltiel S
(1997) The cluster of basic amino acids in vitronectin
contributes to its binding of plasminogen activator
inhibitor-1: evidence from thrombin-, elastase- and
plasmin-cleaved vitronectins and anti-peptide antibod-
ies. Biochem J 325, 339–349.
47 Rossignol P, Bouton MC, Jandrot-Perrus M, Bryckaert
M, Jacob MP, Bezeaud A, Guillin MC, Michel JB &
Meilhac O (2004) A paradoxical pro-apoptotic effect of
thrombin on smooth muscle cells. Exp Cell Res 299,
279–285.
48 Kirsberg C, Rumjanek VM & Monteiro RQ (2005)
Assembly and regulation of prothrombinase complex

on B16F10 melanoma cells. Thromb Res 115,
123–129.
49 Furie B & Furie BC (2006) Cancer-associated
thrombosis. Blood Cells Mol Dis 36, 177–181.
50 Lobb RR (1988) Thrombin inactivates acidic fibroblast
growth factor but not basic fibroblast growth factor.
Biochemistry 27, 2572–2578.
51 Lancellotti S, Rutella S, De Filippis V, Pozzi N, Rocca
B & De Cristofaro R. (2008) Fibrinogen-elongated
gamma chain inhibits thrombin-induced platelet
response, hindering the interaction with different
receptors. J Biol Chem 283, 30193–30204.
52 Gorospe M, Cirielli C, Wang X, Seth P, Capogrossi
MC & Holbrook NJ (1997) p21(Waf1 ⁄ Cip1) protects
against p53-mediated apoptosis of human melanoma
cells. Oncogene 14, 929–935.
53 Duarte M, Kolev V, Soldi R, Kirov A, Graziani I,
Oliveira SM, Kacer D, Friesel R, Maciag T &
Prudovsky I (2006) Thrombin induces rapid
PAR1-mediated non-classical FGF1 release. Biochem
Biophys Res Commun 350, 604–609.
54 Furue MK, Na J, Jackson JP, Okamoto T, Jones M,
baker D, Hata R, Noore HD, Sato JD & Andrews PW
(2008) Heparin promotes the growth of human
embryonic stem cells in a defined serum-free medium.
PNAS 105, 13409–13414.
55 Aguzzi MS, Giampietri C, De Marchis F, Padula F,
Gaeta R, Ragone G, Capogrossi MC & Facchiano A
(2004) RGDS peptide induces caspase 8 and caspase 9
activation in human endothelial cells. Blood 103,

4180–4187.
Supporting information
The following supplementary material is available:
Fig. S1. FGF-2 analysis by western blotting.
This supplementary material can be found in the
online version of this article.
Please note: Wiley-Blackwell is not responsible for
the content or functionality of any supplementary
materials supplied by the authors. Any queries (other
than missing material) should be directed to the corre-
sponding author for the article.
P. Totta et al. FGF-2 is a thrombin substrate
FEBS Journal 276 (2009) 3277–3289 ª 2009 The Authors Journal compilation ª 2009 FEBS 3289