Thrombin peptide, TP508, stimulates angiogenic responses in animal

models of dermal wound healing, in chick chorioallantoic membranes,

and in cultured human aortic and microvascular endothelial cells

Andrea M. Norfleet, John S. Bergmann, Darrell H. Carney*

Chrysalis BioTechnology, Inc., 2200 Market Street, Suite 600, Galveston, TX, USA

Department of Human Biological Chemistry and Genetics, University of Texas Medical Branch, Galveston, TX, USA

Received 1 September 2001; accepted 1 October 2001

Abstract

The a-thrombin peptide, TP508, accelerates the healing of full-thickness wounds in both normal and ischemic skin. In wounds treated

with TP508, a pattern of increased vascularization is consistently observed both grossly and microscopically when compared to wounds

treated with saline. One possible mechanism by which the peptide accelerates wound healing is by promoting revascularization of granulation

tissue at the injured site. To evaluate the angiogenic potential of TP508, the peptide was tested in the chick embryo chorioallantoic membrane

(CAM), where it increased the density and size of CAM blood vessels relative to controls. Additionally, TP508 stimulated chemokinesis and

chemotaxis in a dose-dependent fashion in cultured human aortic and human microvascular endothelial cells. Taken together, these in vivo

and in vitro data support an angiogenic role for TP508 in wound healing. A working model is presented to explain how this 23-amino-acid

peptide, which lacks proteolytic activity, is generated during wound healing and contributes to the nonproteolytic functions associated with

a-thrombin during tissue repair. D 2002 Elsevier Science Inc. All rights reserved.

Keywords: Growth factors; Tissue repair; Synthetic peptides; Hemostasis

1. Cellular responses to A-thrombin: proteolytic and

nonproteolytic mechanisms

The functions of A-thrombin in response to an injury are

complex, evolving in a dynamic—and sometimes contra-

dictory—fashion as the microenvironment of the wound

changes over the course of healing (Grand et al., 1996;

Strukova, 2001). Initially, for example,A-thrombin is pivotal

in formation of the fibrin clot, aggregation of platelets, and

stimulation of proinflammatory responses. Later,A-thrombin

triggers events that limit its own formation, lead to dissolu-

tion of the clot, and contribute to cessation of inflammatory

cell activity. We are just beginning to gain an understanding

of howA-thrombin achieves this type of functional versatility

and how the diverse effects of thrombin contribute to the

intricate network of responses required for tissue repair.

One facet of A-thrombin’s multifunctionality is related to

its well-characterized activity as a serine protease. In this ca-

pacity, A-thrombin utilizes its proteolytic activity to catalyze

the hydrolysis of specific substrate molecules to yield unique

products, as exemplified by the cleavage of fibrinogen

to fibrin. In a less classical role, A-thrombin triggers intra-

cellular signaling pathways by hydrolytically removing

a portion of the amino-terminal exodomain of specific

G-protein-coupled receptors, called proteolytically activated

receptors (PARs). Three PAR subtypes (PAR-1, -3, -4) are

activated by A-thrombin via this unique mechanism, which

requires enzymatic cleavage of the PAR by the serine

protease (Vergnolle et al., 2001).

Another dimension to the complex biology ofA-thrombin

is a group of functions that are not dependent upon protease

activity. Compelling evidence for such interactions is based

on experiments in which cells respond to A-thrombin even

when the active site of the enzyme is rendered inert by

chemical blockade or genetic mutation (Bar-Shavit et al.,

1983, 1986; Carney et al., 1992a,b; Jenkins et al., 1995).

Moreover, this type of observation has led to the discovery

0306-3623/02/$ – see front matter D 2002 Elsevier Science Inc. All rights reserved.

PII: S0306 -3623 (01 )00118 -5

* Corresponding author. Chrysalis BioTechnology, 2200 Market Street,

Suite 600, Galveston, TX 77550, USA. Tel.: +1-409-750-9251; fax: +1-

409-750-9253.

E-mail address: dcarney@chrysalisbio.com (D.H. Carney).

General Pharmacology 35 (2002) 249–254

and characterization of peptides corresponding to unique

regions of A-thrombin that lie outside the active site, yet

exhibit biological activity in vitro and in vivo (Bar-Shavit

and Wilner, 1986; Glenn et al., 1988).

In our working model of A-thrombin activity, events eli-

cited both by proteolytic cleavage and by nonproteolytic

binding comprise a dual signaling mechanism that is required

to achieve the full physiological response to A-thrombin.

Early studies showed that thrombin stimulation of cell

proliferation required two sets of signals: one generated by

proteolytic cleavage and one generated by high-affinity

interaction of A-thrombin (or a thrombin receptor monoclo-

nal antibody) with specific receptor sites (Carney et al.,

1986). We postulated that the signaling events initiated by

nonproteolytic or high-affinity interactions of thrombin are

mediated by separate receptor proteins, termed nonproteo-

lytically activated receptors or N-PARs. It is notable that

Factor Xa, another serine protease involved in the coagu-

lation cascade, has been found to enzymatically cleave and

activate PAR-2 (Cirino et al., 2000), as well as to specifically

bind and activate the EPR-1 receptor without proteolytically

altering it (Altieri, 1995). Thus, it may be that a number of

coagulation cascade proteases have evolved the ability to

initiate distinct physiological responses by cleavage of one

type of receptor and binding of another, thereby triggering

two completely different sets of intracellular signals and/or

allowing cross-talk between the two types of receptors.

To identify the functional binding domains of the A-

thrombin molecule involved in N-PAR activation and dual

signaling, we synthesized and tested a number of molecules

representing different regions of A-thrombin. We became

interested in a specific 23-amino-acid peptide corresponding

to a highly conserved region of the B-chain of the human

prothrombin sequence that extends from amino acids 508 to

530 that became known as thrombin peptide 508, or TP508.

Although the peptide has no proteolytic activity, we initially

found that the 23-mer competes for high-affinityA-thrombin

binding sites on fibroblasts and enhances proliferation of

fibroblasts co-stimulated with proteolytically active A-

thrombin or molecules that activate downstream signal events

normally generated by PAR-1 activation (Carney et al., 1984,

1986, 1992a,b). TP508 has also proved to be chemotactic for

neutrophils (Moller et al., in press), keratinocytes (Sower et

al., 1999), and, as discussed below, endothelial cells. In an

effort to understand how these effects relate to A-thrombin’s

role in tissue repair following injury, we began to test TP508

in animal models of dermal wound healing.

2. A-Thrombin peptide, TP508, as a possible angiogenic

factor in wound healing

In early studies in our laboratory (Carney et al., 1992a,b),

full-thickness incisions were created in the dorsal skin of

normal rats and immediately treated with a single topical

application of TP508 in saline. By Postwounding Day 7,

TP508-treated incisions exhibited up to 80% greater tensile

strength than saline controls, corresponding to a forward

shift in the time course of healing of over 4 days. Radio-

angiography and histology of the Day 7 wounds revealed

that TP508 treatment augmented revascularization of the

incisional wounds, which we speculated could have con-

tributed to the acceleration of healing.

Further studies using a full-thickness excisional model of

dermal wounding confirmed the ability of TP508 to accel-

erate the healing process (Stiernberg et al., 2000): Open

excisions that had been treated with a single topical applica-

tion of TP508 on Day 0 were 35% smaller than controls by

Day 7. Visual inspection of the underside of the dermis

showed an enhanced network of vessels directed towards

the TP508-treated wounds, compared to the saline-treated

wounds. This gross finding was consistent with microscopic

observations of Day 7 wound histology, which indicated

that the granulation tissue of TP508-treated wounds con-

tained larger blood vessels than corresponding control

wounds. In this study, we also found that between 12- and

24-h postwounding, a greater number of leukocytes had

infiltrated the TP508-treated wounds compared to controls.

In an experimental model of ischemic wound healing

(Norfleet et al., 2000), TP508 was again shown to accelerate

wound closure, to augment neovascularization of the granu-

lating wound tissue, and to enhance early leukocyte recruit-

ment. To produce ischemic wounds, full-thickness excisions

were made within a cranially based pedicle flap, a three-

sided skin flap to which all blood vessels are severed except

along the cranial (attached) edge. This procedure results in a

gradient of blood flow that decreases from the cranial to the

caudal end of the flap—and, therefore, from the cranial to

the caudal margin of the excisional wounds within the skin

flap. The decreased perfusion and concomitant increased

hypoxia at the caudal wound edge, relative to the cranial

edge of a given wound, were associated with the presence of

a greater number of large neovessels in the granulation

tissue at Postwounding Day 7. As shown in Fig. 1, TP508

treatment (lower panel) produced a further increase in the

number of large vessels in the granulation tissue at the

caudal margin of ischemic wounds as compared to those

treated with saline (upper panel). By Day 10 postwounding,

the numbers and sizes of the blood vessels did not differ

significantly in control and treated wounds. In fact, some

involution of the neovessels at the margins of the wounds

has occurred by this point, demonstrating that TP508 does

not induce an abnormal, persistent angiogenic response. In

this same study, we found that TP508 enhanced the infiltra-

tion of inflammatory cells into the ischemic wounds by 24 h.

Moreover, the granulation tissue of saline-treated ischemic

wounds was punctuated with high numbers of activated

neutrophils 14 days after wounds, while TP508-treated

wounds contained few of these cells, similar to Day 14

wounds in normal (nonischemic) skin.

Taken together, the observations from dermal wound

healing studies indicated that TP508 treatment results in

A.M. Norfleet et al. / General Pharmacology 35 (2002) 249–254250

enhance neovascularization of the wounded tissue by 7 days

postwounding. While it is not possible from these in vivo

systems to decipher the mechanism by which TP508 pro-

duces this apparent angiogenic response, several general

hypotheses emerge based on current models of the angio-

genic cascade (Augustin, 1998). For example, it is possible

that TP508 directly attracts endothelial cell precursors from

the blood to the wound in a manner similar to its recruitment

of neutrophils and monocytes. Alternatively, TP508 may

directly stimulate endothelial cells from damaged or acti-

vated vessels to migrate and/or proliferate. It is important to

consider, however, that additional factor(s) released from

various cell types at the injured site may be required as co-

stimulators, setting the proper cellular context for TP508

responsiveness. Another mode of TP508 action may pro-

ceed indirectly through stimulation of wound cells to

increase production and/or release of angiogenic factors.

The finding that TP508 augments early recruitment of

inflammatory cells, well-known sources of angiogenic

growth factors and cytokines, suggests that this indirect

mechanism may account in some part for the effect of the

peptide on wound revascularization. We therefore undertook

experiments to identify angiogenic properties of TP508 that

could be measured in an in vivo angiogenic assay or in

cultured cells in vitro.

First, to determine if TP508 could promote angiogenesis

in a context other than that of a healing wound, we initiated

experiments in the chicken chorioallantoic membrane

(CAM) model, a system used extensively to evaluate the

angiogenic capacity of various growth factors (Fett et al.,

1987). For these CAM experiments, an agar disk, impreg-

nated with PBS alone or PBS with TP508, was placed on

the CAM of 9-day-old chick embryos for 96 h. As shown in

Fig. 2, relative to control disks (top panel), the density and

size of blood vessels surrounding TP508-containing disks

(bottom panel) were significantly increased. As has been

typically seen with other growth factors (Fett et al., 1987;

Yang and Moses, 1990), these vessels were arranged in a

radial pattern emanating from the margin of the disk. The

data obtained from two separate experiments (not shown)

suggested that the TP508 effect is dose-dependent between

0.1 and 1.0 mg, although evaluation of earlier time points

and a broader range of doses would be needed to clarify this

preliminary result. Nevertheless, these findings support the

Fig. 1. Large blood vessels are formed in the granulation tissue of ischemic skin wounds treated with TP508. Cross-sections of Day 7 wounds were stained

using Movat’s pentachrome technique. Photomicrographs were obtained at a magnification of 4� near the caudal margin of each wound; the yellow-stained

areas on the left side of each panel represent the nonwounded dermis. Top panel: saline-treated; bottom panel: TP508-treated.

A.M. Norfleet et al. / General Pharmacology 35 (2002) 249–254 251

hypothesis that the enhanced revascularization seen in

TP508-treated wounds involves an angiogenic response to

the peptide.

Like the dermal wound models, the CAM assay is not

well suited to answer mechanistic questions. In both experi-

mental systems, many cell populations are present, includ-

ing monocytes/macrophages, mast cells, lymphocytes,

connective tissue cells, pericytes, and endothelial cells. All

of these cells have been shown to influence the formation of

new vessels by secreting soluble angiogenic and antiangio-

genic molecules, as well as extracellular matrix proteins and

proteolytic enzymes that can promote new vessel formation

or vessel remodeling (Schwartz and Liaw, 1993; Augustin,

1998). In these CAM assays, delivery of test agents is

accomplished by depositing the agent in a carrier, which

can, by itself, elicit a mild inflammatory reaction at the drug

delivery site; this carrier effect, accompanied by the infiltra-

tion of monocytes and neutrophils, could mimic or syner-

gize with angiogenic activity of the test agent. Therefore, to

determine if the effects of TP508 on wound revasculariza-

tion and CAM angiogenesis were related to direct activity of

the peptide on endothelial cells, in vitro studies were un-

dertaken using cultured human microvascular (HMVE) and

human aortic (HAE) endothelial cells.

In initial experiments, we assessed the ability of TP508

to stimulate migration of endothelial cells, since we have

observed chemotactic effects of the peptide in other cell

types (Moller et al., in press; Sower et al., 1999). Using a

modification of a system designed to assess nondirectional

chemokinesis (Hoying and Williams, 1996), either HMVE

or HAE cells were cultured in a defined circular area created

with silicone restraining chambers. After removal of the

restraining chambers, the cells were incubated for 72 h in

media with or without TP508 and the distance cells

migrated outward from the original circular area was meas-

ured. As shown in Fig. 3, a concentration of 40 nM TP508

stimulated chemokinetic activity up to 40% above controls

in HMVE cells and up to 15% in HAE cells. Higher

micromolar concentrations of TP508 appeared to inhibit

cell movement in both cell types, displaying a biphasic

response common for agents that stimulate cell migration. It

is also important to note that under the incubation conditions

employed in the chemokinesis assays, no effect of TP508 on

cell proliferation was observed (not shown).

In further experiments, the ability of TP508 to trigger

HMVE or HAE cell chemotaxis— the directed movement of

cells along a concentration gradient—was examined using a

transwell migration assay (Goldberger and Septak, 1998).

Fluorescently labeled cells were placed in a microporous

insert and allowed to migrate towards wells containing

TP508 at various concentrations. As shown in Fig. 4 (topFig. 2. Response of CAM vasculature to TP508. An agar disk impregnated

with either PBS (top panel) or 1 mg of TP508 in PBS (bottom panel) was

placed on the CAM of 9-day-old chick embryos for 96 h. The CAM area

around the disk was photographed under a dissecting microscope at a

magnification of 2�.

Fig. 3. TP508 stimulates chemokinesis of cultured endothelial cells. HMVE

and HAE cells (Clonetics, San Diego, CA) were plated on Thermonex

coverslips in silicone-restraining chamber with a diameter of 6 mm. After

48 h, the chamber was removed and fresh medium with or without TP508

was added to the cells. After an additional 72 h, the cells were fixed on the

coverslips, then stained with hematoxylin and eosin. From photomicro-

graphs, the distance the cells had migrated from the original 6-mm circle

was measured. Each bar represents the mean ± S.E. of three samples.

A.M. Norfleet et al. / General Pharmacology 35 (2002) 249–254252

panel), HAE cells demonstrated peptide-stimulated chemo-

taxis; after 3 h, the maximum effect of TP508 was observed

at the lowest concentration tested, while higher concentra-

tions inhibited cell migration. The bottom panel of Fig. 4

depicts the chemotactic effects of TP508 on HMVE cells.

After a 2-h exposure to 0.3 mM TP508, HMVE cells

exhibited maximum chemotaxis in the transwell migration

assay with a sevenfold increase over control. Thus, it appears

that TP508 may have a direct effect on endothelial cells

consistent with its effects on angiogenesis in the CAM and

its enhancement of revascularization in dermal wounds. As

described above, however, it is likely that the effects of

TP508 may involve both these direct effects as well as the

indirect effects involving cytokines and other factors re-

leased by inflammatory and resident cells responding to

the peptide.

If TP508 has such dramatic effects on wound healing and

revascularization, one is forced to speculate that this binding

domain region of a-thrombin may have evolved as a natural

initiator of tissue repair processes. It is well known that

a-thrombin is sequestered within fibrin clots where it can

retain activity for extended periods of time (Wilner et al.,

1981). We speculate that, as thrombin (and the fibrin clot) is

degraded at the injured site by wound proteases, including

those produced by recruited neutrophils and monocytes/

macrophages, bioactive ‘‘degradation products’’ correspond-

ing to functional domains of thrombin are generated and

released from the clot. As these peptides diffuse into sur-

rounding tissues, they may exert chemotactic effects, attract-

ing more inflammatory cells and other cell types to the

provisional matrix of the wound where they then contribute

to the healing process. Because a-thrombin circulates in the

plasma at micromolar concentrations (Mann, 1999) and

because thrombin sequestration within the clot further

increases its local concentration (Wilner et al., 1981), sig-

nificant amounts of a peptide like TP508 could be available

at the wound site. Release of thrombin-derived peptides from

the clot most likely entails a gradual process that takes place

over several days. In contrast, in dermal wounds treated with

exogenous, topically applied TP508, cells are exposed to this

naturally stimulatory peptide at an earlier time and in higher

concentrations than would normally occur, thereby acceler-

ating the healing process. Additional studies are underway to

better define the role of TP508 and related peptides that may

be released from blood clots and to determine if the these

peptides trigger similar angiogenic events in other soft and

hard tissues during the process of repair.

References

Altieri, D.C., 1995. Proteases and protease receptors in modulation of leu-

kocyte effector functions. J. Leukocyte Biol. 58, 120–127.

Augustin, H.G., 1998. Antiangiogenic tumour therapy: will it work? Trends

Pharmacol. Sci. 19, 216–222.

Bar-Shavit, R., Wilner, G.D., 1986. Mediation of cellular events by throm-

bin. Int. Rev. Exp. Pathol. 29, 213–241.

Bar-Shavit, R., Kahn, A., Fenton, J.W., Wilner, G.D., 1983. Receptor-medi-

ated chemotactic response of a macrophage cell line (J774) to thrombin.

Lab. Invest. 49, 702–707.

Bar-Shavit, R., Kahn, A.J., Mann, K.G., Wilner, G.D., 1986. Growth-pro-

moting effects of esterolytically inactive thrombin on macrophages. J.

Cell. Biochem. 32, 261–272.

Carney, D.H., Stiernberg, J., Fenton, J.W., 1984. Initiation of proliferative

events by human alpha-thrombin requires both receptor binding and

enzymic activity. J. Cell. Biochem. 26, 181–195.

Carney, D.H., Herbosa, G.J., Stiernberg, J., Bergmann, J.S., Gordon, E.A.,

Scott, D.L., Fenton, J.W., 1986. Double-signal hypothesis for throm-

bin initiation of cell proliferation. Semin. Thromb. Haemostasis 12,

231–240.

Carney, D.H., Mann, R., Redin, W.R., Pernia, S.D., Berry, D., Heggers, J.P.,

Hayward, P.G., Robson, M.C., Christie, J., Annable, C., Fenton, J.W.,

Glenn, K.C., 1992a. Enhancement of incisional wound healing and

Fig. 4. TP508 displays chemotactic activity. HAE and HMVE cells

(Clonetics) were labeled in situ with Calcein-AM (Molecular Probes,

Eugene, OR), then 104 cells were aliquoted into 8-mm Falcon HTS

FluoroBlock inserts (Becton Dickinson, Sparks, MD). The inserts were

placed into 24-well plates that contained medium alone or with varying

concentrations of TP508. The fluorescent signal of cells migrating from the

insert to the lower chamber was detected with a Bio-Tek FL600 plate reader

(Winooski, VT). Data shown are the means ± S.E. of two to three wells.

HAE at 3 h (top panel); HMVE cells at 2 h (bottom panel).

A.M. Norfleet et al. / General Pharmacology 35 (2002) 249–254 253

neovascularization in normal rats by thrombin and synthetic thrombin

receptor-activating peptides. J. Clin. Invest. 89, 1469–1477.

Carney, D.H., Redin, W., McCroskey, L., 1992b. Role of high-affinity

thrombin receptors in postclotting cellular effects of thrombin. Semin.

Thromb. Haemostasis 18, 91–103.

Cirino, G., Napoli, C., Bucci, M., Cicala, C., 2000. Inflammation-coagu-

lation network: are serine protease receptors the knot? Trends Pharma-

col. Sci. 21, 170–172.

Fett, J.W., Bethune, J.L., Vallee, B.L., 1987. Induction of angiogenesis by

mixtures of two angiogenic proteins, angiogenin and acidic fibroblast

growth factor, in the chick chorioallantoic membrane. Biochem. Bio-

phys. Res. Commun. 146, 1122–1131.

Glenn, K.C., Frost, G.H., Bergmann, J.S., Carney, D.H., 1988. Synthetic

peptides bind to high-affinity thrombin receptors and modulate throm-

bin mitogenesis. Pept. Res. 1, 65–73.

Goldberger, A.S., Septak, M., 1998. A fluorescence blocking membrane

insert enhances analysis of cell motility assays. Becton Dickinson Tech.

Bull. 428, 1–5.

Grand, R.J., Turnell, A.S., Grabham, P.W., 1996. Cellular consequences of

thrombin-receptor activation. Biochem. J. 313, 353–368.

Hoying, J.B., Williams, S.K., 1996. Measurement of endothelial cell migra-

tion using an improved linear migration assay. Microcirculation 3,

167–174.

Jenkins, A.L., Howells, G.L., Scott, E., LeBonniec, B.F., Curtis, M.A.,

Stone, S.R., 1995. The response to thrombin of human neutrophils:

evidence for two novel receptors. J. Cell. Sci. 108, 3059–3066.

Mann, K.G., 1999. Biochemistry and physiology of blood coagulation.

Thromb. Haemostasis 82, 165–174.

Moller, M.L., Keherly, M.J., Carney, D.H., 2001. The thrombin peptide

TP508 is a potent chemotactic agent for human neutrophils (PMNs).

Mol. Biol. Cell 12, 257a.

Norfleet, A.M., Huang, Y., Sower, L.E., Redin, W.R., Fritz, R.R., Carney,

D.H., 2000. Thrombin peptide TP508 accelerates closure of dermal

excisions in animal tissue with surgically induced ischemia. Wound

Repair Regener. 8, 517–529.

Schwartz, S.M., Liaw, L., 1993. Growth control and morphogenesis in the

development and pathology of arteries. J. Cardiovasc. Pharmacol. 21,

S31–S49.

Sower, L.E., Huang, Y., Norfleet, A.M., Carney, D.H., 1999. Thrombin

peptide, TP508, induces proliferation and migration of keratinocytes.

Mol. Biol. Cell 10, 186a.

Stiernberg, J., Norfleet, A.M., Redin, W.R., Warner, W.S., Fritz, R.R.,

Carney, D.H., 2000. Acceleration of full-thickness wound healing in

normal rats by the synthetic thrombin peptide, TP508. Wound Repair

Regener. 8, 204–215.

Strukova, S.M., 2001. Thrombin as a regulator of inflammation and repar-

ative processes in tissues. Biochemistry (Moscow) 66, 8–18.

Vergnolle, N., Wallace, J.L., Bunnett, N.W., Hollenberg, M.D., 2001. Pro-

tease-activated receptors in inflammation, neuronal signaling and pain.

Trends Pharmacol. Sci. 22, 146–152.

Wilner, G.D., Danitz, M.P., Mudd, M.S., Hsieh, K.H., Fenton, J.W., 1981.

Selective immobilization of alpha-thrombin by surface-bound fibrin. J.

Lab. Clin. Med. 97, 403–411.

Yang, E.Y., Moses, H.L., 1990. Transforming growth factor beta 1-induced

changes in cell migration, proliferation, and angiogenesis in the chicken

chorioallantoic membrane. J. Cell. Biol. 111, 731–741.

A.M. Norfleet et al. / General Pharmacology 35 (2002) 249–254254