The common pathway for alpha- and gamma-thrombin-induced platelet activation is independent of GPIb: a study of Bernard-Soulier platelets

Phosphatidylserine exposure and other apoptotic-like events in Bernard-Soulier syndrome platelets

In the Bernard-Soulier syndrome (BSS), the giant platelets are said to have increased phosphatidylserine (PS) surface exposure in the resting state and shortened survival in the circulation. When normal platelets are activated, they undergo many biochemical and morphological changes, some of which are apoptotic. Herein, we investigated apoptotic-like events in BSS platelets upon activation, specifically, PS exposure, microparticle (MP) formation, cell shrinkage, and loss of mitochondrial inner membrane potential (DeltaPsi(m)). Platelets from two unrelated BSS patients were examined in whole blood; agonists used were collagen, thrombin, PAR1- or PAR4-activating peptides (APs), or combinations of collagen with thrombin, and the PAR-APs. Flow cytometry was used to measure PS exposure (annexin A5 binding), platelet-derived MPs (forward scatter; events <0.75 microm size), and DeltaPsi(m) (TMRM fluorescence). PS exposure was increased on resting and activated BSS platelets, and this was independent of the platelet size. MP formation by BSS platelets was generally enhanced. Cell shrinkage occurred on activation to form smaller, PS-exposing platelets in BSS and controls. A proportion of PS-exposing BSS and control platelets exhibited DeltaPsi(m) loss, but unlike controls, there was also loss of DeltaPsi(m) in the BSS platelets not exposing PS. Thus, BSS platelets undergo apoptotic-like events upon activation, with PS exposure and MP formation being enhanced. These events may play a role in the shortened survival in BSS, as well as affecting thrombin generation.

Synergistic Effect of Thrombin on Collagen-Induced Platelet Procoagulant Activity Is Mediated Through Protease-Activated Receptor-1

OBJECTIVE

In the blood coagulation process, the rate of thrombin formation is critically dependent on phosphatidylserine (PtdSer) at the surface of activated platelets. Thrombin synergistically enhances the collagen-induced platelet procoagulant response. The objective of this study is to elucidate the mechanism of this synergistic action with a focus on the intracellular Ca2+ concentration ([Ca2+]i) and the various platelet receptors for thrombin.

METHODS AND RESULTS

We demonstrate that procoagulant activity is related to a sustained increased [Ca2+]i, which in turn depends on extracellular Ca2+ influx. Increased PtdSer exposure coincides with increased [Ca2+]i and was observed in a subpopulation (approximately 14%) of the platelets after stimulation with thrombin plus collagen. 2D2-Fab fragments against the thrombin binding site on GPIbalpha made clear that this receptor did not signal for platelet procoagulant activity. Inhibition of protease-activated receptor 1 (PAR-1) and PAR-4 by selective intracellular inhibitors and selective desensitization of these receptors revealed that PAR-1, but not PAR-4, activation is a prerequisite for both sustained elevations in [Ca2+]i and procoagulant activity induced by collagen plus thrombin.

CONCLUSIONS

The interaction of thrombin with PAR-1 mediates a synergistic effect on collagen-induced procoagulant activity by inducing a sustained elevation in [Ca2+]i in a subpopulation of platelets.

Redistribution of glycoprotein Ib within platelets in response to protease-activated receptors 1 and 4: roles of cytoskeleton and calcium

Thrombin activates human platelets by hydrolyzing the protease-activated receptors PAR-1 and PAR-4, exposing new N-terminal sequences which act as tethered ligands, and binding to glycoprotein (GP) Ib, whose surface accessibility transiently decreases when platelets are stimulated by the enzyme. In an attempt to better understand this latter process, we used the peptides SFLLRNPNDKYEPF (PAR-1-AP or TRAP) and AYPGKF (PAR-4-AP) to study whether hydrolysis of both PAR receptors leads to GPIb redistribution. Both peptides induced surface clearance of GPIb with a maximum at 2 min and 5 min for PAR-1-AP and PAR-4-AP, respectively, followed by a slow return to the surface with levels normalizing between 30 and 60 min. Translocation was associated with the formation of clusters of GPIb as revealed by fluorescence microscopy. This transient redistribution of GPIb was blocked by cytochalasin D and in large part by the membrane permeable Ca2+ chelator, BAPTA. The inhibitor of phosphatidylinositol 3-kinase and myosin light chain kinase, wortmannin, did not significantly modify internalization of GPIb, although its return to the surface was delayed for PAR-1-AP. PAR receptor-mediated association of GPIb to the insoluble cytoskeleton was blocked by cytochalasin D, while BAPTA alone increased and stabilized the presence of GPIb. Globally, immunoprecipitation experiments and analysis of the cytoskeleton confirmed that GPIb translocation is powered by a contractile mechanism involving Ca2+ mobilization, actin polymerization, and myosin incorporation into the cytoskeleton and that both PAR-1 and PAR-4 can activate this process.

Glycoprotein Ib-mediated platelet activation. A signalling pathway triggered by thrombin

Platelet activation by thrombin plays a major role in the development of haemostasis and thrombosis. Thrombin activates human platelets by cleaving the N-terminal region of G-protein-coupled protease-activated receptors (PARs). On the other hand, the platelet membrane glycoprotein GPIb acts as a thrombin-binding site and promotes platelet activation by low thrombin concentrations. We present here new evidence in favour of a thrombin receptor function for GPIb. We have selected conditions in which thrombin-GPIb interactions were enhanced by thrombin immobilization. Activation was studied independently of PAR cleavage by using active-site-blocked thrombin. We show that immobilized, proteolytically inactive thrombin induces platelet adhesion and spreading, dense granule secretion and integrin alphaIIbbeta3-dependent platelet-platelet interactions. The pathway must be dependent on GPIb because it is deficient in platelets from a patient with Bernard Soulier syndrome and inhibited by a monoclonal antibody to GPIb (SZ2) or by an excess of glycocalicin. Secreted ADP plays a major role in GPIb-dependent thrombin-induced platelet activation which is, in addition, regulated by cAMP concentration. Thrombin-induced GPIb-dependent platelet activation leads to tyrosyl phosphorylation of several proteins. Inhibition of platelet-platelet interactions and protein tyrosine phosphorylations by inhibitors of phosphatidylinositol 3-kinases and protein kinase C implies that activation of the latter are important steps of the GPIb-coupled signalling pathway triggered by thrombin.

Thrombin-induced platelet PAR4 activation: role of glycoprotein Ib and ADP

Thrombin activates human platelets via the cleavage of two protease-activated G-protein coupled receptors (PARs), PAR1 and PAR4 that respond to low and high concentrations of thrombin, respectively. The aim of the present study was to examine the relative contributions of GPIbalpha and ADP receptors in response to thrombin-induced PAR1 and PAR4 stimulation. Platelet responses (aggregation, secretion and calcium mobilization) elicited by low thrombin concentrations were impaired when thrombin interaction with GPIbalpha was blocked. In contrast, blockade of thrombin interaction with GPIbalpha had no effect when PAR4-coupled responses were specifically elicited by high thrombin concentrations in the presence of PAR1 antagonists or after PAR1 desensitization. These results confirmed that unlike PAR1, PAR4 does not require GPIbalpha as a cofactor for thrombin-mediated activation. Both apyrase and selective antagonists of P2Y1 and P2Y12 inhibited PAR1-coupled responses but did not modify PAR4-coupled responses, indicating that in contrast to PAR1, PAR4 signals are not reinforced by ADP secretion and binding to the platelets. These results provide the direct evidence that, in human platelets, GPIbalpha and ADP act in synergy to amplify PAR1 coupled responses while PAR4 is activated independently of GPIbalpha and ADP.

Human factor XII binding to the glycoprotein Ib-IX-V complex inhibits thrombin-induced platelet aggregation

Factor XII deficiency has been postulated to be a risk factor for thrombosis suggesting that factor XII is an antithrombotic protein. The biochemical mechanism leading to this clinical observation is unknown. We have previously reported high molecular weight kininogen (HK) inhibition of thrombin-induced platelet aggregation by binding to the platelet glycoprotein (GP) Ib-IX-V complex. Although factor XII will bind to the intact platelet through GP Ibalpha (glycocalicin) without activation, we now report that factor XIIa (0. 37 microm), but not factor XII zymogen, is required for the inhibition of thrombin-induced platelet aggregation. Factor XIIa had no significant effect on SFLLRN-induced platelet aggregation. Moreover, an antibody to the thrombin site on protease-activated receptor-1 failed to block factor XII binding to platelets. Inhibition of thrombin-induced platelet aggregation was demonstrated with factor XIIa but not with factor XII zymogen or factor XIIf, indicating that the conformational exposure of the heavy chain following proteolytic activation is required for inhibition. However, inactivation of the catalytic activity of factor XIIa did not affect the inhibition of thrombin-induced platelet aggregation. Factor XII showed displacement of biotin-labeled HK (30 nm) binding to gel-filtered platelets and, at concentrations of 50 nm, was able to block 50% of the HK binding, suggesting involvement of the GP Ib complex. Antibodies to GP Ib and GP IX, which inhibited HK binding to platelets, did not block factor XII binding. However, using a biosensor, which monitors protein-protein interactions, both HK and factor XII bind to GP Ibalpha. Factor XII may serve to regulate thrombin binding to the GP Ib receptor by co-localizing with HK, to control the extent of platelet aggregation in vivo.

Human Kininogens Regulate Thrombin Binding to Platelets Through the Glycoprotein Ib-IX-V Complex

We and others have shown that both high and low molecular mass kininogens are able to inhibit the thrombin-induced aggregation of gel-filtered platelets, indicating that the locus for inhibition resides in the heavy chain. The inhibitory site is present in domain 3, confined to the C-terminal portion of the region encoded by exon 7 (K270-G292), and the minimal effective sequence is a heptapeptide (L271-A277; Kunapuli et al, J Biol Chem 271:11228, 1996). Kininogens inhibit thrombin binding to platelets and thus inhibit thrombin-induced aggregation. The molecular mechanism by which kininogens inhibit thrombin-induced aggregation of platelets is unknown. Thrombin has previously been shown to bind to two receptors on the platelet surface, glycoprotein (GP) Ib-IX-V complex and the hepta-spanning transmembrane receptor coupled to G protein(s). We now show that, unlike its effect on normal platelets, kininogen (2 μmol/L) did not inhibit the thrombin-induced aggregation of Bernard-Soulier platelets, which lack the GP Ib-IX-V complex, suggesting that kininogen interacts either directly or indirectly with that complex and restricts access by thrombin to this receptor. We further show that both recombinant K270-G292 polypeptide and the synthetic peptide L271-A277 derived from high molecular mass kininogen lower thrombin binding to platelets in a manner similar to monoclonal antibodies to or ligands (von Willebrand factor and echicetin) of GP Ib-IX. The anti–GP Ib-IX-V complex antibodies, TM-60 and SZ 2, can inhibit 125I-high molecular mass kininogen binding to platelets. Conversely, kininogen could block the binding of biotinylated TM-60 or of 125I-SZ 2. Kininogen inhibited the binding of biotinylated thrombin bound to a mouse fibroblast cell line transfected with the GP Ib-IX-V complex. These results indicated that kininogen binds to the GP Ib-IX-V complex modulating thrombin binding to platelets and the consequent platelet aggregation. Kininogen can thus serve as an important regulator of the early stages of platelet stimulation by thrombin.

Human Kininogens Regulate Thrombin Binding to Platelets Through the Glycoprotein Ib-IX-V Complex

We and others have shown that both high and low molecular mass kininogens are able to inhibit the thrombin-induced aggregation of gel-filtered platelets, indicating that the locus for inhibition resides in the heavy chain. The inhibitory site is present in domain 3, confined to the C-terminal portion of the region encoded by exon 7 (K270-G292), and the minimal effective sequence is a heptapeptide (L271-A277; Kunapuli et al, J Biol Chem 271:11228, 1996). Kininogens inhibit thrombin binding to platelets and thus inhibit thrombin-induced aggregation. The molecular mechanism by which kininogens inhibit thrombin-induced aggregation of platelets is unknown. Thrombin has previously been shown to bind to two receptors on the platelet surface, glycoprotein (GP) Ib-IX-V complex and the hepta-spanning transmembrane receptor coupled to G protein(s). We now show that, unlike its effect on normal platelets, kininogen (2 μmol/L) did not inhibit the thrombin-induced aggregation of Bernard-Soulier platelets, which lack the GP Ib-IX-V complex, suggesting that kininogen interacts either directly or indirectly with that complex and restricts access by thrombin to this receptor. We further show that both recombinant K270-G292 polypeptide and the synthetic peptide L271-A277 derived from high molecular mass kininogen lower thrombin binding to platelets in a manner similar to monoclonal antibodies to or ligands (von Willebrand factor and echicetin) of GP Ib-IX. The anti–GP Ib-IX-V complex antibodies, TM-60 and SZ 2, can inhibit 125I-high molecular mass kininogen binding to platelets. Conversely, kininogen could block the binding of biotinylated TM-60 or of 125I-SZ 2. Kininogen inhibited the binding of biotinylated thrombin bound to a mouse fibroblast cell line transfected with the GP Ib-IX-V complex. These results indicated that kininogen binds to the GP Ib-IX-V complex modulating thrombin binding to platelets and the consequent platelet aggregation. Kininogen can thus serve as an important regulator of the early stages of platelet stimulation by thrombin.

Differences between contrast media in the inhibition of platelet activation by specific platelet agonists

RATIONALE AND OBJECTIVES

The authors evaluated the ability of three x-ray contrast agents--a nonionic monomeric agent (iohexol), a nonionic dimeric agent (iodixanol), and an ionic dimeric agent (ioxaglate)--to either directly activate platelets or inhibit a platelet agonist from activating platelets.

METHODS

Fluorescence spectroscopy was used to detect the effect of contrast media on platelet activation. In this method, the platelet is first exposed to a fluorescent probe, which is de-esterified and trapped to Fluo-3 within the platelet. In the presence of calcium, the fluorescence emission from Fluo-3 is increased 80-fold. Thus, the increase in the free platelet calcium associated with platelet activation can be used to indicate platelet activation.

RESULTS

None of the agents were shown to directly activate platelets. However, wide differences in the ability of contrast media to inhibit platelet activation by a specific agonist were observed. Activation of platelets by epinephrine or arachidonic acid was not affected by any of the three contrast media studied. All three agents partially inhibited collagen activation of platelets, with ioxaglate the more potent inhibitor. Ioxaglate was the only agent to inhibit thrombin activation of platelets. Inhibition of adenosine diphosphate platelet activation was more extensive with ioxaglate than with iodixanol; iohexol produced no inhibition at all.

CONCLUSION

Direct activation of platelets by contrast media was not observed. Of greater importance is the finding that ionic contrast media, but not nonionic contrast media, inhibit thrombin activation of platelets by binding to the anion-binding exosite I, thus preventing thrombin from binding to and activating the platelet.

Defective thrombin-induced calcium changes and aggregation of Bernard-Soulier platelets are not associated with deficient moderate-affinity receptors

Cloning of the moderate-affinity, serpentine thrombin receptor has helped clarify the mechanism of thrombin-induced platelet activation. Proteolytic cleavage by thrombin generates a new amino terminal that autostimulates the receptor, leading to activation of multiple signaling pathways and the platelet response. The function of other thrombin receptors, such as high-affinity glycoprotein Ib (GPIb), on platelets and their relationships to the moderate-affinity receptor remain unclear. The present study examined the role of the moderate-affinity thrombin receptor in Bernard-Soulier syndrome (BSS) platelets, which contain low amounts of GPIb. Platelets from four BSS subjects displayed normal aggregation profiles and cytosolic calcium changes in response to moderate or high concentrations of thrombin. In contrast, the BSS platelet aggregation response was delayed and calcium changes were absent in response to low thrombin concentrations. Platelets from an asymptomatic BSS heterozygote displayed an activation profile similar to those of control individuals. Specific activation of the moderate-affinity receptor by a synthetic peptide caused similar aggregation in platelets from all individuals. The synthetic peptide also elicited calcium responses in BSS platelets. Platelets from the BSS subjects and from an individual with the May-Hegglin anomaly showed increased expression of the moderate-affinity thrombin receptor by flow-cytometric analyses. These results suggest that BSS platelets possess high levels of a functional moderate-affinity thrombin receptor, probably due to large platelet size, and provide indirect evidence that a high-affinity thrombin receptor is associated with GPIb.

Disorders of platelet function

Qualitative platelet disorders are described and reviewed above. The acquired platelet function defects are very common, and sometimes result in hemorrhage, especially in association with trauma or surgery. However, the specific biochemical defect is absent, and no characterized platelet abnormalities have been recognized. On the other hand, the hereditary qualitative platelet defects are rare, but the platelet abnormalities are characteristic. The study of these patients had led to an increased understanding of the normal primary hemostatic mechanism. Recently, the molecular basis analysis of the platelet defects has been developed. This will help us understand the molecular events involved in platelet adhesion and aggregation.

Thrombin interaction with a recombinant N-terminal extracellular domain of the thrombin receptor in an acellular system

The cDNA of the human endothelial cell thrombin receptor has been cloned and a chimeric fusion protein consisting of glutathione-S-transferase (GST) and the portion 25-97 corresponding to the N-terminal first extracellular domain of the thrombin receptor (TRE) has been expressed in Escherichia coli. Introduction of a factor Xa cleavage site in the fusion protein allowed purification of TRE after removal from the GST carrier protein. Purified GST-TRE or TRE have been tested in solution for their ability to interact with thrombin. alpha-Thrombin cleaved the fusion protein at position Arg-41-Ser-42 of TRE in a time- and concentration-dependent manner and GST-TRE competed with the tripeptidic substrate S-2238 for hydrolysis by thrombin (Ki = 0.5 microM). gamma-Thrombin that lacks the anion-binding exosite was 100-fold less potent than alpha-thrombin at cleaving GST-TRE. TRE competed with polymerizing fibrin monomers for binding to thrombin (Ki = 7.5 microM). The cleavage of GST-TRE by alpha-thrombin was inhibited by several alpha-thrombin exosite ligands such as the C-terminal peptide of hirudin, thrombomodulin and fibrin(ogen) fragment E. In contrast, platelet glycocalicin did not inhibit GST-TRE cleavage. In conclusion, the use of purified soluble GST-TRE allowed us to derive an affinity constant for thrombin interaction with the N-terminal domain of the receptor and to confirm the location of the cleavage site at Arg41-Ser-42 of the receptor. The importance of the thrombin anion-binding exosite for thrombin receptor recognition is highlighted by the low reactivity of gamma-thrombin for GST-TRE and by competition experiments, which in addition indicate that binding sites for fibrin(ogen), thrombomodulin and GST-TRE are overlapping. In contrast, binding of thrombin to GST-TRE and glycocalicin are not mutually exclusive, indicating that glycocalicin and TRE interact with discrete subsites within the large groove that constitutes the anion-binding exosite.

Thrombin-receptor agonist peptides, in contrast to thrombin itself, are not full agonists for activation and signal transduction in human platelets in the absence of platelet-derived secondary mediators

Synthetic thrombin receptor peptides (TRPs), comprising the first 6-14 amino acids of the new N-terminus tethered ligand of the thrombin receptor that is generated by thrombin's proteolytic activity, were reported to activate platelets equally with thrombin itself and are considered to be full agonists [Vu et al. (1991) Cell 64, 1057-1068]. Using aspirin plus ADP-scavengers or the ADP-receptor antagonist adenosine 5'-[alpha-thio]triphosphate to prevent the secondary effects of the potent agonists that are normally released from stimulated platelets (i.e. ADP and thromboxane A2), we assessed the direct actions of thrombin and TRPs (i.e. TRP42-47 and TRP42-55). Compared with thrombin, under these conditions, TRPs: (1) failed to aggregate platelets completely; (2) produced less activation of glycoprotein (GP)IIb-IIIa; (3) did not cause association of GPIIb and pp60c-src with the cytoskeleton; and (4) caused less alpha-granule secretion, phosphorylation of cytoplasmic phospholipase A2, arachidonic acid release and phosphatidyl inositol (PtdOH) production. Furthermore, TRPs induced transient increases in protein phosphorylation mediated by protein kinase C and protein tyrosine phosphorylation, whereas these same responses to thrombin were greater and more sustained. Hirudin added after thrombin accelerated protein dephosphorylation, thereby mimicking the rate of spontaneous dephosphorylation seen after stimulation by TRPs. Platelets totally desensitized to very high concentrations of TRPs, by prior exposure to maximally effective concentrations of the peptides, remained responsive to alpha- and gamma-thrombins. Thrombin-stimulated PtdOH production in permeabilized platelets desensitized to TRPs was abolished by guanosine 5'-[beta-thio]diphosphate (GDP[beta S]), as in normal platelets. These results are discussed in terms of the allosteric Ternary Complex Model for G-protein linked receptors [Samama et al. (1993) J. Biol. Chem. 268, 4625-4636]. We conclude that: (1) TRPs are partial agonists for the thrombin receptor and produce incomplete receptor desensitization in keeping with their lower intrinsic activity; (2) thrombin's effects in platelets, even in TRP-desensitized platelets, are entirely mediated through the recently cloned G-protein linked receptor, and (3) thrombin's ability to produce sustained signals, compared with TRPs, may require the continued progressive proteolytic activation of naive thrombin receptors.

Phospholipase C activity in platelets from Bernard-Soulier syndrome patients

The levels of glycoprotein (GP) Ib and GPV and phospholipase C activity were measured in platelets from three Bernard-Soulier syndrome patients. The patients' platelets had 46%, 46%, and 24% of control levels of GPIb alpha and 43%, trace, and 13% of control levels of GPV as determined by immunoblot analysis. Stimulation by thrombin, trypsin, the thromboxane analogue U46619, and the combination of U46619 and trypsin caused the formation of [32P]phosphatidic acid, an index of phospholipase C activity, in [32P]orthophosphate-prelabeled platelets. With all agonists, however, the formation of [32P]phosphatidic acid was markedly reduced in Bernard-Soulier syndrome platelets compared with control platelets. These data indicated a postreceptor defect in phospholipase C activation in Bernard-Soulier syndrome platelets and confirmed earlier observations of potential proteolytic and nonproteolytic mechanisms of platelet activation.

Activation of the cloned platelet thrombin receptor decreases the pertussis-toxin-dependent ADP-ribosylation of the membrane and soluble inhibitory guanine-nucleotide-binding-alpha proteins. Inhibition by the prostacyclin analog, iloprost

Thrombin binds at least to two sites of the platelet surface; to the recently cloned thrombin receptor [Vu, T. K., Hung, D. T., Wheaton, V. I. & Coughlin, S. R. (1991) Cell 64, 1057-1068] and to glycoprotein Ib. In the present study, the decrease of pertussis-toxin-dependent ADP-ribosylation of membrane and soluble inhibitory guanine-nucleotide-binding alpha (Gi alpha) proteins was measured after platelet stimulation with a thrombin-receptor-activating peptide (TRAP), and compared to stimulation with thrombin. Stimulation of intact platelets with TRAP decreased the pertussis-toxin-dependent ADP-ribosylation of the major membrane 41-kDa Gi alpha protein and the minor soluble 40 kDa Gi alpha protein recently described in platelets [Gennity, J. M. & Siess, W. (1991) Biochem. J. 279, 643-650]. The kinetics and extent of the decrease of pertussis-toxin-dependent ADP-ribosylation after stimulation of TRAP were similar to the effect of thrombin. The decrease of pertussis-toxin-dependent ADP-ribosylation of the soluble Gi alpha protein was more pronounced and observed at lower agonist concentrations than the decrease of the membrane Gi alpha protein. Desensitization of the thrombin receptor by incubating platelets with a low concentration of TRAP reduced the subsequent decrease of pertussis-toxin-dependent ADP-ribosylation of Gi alpha proteins, evoked by TRAP or thrombin. Platelet stimulation with gamma-thrombin that does not bind to glycoprotein Ib also showed a decrease in the pertussis-toxin-dependent ADP-ribosylation of the soluble and membrane Gi alpha proteins. Treatment of platelets with the stable prostacyclin analog, iloprost, reduced the decrease of pertussis-toxin-dependent ADP-ribosylation of Gi alpha proteins induced by TRAP or thrombin. Among other platelet stimuli tested (endoperoxide/thromboxane analog U44619, collagen, ADP, vasopressin), only U44619 decreased the pertussis-toxin-dependent ADP-ribosylation of the soluble and membrane Gi alpha proteins to a degree comparable to TRAP. It is concluded that the thrombin-induced activation of both the membrane and soluble Gi alpha proteins in platelets occurs via stimulation of the recently cloned thrombin receptor and is independent of the binding of thrombin to glycoprotein Ib. Furthermore, the coupling thrombin receptor/Gi protein is reduced by intracellular cAMP.

Enhancement of the synthesis and secretion of nerve growth factor in primary cultures of glial cells by proteases: a possible involvement of thrombin

Newborn rat brain astrocytes cultured in vitro in a chemically defined medium are shown to secrete enhanced levels of nerve growth factor (NGF) when they are exposed to various types of proteases. Proteolytic enzymes such as alpha-thrombin or collagenase induce a continuous, dose-dependent enhancement of the levels of cell-secreted NGF. Incubation of astrocytes for a 24-h period with 300 ng/ml of alpha-thrombin (approximately 9 nM, or 1 U/ml) results in an increase of the levels of cell-secreted NGF by a factor of three- to fourfold, and at doses 10 times higher, stimulation by a factor of up to four- to fivefold was observed. This phenomenon reflects an enhancement of the cellular pool of NGF mRNA, already noticeable after 3 h of treatment, which is preceded by a temporary activation of protooncogenes encoding transcription factors of the AP-1 family, such as c-fos, c-jun or junB. Trypsin, plasmin, alpha-chymotrypsin, or elastase also enhanced, to different extents, the levels of cell-secreted NGF. However, unlike alpha-thrombin or collagenase, these enzymes cause, above a critical concentration, an extensive cell detachment from the solid support, and this is accompanied by a decrease of their activity on the production of NGF, so that their dose-response curves are bell shaped. Stimulation was maximal at those concentrations that cause a limited loosening of the cell-substratum interactions, as evidenced by a retraction of some cell processes after 24 h of treatment. Studies of the effect of alpha-thrombin indicate that the proteolytic activity itself is required to enhance the production of NGF by astrocytes. Inactivation of alpha-thrombin with D-phenyl-alanyl-L-propyl-L-arginine chloromethyl ketone, phenylmethylsulfonyl fluoride, antithrombin III, or hirudin results in a marked decrease of the stimulatory effect. Furthermore, the prolonged presence of alpha-thrombin is required to elicit a maximal effect on the levels of extracellular NGF, which was observed after 48 h of treatment. It is known that some effects of alpha-thrombin require binding to the cell surface. We found that gamma-thrombin, which still has some proteolytic activity but has lost its ability to bind to the cell surface, is almost as potent as alpha-thrombin in promoting the release of NGF. It is concluded that the effect of thrombin on NGF synthesis is essentially mediated by its proteolytic activity.(ABSTRACT TRUNCATED AT 400 WORDS)