Suppression of Plasminogen Activator Inhibitor-1 Release From Human Cerebral Endothelium by Plasminogen Activators
A Factor Potentially Predisposing to Intracranial Bleeding
Background Intracranial bleeding is the most catastrophic potential complication of treatment with thrombolytic agents. To identify potential factors that may contribute to this problem, we characterized elaboration by human brain endothelial cells of plasminogen activator inhibitor-1 (PAI-1) and measured PAI-1 mRNA levels.
Methods and Results When human cerebral microvascular endothelial cells (HCMEC), pial arterial endothelial cells, and middle meningeal arterial endothelial cells were exposed to 10 to 1000 ng/mL recombinant tissue-type plasminogen activator (RTPA), urokinase-type plasminogen activator (UPA), or streptokinase/plasminogen (37 U streptokinase plus 2 μmol/L plasminogen) for 24 hours, they exhibited concentration-dependent decreases in elaboration of PAI-1 of 65±3%, 48±3%, and 59±8%. UPA and streptokinase/plasminogen elicited decreases of 33±8% and 35±4%, respectively, that were specific with respect to the protease agonists as to total protein synthesis and cell type; ie, neither human umbilical vein endothelial cells nor cerebral pericytes exhibited inhibition of PAI-1 elaboration. No decrease in HCMEC PAI-1 elaboration was induced by coagulation factor Xa (10 nmol/L). A 2.7±0.5-fold increase was induced by α-thrombin (10 nmol/L). PAI-1 secretion from HCMEC decreased within 4 hours of exposure to 100 ng/mL RTPA. In HCMEC exposed to RTPA for 8 hours, PAI-1 mRNA decreased from 176±20 to 43±2.2 pg/μg RNA.
Conclusions These results indicate that brain endothelial cells exposed to RTPA exhibit paradoxically diminished elaboration of PAI-1. This property may render brain vasculature vulnerable to attack by serine proteases, thereby predisposing to injury and initiating an underlying subsequent intracerebral hemorrhage in patients given plasminogen activators for treatment of coronary thrombosis.
Despite the remarkable success of coronary thrombolysis with plasminogen activators in salvaging ischemic myocardium and improving prognosis after acute myocardial infarction, ICB remains a dreaded, albeit rare, complication. Cerebral bleeding often occurs after the exogenous activators have been cleared from the blood, perhaps because minute bleeding initiated earlier leads to massive hemorrhage or because proteolytic injury to vessels sets the stage for later bleeding as a result of continuing, unopposed (see below) endogenous fibrinolysis. We have hypothesized that ICB is attributable, at least in part, to occult derangements in brain vasculature, rendering it peculiarly susceptible to degradation by serine proteases,1 consistent with the striking correlations between risk of ICB and advanced age, uncontrolled and long-standing hypertension, deposition of β-amyloid pleated proteins in the microvasculature, amyloid plaques typical of Alzheimer's disease, and previous cerebrovascular accidents. This hypothesis also is consistent with the lack of induction of ICB in normal laboratory animals, even with massive doses of TPA.1 2
Both procoagulant and fibrinolytic proteins are synthesized by endothelial cells (reviewed in Reference 3). Endothelial cell membranes provide an essential surface for the assembly of complexes involved in both systems. Thus, overall and particularly local regulation of proteolysis occurs at the level of the endothelial cell.
To identify factors that may serve as targets for reduction of the risk of induction of ICB in patients given fibrinolytic drugs, we sought in the present study to characterize the specific biological factors or phenomena that may render brain vessels susceptible to plasminogen activator–induced proteolysis; activation of matrix zymogens by conversion of tissue plasminogen to plasmin known to activate collagenase, stromelysin, and other metalloproteases; augmented degradation of matrix proteins; and consequently injury. Our results implicate anomalous suppression of synthesis of PAI-1 by brain endothelium exposed to plasminogen activators as a potentially modifiable factor predisposing to ICB.
Reagents and Supplies
Collagenase (type 1) was obtained from Worthington; rabbit anti–factor VIII/von Willebrand factor and fluorescein-labeled goat anti-rabbit IgG were obtained from Cappel Laboratories; fluorescein-labeled Ulex europaeus agglutinin-1 lectin came from Vector Laboratories; antibodies to glial fibrillary acidic protein were obtained from Boehringer Mannheim; anti–α-actin and anti–β-actin were purchased from Sigma; medium M-199, calf serum, neomycin, nystatin, basal Eagle's medium, vitamins, amino acids, penicillin/streptomycin, l-glutamine, EDTA, trypsin, HBSS, and RNA calibration mixture were obtained from GIBCO/BRL; Pentex fatty acid–poor, fraction V BSA was purchased from Miles Laboratories, Inc; BCA protein assay reagent came from Pierce; HEPES was obtained from United States Biochemical Corp; and cycloheximide and other reagents were obtained from Sigma Chemical Co unless otherwise specified.
Modified medium M-199, pH 7.2 to 7.4, was formulated with 9.87 g/L medium M-199 powder, 1% basal medium Eagle's vitamin solution (×100), 5 mmol/L glucose, 0.1 g/L neomycin (1120 U), 26.2 mmol/L NaHCO3, 2 mmol/L l-glutamine, and 20% FBS. Serum-free medium was formulated as modified medium M-199, 0.35% BSA, and 20 mmol/L HEPES, pH 7.2 to 7.4. PBS–Tween 80 contained 137 mmol/L NaCl, 6 mmol/L Na2HPO47H2O, 2.7 mmol/L KCl, 1.5 mmol/L KH2PO4, pH 7.4, and 1.0% Tween 80.
HCMEC were obtained from fragments of cerebral cortex at the time of clinically mandated neurosurgical procedures at Fletcher Allen Health Care (previously the Medical Center Hospital of Vermont), Burlington, Vt, consisting exclusively of gray matter shown histologically to exhibit no pathological changes. All large blood vessels were removed before isolation of cells. The tissue was obtained with the assistance of a neurological team. All patients gave written informed consent.
HCMEC were isolated by subjecting finely minced tissue to repeated cycles of digestion with collagenase,4 after which the supernatant containing microvessels was forced through a 70-μm mesh to separate the cells from undigested tissue fragments. Cells were pelleted and seeded in medium M-199 supplemented with 20% heat-inactivated FBS, 100 μg/mL neomycin, 10 μg/mL amphotericin B, 20 U/mL nystatin, and 2 mmol/L l-glutamine. As the cells began to proliferate, microvascular endothelium was isolated further by scraping of contaminating cell types from the surface of the culture dish with a rubber policeman. Cells were identified as endothelial on the basis of the following four criteria: positive immunocytochemical staining for factor VII/von Willebrand factor antigen5 ; binding of Ulex europaeus agglutinin-1 lectin, a specific human endothelial cell marker maintained throughout cell passage6 ; negative staining for glial fibrillary acidic protein, confirming that the cells were not of astrocytic origin7 ; and the capacity to produce prostacyclin delineated as described previously.4 8 The HCMEC used in this study were derived from nine donors, five men and four women, ranging in age from 27 to 82 years.
Pericytes were derived from human brain microvessels as described above and distinguished from endothelial cells on the basis of their larger size, irregular morphology, and overlapping growth pattern. Colonies of pericytes were demarcated on the culture vessel with an inverted phase microscope, and all other cells were scraped from the culture dish with a rubber policeman. Purity of human pericyte cultures was determined on the basis of failure of the cells to reach confluence or exhibit contact inhibition, consistently large irregular morphology, and typically slow rate of growth. Pericytes were distinguished further from endothelial and smooth muscle cells on the basis of a lack of staining by factor VIII/von Willebrand factor antigen and glial fibrillary acidic protein antigen and positive staining for both α-smooth muscle and β–nonmuscle actin isoforms as previously described.9 10 They were subcultured in modified medium M-199 and used between passages 1 and 3.
Human pial artery and middle meningeal artery endothelial cells were derived from explants. Segments of pial and middle meningeal arteries were provided by the Totman Vascular Research Laboratory, University of Vermont, with the assistance of Drs Rosemary Bevan and Carrie Walters, from patients who provided written informed consent in each case. After the tissue had been minced finely and anchored to scored culture dishes, the explanted segments began to proliferate. Endothelial cells were selected initially on the basis of morphology and growth patterns. All other cells were scraped away from the culture dish. With subcultivation, the cells were characterized as endothelial as described above for HCMEC. Cells were cultured in medium M-199 and used between passages 2 and 4.
Primary cultures of HUVEC were established11 with the use of digestion with collagenase. On isolation, cells were grown to confluence in modified medium M-199 and used at passages 1 and 2.
HCMEC were plated onto 24-well (2.9 cm2 per well) culture plates at an initial seeding density of 5×104 cells per well. Cells were grown to confluence in M-199 medium supplemented with 20% heat-inactivated FBS, 2 mmol/L l-glutamine, 100 μg/mL neomycin, and 20 U/mL nystatin. Before each experiment, serum-containing medium was removed, the cell monolayer was washed twice with PBS, and serum-free medium with or without RTPA or other agents was added at designated intervals. When conditioned medium was harvested, cellular debris was removed by centrifugation. PBS–Tween 80 (final concentration, 0.01%) was added to each sample, and samples were stored at −20°C until assay. Cells were lysed with 1 mL per well of 5 mol/L guanidine isothiocyanate and stored at −80°C until isolation of RNA.
The concentration of UPA antigen was measured with a two-site ELISA as previously described.12 PAI-1 in serum-free medium was measured with a monoclonal antibody–based specific ELISA for latent, active, and complexed PAI-1 with a sensitivity of 0.02 ng PAI-1 per 1 mL as described previously.13 All reagents used in the UPA and PAI-1 ELISAs were supplied by Prof De´sire´ Collen, Catholic University of Leuven, Belgium. Total protein concentration was determined with the use of a bicinchoninic acid protein assay with BSA calibrated according to the manufacturer's protocol. Viability of cells was determined by light microscopy, ie, demonstration of exclusion of trypan blue. Five hundred cells from at least five representative fields per culture well were scored for viability in both control and RTPA-exposed cultures, and viability was expressed as the number of cells excluding dye divided by the total number of cells counted times 100%.
Inhibition of protein synthesis was accomplished and verified by metabolic labeling procedures with 5 μg/mL cycloheximide in serum-free medium for 24 hours in both control and experimental cell cultures.
RTPA (Activase) was a gift from Genentech, Inc. The material used comprised both single-chain and two-chain TPA in a 70:30 ratio and in a vehicle of 200 mmol/L arginine hydrochloride, 50 mmol/L sodium monobasic and dibasic phosphate, and 0.01% Tween 80, pH 7.2.
Streptokinase (Kabikinase) was purchased as lyophilized powder in vials of 750 000 IU from KabiVitrum and reconstituted in 5 mL PBS to yield a final concentration of 150 000 U/mL. For experiments, 36 IU/mL streptokinase was combined with 2 μmol/L human plasminogen before addition of the complex to cells.
SCUPA was provided by Collaborative Research. Its amidolytic activity with the chromogenic substrate S2444 (American Diagnostics) was consistent with a maximum of 1.3% content of two-chain urokinase. After conversion to the two-chain form by treatment with plasmin, amidolytic activity was 119 200 IU/mg as calibrated with the international reference preparation for urokinase (obtained from Dr P.J. Gaffney, National Institute for Biological Standards and Control). α-Thrombin (3200 IU/mg) was provided by Dr Kenneth G. Mann (University of Vermont) and prepared by the method of Lundblad et al.14 Human coagulation factor Xa was provided by Dr Paul Haley, Hematologic Technologies. Human plasmin, prepared by the method of Claeys et al,15 and plasminogen, prepared by the method of Lijnen et al,16 were purified from human plasma.
RTPA modified with the chloromethylketone FPRCK was provided by Genentech. Its catalytic activity was only 0.4% that of RTPA. Before use in experiments, FPRCK-modified RTPA (1 mg/mL) was treated with a fivefold molar excess of FPRCK for 4 hours at room temperature and dialyzed extensively against buffer containing 200 mmol/L l-arginine hydrochloride, 50 mmol/L NaH2PO4, pH 7.2, and 0.01% Tween 80. Residual activity was <0.01%.
Quantification of PAI-1 mRNA by Slot Blot and Northern Blot Analyses
Total cellular RNA was extracted from cells by lysis with 1 mL per well of 5 mol/L guanidinium isothiocyanate and 25 mmol/L sodium citrate, pH 7.0, containing 0.5% Sarkosyl and 8% β-mercaptoethanol, as described by Chomczynski and Sacchi,17 followed by cold phenol extraction. The final RNA pellet was resuspended in 50 μL H2O, and the concentration of RNA was determined by absorbance at 260 nm.
RNA was denatured with 6 mol/L glyoxal for 1 hour at 50°C. An aliquot of 5.0 μg was applied to a nylon membrane (Zetaprobe, Bio-Rad) with a slot blot filtration apparatus (Schleicher & Schuell). Concentrations of PAI-1 mRNA ranging from 1000 to 0.05 pg/mL were applied to each membrane to generate a standard curve. PAI-1 mRNA was prepared with the use of an in vitro SP6 polymerase system (Riboprobe, Promega) and quantified by absorbance at 260 nm.18 A 2000-bp-long PAI-1 cDNA, containing most of the coding sequence and 700 bp of the 3′ untranslated sequence,19 was cloned into pSP65. Human spleen RNA was used as a control to determine nonspecific hybridization. After application of the RNA, the membrane was baked for 2 hours at 80°C under vacuum and prehybridized for at least 4 hours at 65°C in 50% formamide; 5× SSC; 10× Denhardt's solution; 50 mmol/L sodium phosphate buffer, pH 7.6; 5% dextran sulfate; 1 mmol/L EDTA; 1% SDS containing heat-denatured tRNA (500 μg/mL); and sonified, heat-denatured salmon sperm DNA (200 μg/mL).
Hybridization was performed overnight at 65°C in the same solution containing specific [32P]-RNA probes for PAI-1 and 18S rRNA. Antisense RNA probes were generated by either SP6 polymerase transcription of linearized pSP65 plasmids containing nucleotides 1045 through 1481 of PAI-1 according to the published cDNA sequence19 or T7 polymerase transcription of a linearized pT7 RNA 18S template (Ambion). The template contained an 80-bp antisense fragment of a highly conserved region of the human 18S rRNA gene. The fresh probes were prepared with a Promega Biotec transcription kit and exhibited specific activity of 0.5 to 2.0×109 cpm/μg RNA. They were heat denatured and added to the hybridization mixture at a concentration of 106 cpm/mL. Membrane washing at 65°C was performed as described by the manufacturer of Zetaprobe, with a final wash of 0.05× SSC at 65°C. Autoradiography was performed with the use of Reflection NEF autoradiographic film (Du Pont/New England Nuclear) at −72°C.
PAI-1 mRNA applied to the membrane was quantified by measurement of the radioactivity present in each individual sample (slot) by liquid scintillation counting. A total of three samples (5 μg total RNA isolated at each time point and from cells subjected to each condition) were loaded onto the membrane. Radioactivity in the slots containing the known PAI-1 standards (1000 to 0.05 pg/mL) was determined with the same procedure. Counts per minute from the samples were compared with those from standards, and the amount of the experimental PAI-1 mRNA was thereby determined. Each PAI-1 mRNA sample was normalized with respect to 18S rRNA. Thus, counts per minute for PAI-1 mRNA was divided by counts per minute for the 18S rRNA of the sample, multiplied by the mean counts per minute of all 18S rRNA samples for the experimental group, and divided by 5 μg to determine picograms per microgram of total RNA. Data were expressed as picograms of PAI-1 mRNA per microgram of total RNA (means±SD for three samples in each case).
For Northern blot analysis, glyoxal-treated RNA (5 μg) was electrophoresed on a 1% agarose gel and subjected to capillary transfer to a Zetaprobe membrane. Hybridizations were performed as described for slot blots. The molecular sizes of PAI-1 mRNA and 18S rRNA were determined with an RNA calibration mixture.
Results are expressed as mean±SD. Significance (P<.0001) was determined by one-way and one-way two-tailed ANOVAs (Graph Pad Software Co).
Effect of RTPA on the Synthesis and Secretion of Fibrinolytic System Components
Under basal conditions HCMEC secrete 3.0±1.0 ng TPA per 1×106 cells, 0.8±0.3 ng UPA per 1×106 cells, and 800±360 ng PAI-1 per 1×106 cells into serum-free conditioned medium over 24 hours.16 Exposure of HCMEC to selected concentrations of RTPA (from 0 to 1000 ng/mL) for 24 hours in serum-free medium evoked marked, concentration-dependent decreases in secretion of PAI-1 (Fig 1⇓). The concentrations of RTPA were nontoxic to cerebral microvascular endothelium in terms of retained capacity of the cells to exclude trypan blue. A 23±2% decrease in accumulation of PAI-1 in conditioned medium was seen with exposure of the cells to 5 ng/mL RTPA, 30±8% with 10 ng/mL RTPA, and 43±14% with 50 ng/mL RTPA. Maximal depression (60±15%) was seen with 100 ng/mL RTPA.
When the cells were evaluated with respect to secretion of UPA (Fig 1⇑), no significant change was evident. Both control and RTPA-exposed cells secreted similar amounts of UPA (<1 to 2 ng UPA per 1×106 cells) over 24 hours, implying that the decreased elaboration of PAI-1 in HCMEC exposed to RTPA is specific.
When HCMEC lysates were assayed for PAI-1 and UPA, neither was detectable in appreciable quantities. Depression of PAI-1 secretion into media occurred regardless of whether the cells had been exposed to RTPA in serum-free or serum-containing medium.
As judged from results with conditioned medium of HCMEC harvested 0, 2, 4, 8, 16, or 24 hours after exposure of the cells to 100 ng/mL RTPA, control and RTPA-exposed cells elaborated comparable amounts of PAI-1 in the first 4 hours, 260±50 and 296±50 ng PAI-1, respectively. Subsequently, however, elaboration of PAI-1 by control cells increased (over the entire 32-hour interval in each experiment) to 1300±33 ng PAI-1 per 1×106 cells in 32 hours. By contrast, elaboration was significantly less after incubations of 8, 16, 24, and 32 hours in media from cells exposed to RTPA (Fig 2⇓).
Incubation of control and RTPA-exposed HCMEC with 5 and 10 μg/mL of cycloheximide used to inhibit protein synthesis led to a marked reduction in PAI-1 antigen accumulation in media over 24 hours (Table 1⇓). The concentrations of cycloheximide had no adverse effects on cell viability as judged from persistent exclusion of trypan blue.
To determine whether the RTPA vehicle influenced results, we exposed HCMEC to vehicle alone for 24 hours. Vehicle had no effect on elaboration of PAI-1 (560±24 [vehicle] compared with 599±63 ng PAI-1 per 1×106 cells [controls]). Results differed markedly from those with cells exposed to 100 ng/mL RTPA (225±22 ng PAI-1 per 1×106 cells over the same interval).
To determine whether the depression in PAI-1 elaboration seen with HCMEC exposed to RTPA was dependent on an intact catalytically active site in the RTPA, HCMEC were exposed to FPRCK-modified RTPA for 24 hours. Results showed that FPRCK-treated RTPA elicited nearly the same magnitude of depression of PAI-1 elaboration as did the nonaltered RTPA, despite indicating that the PAI-1 response was not dependent on retention of a catalytically active site in the RTPA molecule. Control cells were found to secrete 505±17.5 ng PAI-1 per 1×106 cells, whereas those exposed to 100 ng/mL RTPA and FPRCK-treated RTPA secreted 165±20 and 289±105 ng PAI-1 per 1×106 cells, respectively.
To define the specificity of the PAI-1 response in brain endothelial cells exposed to RTPA, we compared responses in three types of human brain endothelial cells (cerebral microvascular, middle meningeal arterial, and pial arterial), human cerebral pericytes, and HUVEC. As Table 2⇓ shows, exposure of all three types of human cerebral endothelial cells to RTPA decreased elaboration of PAI-1. A 30±8% reduction was seen with cerebral microvascular endothelium with 10 ng/mL RTPA; a 60±1.5% reduction was seen with 100 ng. With middle meningeal arterial cells, reductions of 40±1.3% with 10 ng and 52±3% with 100 ng RTPA were seen. With pial arterial endothelium, reductions were 47±6% and 58±12% with 10 and 100 ng, respectively. In contrast, the elaboration of PAI-1 by cerebral pericytes and HUVEC was not significantly reduced when the cells were exposed to RTPA.
To define the specificity of PAI-1 response in HCMEC with respect to RTPA, we exposed cells to RTPA and other plasminogen activators, including SCUPA and SK/Plg, and to the serine proteases human coagulation factor Xa and human α-thrombin. RTPA (100 ng/mL; 1.43 nmol/L), SCUPA (100 ng/mL; 1.85 nmol/L), SK/Plg (36 IU/2 μmol/L; 2.13/2 μmol/L), coagulation factor Xa (1 to 10 nmol/L), or α-thrombin (1 to 10 nmol/L) was added to the media and incubated with the cells for 24 hours. Results (Fig 3⇓) showed that not only RTPA but also SCUPA and SK/Plg decreased PAI-1 elaboration by HCMEC. The magnitude of decrease was 33±8% and 35±4% with SCUPA and SK/Plg, respectively, compared with 59±8% with RTPA. Coagulation factor Xa had no effect; α-thrombin (10 nmol/L) increased in PAI-1 accumulation by 2.7±0.5-fold.
To delineate possible mechanism(s) by which RTPA depressed elaboration of PAI-1 by HCMEC, we examined plasmin as a potential mediator. HCMEC were exposed to 100 ng/mL RTPA with or without 2 μmol/L aprotinin (trasylol), a relatively nonspecific proteinase inhibitor with a high affinity for plasmin. As Table 3⇓ shows, aprotinin did not affect PAI-1 elaboration by HCMEC. However, aprotinin plus RTPA led to an additional diminution of PAI-1 elaboration over 24 hours compared with that induced by RTPA. Thus, RTPA elicited a 78±2.3% decrease, whereas the combination of RTPA and aprotinin elicited a 94±1.4% decrease in PAI-1 accumulation. These data suggest that the effect induced by RTPA is not mediated by plasmin. It is possible that aprotinin may have protected the exogenous RTPA from proteolysis, thus potentiating the depression of PAI-1 elaboration.
Effects of RTPA on Steady-State Levels of PAI-1 mRNA in HCMEC
Total cellular RNA was extracted from HCMEC that had been exposed to 100 ng RTPA for 24 hours. Northern blot analysis of HCMEC RNA was obtained 2, 8, and 24 hours after exposure of the cells to RTPA. An 8-hour exposure to RTPA demonstrated marked decreases in the content of both species of PAI-1 mRNA (3.4 and 2.4 kb). Hybridization with an 18S rRNA probe indicated that equivalent amounts of RNA had been applied to the gel, that the 18S rRNA levels had not been altered by exposure of the cells to RTPA, and that degradation of RNA by ribonucleases had not been induced ex vivo (data not shown).
Slot blot analysis of the same samples (Fig 4⇓) also demonstrated marked depression of PAI-1 mRNA. As Table 4⇓ shows, PAI-1 mRNA was decreased from 176±20 pg/μg total RNA in controls to 43.2±2.2 pg/μg in cells exposed to RTPA for 8 hours.
The results of this study indicate that HCMEC exhibit a response that differs from that of endothelial cells from other sites when exposed to plasminogen activators, including RTPA, SCUPA, and SK/Plg. In contrast to the typical endothelial cell response, ie, increased or sustained secretion of PAI-1 into conditioned media, HCMEC exhibit paradoxical and counterintuitive decreased elaboration of the inhibitor. Elaboration of inhibitors of serine proteinases by endothelial cells may constitute a protective mechanism so that cell surfaces are not subjected to degradation by the proteinases. If so, human cerebral microvasculature would be particularly vulnerable to proteolytic attack by virtue of the suppressed elaboration of PAI-1 by luminal endothelium exposed to high concentrations of plasminogen activators in vivo.
Both PAI-1 elaborated into conditioned media and PAI-1 mRNA harvested from the endothelial cells exhibited the same directional change. Accordingly, the effects induced by exposing the cells to plasminogen activators appear to be mediated, at least in part, at the level of transcription or stabilization of mRNA.
The effects of plasminogen activators on PAI-1 elaboration do not appear to be mediated by plasmin. Inhibition of the active site of RTPA did not obliterate the suppressive effect on PAI-1 elaboration induced by exposure of the cells to RTPA. If the effects observed were attributable to RTPA-dependent generation of plasmin in the extracellular milieu, inactivated RTPA would not have induced the effect. Furthermore, exposure of the cells to RTPA in the presence of aprotinin, a powerful inhibitor of plasmin, did not obliterate the suppression of elaboration of PAI-1 but instead augmented it.
The effects observed do not appear to be specific to a particular plasminogen activator. RTPA, SK/Plg, and SCUPA suppressed elaboration of PAI-1. The effects were at least relatively specific, however, in that other proteinases, such as thrombin and coagulation factor Xa, did not suppress elaboration of PAI-1.
Coronary thrombolysis is the primary therapy for acute myocardial infarction attributable to coronary thrombosis in most settings. Its success is unequivocal, but it is compromised by a modest but definite risk of ICB. Factors responsible for plasminogen activator–associated ICB are obscure, as noted recently in a review of untoward incidents of ICB in some clinical trials.1 ICB is not associated with fibrinolysis in circulating blood per se, as is evident from the lack of induction of brain hemorrhage in experimental animals or human subjects given cobra venom in doses sufficient to almost completely deplete the circulating blood of fibrinogen. With respect to cobra venom, information is available from only a limited number of patients who have been treated. Thus, the nonoccurrence of ICB may be a chance finding. Of note, however, is the fact that administration of massive doses of plasminogen activators to normal laboratory animals fails to induce brain hemorrhage.
As reviewed recently, several characteristics of ICB in patients treated with RTPA merit particular consideration.1 Generally, episodes occur relatively late after administration of the plasminogen activator, often when plasma clearance of the activator and its metabolic products is complete. The incidence of ICB is elevated in subsets of patients in whom occult cerebral vascular disease probably is present, including those with long-standing or poorly controlled systemic arterial hypertension, those of advanced age, and those whose brains exhibit at autopsy microvascular deposition of the amyloid consistent with β-amyloid pleated protein. Patients who have experienced previous cerebrovascular accidents, even many years before administration of a plasminogen activator, appear to be at increased risk for ICB, consistent with the presence of underlying vascular derangements.
If, in fact, ICB associated with the use of plasminogen activators is dependent on occult cerebral vascular disease, its cause may be susceptibility of the diseased vessels to direct proteolytic injury. The extracellular matrix of blood vessels is richly endowed with plasminogen and zymogens, including collagenase and stromelysin, which exhibit metalloproteinase activity when activated by plasmin. Conversion of plasminogen to plasmin within the matrix of vessel walls also can facilitate the action of elastase. It is reasonable to assume that vulnerable cerebral vessels will exhibit deleterious effects in response to the formation of extensive amounts of plasmin within their walls, ie, susceptibility to injury underlying hemorrhage into the brain.
Our results implicate paradoxical and specific properties of brain microvascular endothelium in the etiology of ICB associated with the use of plasminogen activators. The suppression of elaboration of PAI-1 in endothelium exposed to high concentrations of administered plasminogen activators on vascular luminal surfaces may render the vessels particularly susceptible to injury mediated by activation of intramural plasminogen by the exogenous plasminogen activators in the face of its locally diminished inhibition. Thus, one potential therapeutic target for reduction of the risk of ICB in patients treated with plasminogen activators is interdiction of suppression of elaboration of PAI-1 by brain endothelial cells exposed to the activators. Accordingly, further delineation of the intracellular signaling mechanisms underlying suppression of elaboration of PAI-1 by human brain endothelial cells exposed to plasminogen activators offers promise for identification of intracellular moieties that may be amenable to pharmacological modification, thereby ultimately diminishing the risk of ICB.
Selected Abbreviations and Acronyms
|cpm||=||counts per minute|
|FPRCK||=||d-phenylalanyl-l-propyl-l-arginine chloromethyl ketone|
|HCMEC||=||human cerebral microvascular endothelial cells|
|HUVEC||=||human umbilical vein endothelial cells|
|PAI-1||=||plasminogen activator inhibitor type-1|
|RTPA||=||recombinant tissue-type plasminogen activator|
|SCUPA||=||single-chain urokinase plasminogen activator|
|TPA||=||tissue-type plasminogen activator|
|UPA||=||urokinase-type plasminogen activator|
This work was supported in part by PHS grant NS-3032403 and the Collen Foundation. We thank Dr Thomas Orfeo for preparation of purified human plasminogen and plasmin and further purification of FPRCK-RTPA and the University of Vermont Department of Obstetrics and Gynecology for procuring the umbilical cords used for isolation of HUVEC. We also thank Lori Dales for secretarial assistance.
Presented in part at the 68th Scientific Sessions of the American Heart Association, Anaheim, Calif, November 12-16, 1995, and previously published in abstract form (Circulation. 1995;92[suppl I]:I-355).
- Received December 4, 1995.
- Revision received February 6, 1996.
- Accepted February 20, 1996.
- Copyright © 1996 by American Heart Association
Sobel BE. Intracranial bleeding, fibrinolysis, and anticoagulation: causal connections and clinical implications. Circulation. 1994;90:2147-2152.
Bovill EG, Terrin ML, Stump DC, Berke AD, Frederick M, Collen D, Feit F, Gore JM, Hillis LD, Lambrew CT, Leiboff R, Mann KG, Markis JE, Pratt CM, Sharkey SW, Sopko G, Tracy RP, Chesebro J, for the TIMI Investigators. Hemorrhagic events during therapy with recombinant tissue-type plasminogen activator, heparin and aspirin for acute myocardial infarction: results of the Thrombolysis in Myocardial Infarction (TIMI), Phase II Trial. Ann Intern Med. 1991;115:256-265.
Sobel BE. The structure of cardiologic revolutions. Circulation. 1993;87:2047-2054.
Shatos MA, Orfeo T, Doherty JM, Penar PL, Collen D, Mann KG. α-Thrombin stimulates urokinase production and DNA synthesis in cultured human cerebral microvascular endothelial cells. Arterioscler Thromb Vasc Biol. 1995;15:903-911.
Weksler BB, Ley CW, Jaffe EA. Stimulation of endothelial cell prostacyclin production by thrombin, trypsin and the ionophore A23187. J Clin Invest. 1978;62:923-938.
D'Amore PA, Piper HM, eds. Cell Culture Techniques in Heart and Vessel Research. Berlin, Germany: Springer-Verlag; 1990:299-314.
Bouchard BA, Shatos MA, Tracy PB. Human brain pericytes differentially regulate the expression of procoagulant enzyme complexes comprising the extrinsic pathway of blood coagulation. Arterioscler Thromb Vasc Biol. In press.
Shatos MA, Doherty JM, Stump DC, Thompson EA, Collen D. Oxygen radicals generated during anoxia followed by reoxygenation reduce the synthesis of tissue-type plasminogen activator and plasminogen activator inhibitor-1 in human endothelial cell culture. J Biol Chem. 1990;265:20443-20448.
Declerck PJ, Alessi MC, Verstreken M, Kruithof E, Juhan-Vague I, Collen D. Measurement of plasminogen activator inhibitor-1 in biologic fluids with a murine monoclonal antibody-based enzyme-linked immunosorbent assay. Blood. 1988;71:220-225.
Melton DA, Krieg PA, Rebagliati MR, Maneatis T, Zinn K, Green MR. Efficient in vitro synthesis of biologically active RNA and RNA hybridization probes from plasmids containing a bacteriophage SP 6 promoter. Nucleic Acids Res. 1984;12:7035-7055.