INTRODUCTION

Tissue plasminogen activator (t-PA)¹ is a serine protease playing the dominant role in elimination of fibrin from the vasculature by activating the circulating zymogen, plasminogen, to the primary fibrinolytic enzyme, plasmin (1, 2). The major source of circulating t-PA under basal conditions is thought to be the endothelial cell (3, 4). It is generally believed that t-PA follows the constitutive secretory pathway (5). However, t-PA can be released into the circulation within minutes in response to distinct types of stimulation (6-8). The rapidity of this response suggests secretion from stored pools rather than de novo synthesis (9). However, the mechanisms by which this response occurs in vivo and the cellular sources are not known.

Specific stimuli have been identified that induce the acute release of t-PA. Included among these stimuli are exercise (10), mental stress (11, 12), electroconvulsive therapy (13), and surgery (11). Of note, these types of stimuli also activate the sympathoadrenal system causing exocytotic release of amines and proteins from catecholamine storage vesicles of the adrenal medulla and sympathetic neurons (14). In addition, t-PA is present in neuroendocrine tissues (15-19), including the adrenal medulla (20). Therefore, we tested the hypothesis that t-PA is packaged in and released directly from catecholamine storage vesicles, by investigating t-PA expression, subcellular localization, and secretagogue-mediated t-PA release from several chromaffin cell sources. These included rat PC-12 cells, a well-established chromaffin cell line with abundant catecholamine storage vesicles (21), as well as primary bovine adrenal chromaffin cells, and human pheochromocytoma, a catecholamine-producing tumor of the adrenal medulla, and hence a source of human chromaffin cells. Our results demonstrate that t-PA is expressed in chromaffin cells where it is targeted to the regulated pathway of secretion (into catecholamine storage vesicles) and is co-released with catecholamines by chromaffin cell stimulation. These results suggest that catecholamine storage vesicles may serve as a reservoir, and sympathoadrenal activation may be an important physiologic mechanism, for the rapid release of t-PA.

EXPERIMENTAL PROCEDURES

PC-12 cells (21) were cultured in Dulbecco's modified Eagle's medium containing 5% fetal calf serum, 10% horse serum, 100 units/ml penicillin, and 100 µg/ml streptomycin.

Bovine chromaffin cells were isolated from bovine adrenal glands as described (22). The cells were cultured in minimal essential media containing 1% non-essential amino acids, 1% L-glutamine, 10% fetal calf serum, 1% amphotericin B, 100 units/ml penicillin, and 100 µg/ml streptomycin for 3-5 days prior to assay.

Human umbilical vein endothelial cells were isolated from human umbilical veins as described (23).

Human chromaffin granules were isolated from fresh pheochromocytoma tissue as described (24). Briefly, the tissue was minced and homogenized in isotonic (0.3 M) sucrose. Nuclei were sedimented by centrifugation at 1,000 × g for 10 min. The supernatant was removed and centrifuged at 10,000 × g for 20 min. The pellet was centrifuged over a sucrose step gradient (0.3 M, 1.6 M) at 100,000 × g for 60 min to obtain a pure granule pellet. The pure granules were lysed by freeze-thawing and centrifuged at 100,000 × g for 60 min to sediment granule membranes. Chromaffin vesicle lysates were obtained in the supernatant.

Poly(A) RNA was isolated from cells and tissues using guanidinium isothiocyanate and centrifugation through 5.7 M cesium chloride (25) followed by poly(A) selection on an oligo(dT) column.

For Northern blot analysis, 5 µg of poly(A)-enriched RNA were electrophoresed through 1% agarose, 0.66 M formaldehyde (26) and transferred to nylon filters. Hybridization was performed using a ³²P-labeled 1.9-kb human t-PA cDNA KpnI-XbaI fragment (random primer labeled) as probe at 42 °C in 0.02 M Tris-Cl, pH 7.4, 5 × SSC, 5 × Denhardt's solution, 0.5% SDS, 50% formamide, and 100 µg/ml salmon sperm DNA (Sigma). Washes were performed using the following solutions and conditions: 2 × SSC, 0.5% SDS at 22 °C for 15 min twice; 0.1 × SSC, 0.1% SDS at 50 °C for 45 min twice; 0.1 × SSC at 22 °C for 15 min twice.

Metabolic labeling and immunoprecipitation were performed as described previously (27). Briefly, PC-12 cells were starved in methionine-free medium for 1 h. Then PC-12 cell culture medium containing 50 µCi/plate of [³⁵S]methionine (ICN Biochemicals, Inc., Costa Mesa, CA) was incubated with the cells for 4 h at 37 °C. Cells were washed with 10 mM sodium phosphate, 150 mM NaCl, pH 7.2 (PBS) followed by lysis in 0.1 M NaCl, 0.1 M Tris-Cl, pH 8.0, 10 mM EDTA, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride (Sigma), and 50 kallikrein inhibitor units/ml aprotinin (Calbiochem). Cell debris was removed by centrifugation, and 75 µl of each cell supernatant were incubated with 2 µl of normal rabbit serum for 1 h at 4 °C, followed by 10 µl of 50% protein A-Sepharose (Pharmacia Biotech Inc.) for 30 min at 4 °C on a shaker, to reduce nonspecific binding. The mixtures were centrifuged and the supernatants were incubated with 10 µg/ml of an IgG fraction of either normal rabbit serum or rabbit anti-t-PA serum (28) for 18 h at 4 °C and then incubated with 20 µl of 50% protein A-Sepharose for 2 h at 4 °C followed by centrifugation. The pellets were washed, boiled for 5 min, and electrophoresed on 8% sodium dodecyl sulfate-polyacrylamide gels. The gels were soaked in the enhancing solution, Resolution (EM Corp., Chestnut Hill, MA), for 45 min, placed in 5% cold glycerol for 45 min, dried, and exposed at 80 °C to Kodak X-OMAT AR film (Eastman Kodak Co.).

PC-12 cells were labeled for 2 h with 1 µCi/ml [³H]norepinephrine (Amersham Corp.) in PC-12 medium, washed twice in buffer (10 mM HEPES, pH 7.4, 1 mM ascorbic acid, 150 mM NaCl, 2 mM CaCl₂, 5 mM KCl), homogenized in 3 ml of 0.3 M sucrose, 10 mM HEPES, pH 7, layered over a 19-ml continuous sucrose density gradient (0.3-2.5 M), and centrifuged at 100,000 × g for 60 min at 2 °C as described (27, 29-32). Fractions (1 ml) were collected and assayed for [³H]norepinephrine by liquid scintillation counting, sucrose concentration by refractometry, and t-PA using an ELISA kit (Corvas Biopharmaceuticals, San Diego, CA) according to the manufacturer's instructions. (The assay kit detects both free t-PA and t-PA in complex with inhibitors and has a sensitivity to 0.1 ng/ml.)

PC-12 cells and bovine adrenal chromaffin cells were labeled for 2 h with [³H]norepinephrine at 1 µCi/ml in cell culture medium, washed twice with release buffer (10 mM HEPES, pH 7, 150 mM NaCl, 5 mM KCl, 2 mM CaCl₂), and incubated at 37 °C for 30 min in release buffer with or without the following secretagogues: 60 µM nicotine, 55 mM KCl, or 2 mM BaCl₂ as described (27). (Release buffer for experiments with KCl as secretagogue included NaCl at 100 mM, and release buffer was devoid of CaCl₂ when BaCl₂ was used as secretagogue.) After aspirating the release buffer, cells were harvested and lysed in cell lysis buffer (release buffer containing 0.1% Triton X-100). Release buffer and cell lysates were assayed for [³H]norepinephrine by liquid scintillation counting and t-PA by ELISA. Percent secretion was calculated as the amount in release buffer/total (amount in release buffer + amount in cell lysate), and the results were expressed as fold stimulation compared with basal (unstimulated) values.

Fibrin zymography was performed as described previously (33). Briefly, samples were electrophoresed on 10% SDS-polyacrylamide gels according to the system of Laemmli (34). After removal and neutralization of the SDS by soaking in 2.5% Triton X-100, the gels were incubated on fibrin-agar indicator films containing 2 mg/ml bovine fibrinogen (Calbiochem), 0.5 units/ml human -thrombin, and 0.01 mg/ml human plasminogen (purified as described (35)).

Protein content was determined in chromaffin vesicle lysates and pheochromocytoma homogenates using the bicinchoninic acid protein assay reagent (Pierce) in the microtiter plate protocol according to the manufacturer's instructions.

Catecholamines were measured by differential spectrofluorometry (24, 36).

Human recombinant single chain t-PA (rt-PA, Activase) was from Genentech, Inc. (South San Francisco, CA).

Results are reported as mean ± standard error of the mean. Statistical significance was determined by Student's t test or by analysis of variance followed by Student-Newman-Keuls post-hoc tests for multiple comparisons.

RESULTS

Fig. 1 shows Northern blot results in which we investigated t-PA expression in rat PC-12 cells and in tissue from human pheochromocytoma. Hybridization was performed with a ³²P-labeled 1.9-kb human t-PA cDNA KpnI-XbaI fragment as probe. A prominent 2.7-kb band typical of t-PA was present in PC-12 cells and to a lesser extent in rat brain, rat liver, and rat heart as positive controls (37). t-PA message was not detected in rat skeletal muscle as a negative control. Of note, there was marked expression of t-PA mRNA in human pheochromocytoma.

Next, PC-12 cells were investigated for expression of t-PA antigen. PC-12 cells were metabolically labeled with [³⁵S]methionine, harvested, and incubated with an IgG fraction of either a polyclonal anti-t-PA antiserum (Fig. 2, lane 1) or control normal rabbit serum (Fig. 2, lane 2). Immunoprecipitated proteins were electrophoresed through SDS-polyacrylamide gels and subjected to autoradiography. Two major radiolabeled bands were specifically immunoprecipitated by the anti-t-PA antibodies, one with an apparent molecular weight of 70,000, consistent with the molecular weight of free t-PA. Interestingly, a second band with an apparent molecular weight of 110,000 was present, consistent with t-PA in a complex with inhibitor, plasminogen activator inhibitor 1 (PAI-1), or plasminogen activator inhibitor 2 (PAI-2) (38). The t-PA content (determined by ELISA) of the PC-12 cells was 9.4 ± 0.5 ng/10⁶ cells (n = 6), a value ~2-fold greater than the level we determined for human umbilical vein endothelial cells, 4.8 ± 0.4 ng/10⁶ cells (n = 3).

To explore the subcellular localization of the t-PA expressed in chromaffin cells, we first performed sucrose gradient fractionation of PC-12 cell homogenates. PC-12 cells were labeled with [³H]norepinephrine, homogenized, and layered over a continuous sucrose density gradient (0.3-2.5 M sucrose). Following centrifugation, fractions were collected, and t-PA antigen, [³H]norepinephrine, and sucrose concentration were determined. [³H]Norepinephrine and t-PA antigen were co-localized to the same subcellular fractions with a major peak at 1.4 M sucrose (Fig. 3). The 1.4 M sucrose peak is consistent with the buoyant density which we and others (27, 30, 31) have demonstrated previously for chromaffin granules isolated from PC-12 cells. In addition, some [³H]norepinephrine and t-PA were found at the top (fractions 18-22, ~0.3-0.5 M sucrose) and bottom (fraction 5, ~1.9 M sucrose) of the gradient. These results are similar to those of our own previous sucrose gradient studies (27) as well as to those of other investigators (31, 39). The additional peak at the top of the gradient is consistent with release of granular components from vesicles lysed during the homogenization step (before application of the sample to the gradient) (31, 39). The additional peak toward the bottom of the gradient (at ~1.9 M sucrose) is consistent with vesicles trapped in undisrupted cell debris (39).

We also investigated the presence and subcellular localization of t-PA antigen in homogenates of chromaffin tissue obtained from human pheochromocytoma. We compared t-PA concentration in pheochromocytoma tumor homogenates with t-PA concentration in lysates of catecholamine storage vesicles (Table I). t-PA antigen (by ELISA) was present in human pheochromocytoma and was markedly enriched in the chromaffin vesicle fraction (approximately 30-fold enrichment compared with tumor homogenates). The enrichment in t-PA antigen paralleled the enrichment in catecholamines, suggesting that both are localized in the same subcellular compartment, the storage vesicle. Whereas the t-PA was enriched about 30-fold in the vesicle fraction over homogenates, the catecholamines were enriched about 15-fold. This potential difference in enrichment possibly could have arisen as a result of differential leakage of vesicle constituents during the tissue preparation and homogenization steps, with leakage of catecholamines occurring to a greater extent than leakage of larger vesicle protein components (including t-PA). Trapping of catecholamines within the granule requires the activity of the vesicular monoamine transporter (40), the activity of which may be impaired during the tissue preparation and homogenization steps.

Table I. t-PA and catecholamines in human pheochromocytoma: enrichment in chromaffin vesicle fractions Fraction t-PA Catecholamines ng/mg protein µg/mg protein Tumor homogenates (n = 9) 0.37  ± 0.12 14.80  ± 5.04 Chromaffin vesicle lysates (n = 9) 10.70  ± 4.73a 220.58  ± 73.46a a  p < 0.05, compared with tumor homogenates.

The subcellular localization of t-PA was independently verified in functional secretagogue release studies. In these experiments, t-PA antigen and norepinephrine were measured in both releasates and whole cell lysates from [³H]norepinephrine-loaded PC-12 cells under basal conditions and after 30 min of exposure to several well established chromaffin cell secretagogues: 60 µM nicotine (which acts through nicotinic cholinergic receptors), 55 mM KCl (a membrane depolarizing agent), or 2 mM BaCl₂ (a calcium agonist (27)). Significant increases in t-PA secretion were observed in response to each of these secretagogues, approximately 4-fold for nicotine, 5-fold for KCl, and 12-fold for BaCl₂ (Fig. 4). Furthermore, secretion of t-PA paralleled secretion of norepinephrine, consistent with release from the same subcellular pool.

We performed similar functional secretagogue release studies using primary cultures of bovine adrenal chromaffin cells as a source of nontransformed chromaffin cells. t-PA antigen was present (4.7 ± 0.7 ng/10⁶ cells (n = 8)) and was released from the bovine chromaffin cells in response to the three chromaffin cell secretagogues (nicotine, KCl, and BaCl₂) in these cells as well (Fig. 5). Secretagogue-mediated t-PA release paralleled norepinephrine secretion, consistent with exocytotic release of t-PA from catecholamine storage vesicles in this additional chromaffin cell source.

To confirm the authenticity of t-PA in chromaffin cells, we employed the technique of fibrin zymography for t-PA enzymatic activity. This technique determines apparent molecular weights of plasminogen activator activity within a sample (33). We examined secretagogue-mediated cell releasates from both PC-12 and bovine adrenal chromaffin cells in culture, as well as human pheochromocytoma chromaffin vesicle lysates, for t-PA activity. Zones of lysis indicating plasminogen activator activity were detected in secretagogue-mediated releasates from both PC-12 and bovine adrenal chromaffin cells (Fig. 6, lanes 1 and 3) but not in conditioned media of cells incubated with buffer alone (lanes 2 and 4). This plasminogen activator migrated with a M_r (app) of 70,000, consistent with that of the free t-PA standard (lane 6). In addition, plasminogen activator activity, migrating with a M_r (app) of 70,000 was observed in the lysates from pheochromocytoma chromaffin vesicles (lane 5). Thus, t-PA activity was present in the secretagogue-responsive (releasable) pool from chromaffin cells, and the molecular weight of the activity in each case suggested authentic t-PA.

Since the results of the immunoprecipitation experiment (Fig. 2) suggested that both t-PA and a complex with an electrophoretic mobility consistent with a t-PA·plasminogen activator inhibitor complex were synthesized by chromaffin cells, we examined whether the inability to detect t-PA·plasminogen activator inhibitor complexes in the releasates by zymography was due to greater sensitivity of the immunoprecipitation method in detecting these complexes or the absence of t-PA inhibitors from the storage granules. We subjected human chromaffin vesicle lysates to fibrin zymography using a prolonged exposure which yielded a large zone of lysis corresponding to free t-PA that was close to the maximum zone of lysis that could be resolved on the gel (Fig. 7). Under these conditions, a small zone of lysis was detected which migrated with an M_r (app) corresponding to the presence of t-PA·inhibitor complexes. Thus, these data are consistent with the presence of t-PA·inhibitor complexes in addition to free t-PA within the storage granules.

DISCUSSION

The present study demonstrates that t-PA is expressed in and is targeted to the regulated secretory pathway (41) in chromaffin cells. This was demonstrated for chromaffin cells from several sources. Moreover, these data provide the identification of a specific subcellular compartment (the catecholamine storage vesicle) into which the t-PA molecule is sorted and from which t-PA is released by specific stimuli. In addition, these results suggest that catecholamine storage vesicles are a previously unrecognized reservoir and that sympathoadrenal activation, with resultant exocytotic release from these organelles, may be an important physiologic mechanism for the rapid release of t-PA.

We examined the expression of t-PA in three different chromaffin cell sources, including the rat PC-12 cell line, primary bovine chromaffin cells, and human pheochromocytoma. t-PA was expressed in these tissues as assessed by Northern blotting, immunoprecipitation of cells labeled with [³⁵S]methionine, by a specific ELISA for t-PA, and by fibrin zymography. Previous reports have shown expression of t-PA in PC-12 cells (42-44).

Endothelial cells are a well-recognized source of circulating t-PA under basal conditions (3, 4). We compared the t-PA content of chromaffin cells with that of human umbilical vein endothelial cells. The t-PA content of chromaffin cells was quantitatively equal to (4.7 ± 0.7 ng/10⁶ cells for bovine chromaffin cells) or greater than (9.4 ± 0.5 ng/10⁶ cells for PC-12 cells) the t-PA content of human umbilical vein endothelial cells (4.8 ± 0.4 ng/10⁶ cells). Indeed, the t-PA antigen content of the bovine chromaffin cells may have been underestimated by the ELISA (perhaps due to lower cross-reactivity of the anti-human t-PA antiserum with bovine than with human t-PA) since bovine samples yielded much larger zones of lysis on fibrin zymography than human samples with the same apparent t-PA antigen concentration as determined by the ELISA (data not shown).

Secretory proteins expressed in neuroendocrine cells are targeted into one of two pathways, constitutive and regulated (41, 45). Proteins entering the regulated pathway are concentrated and stored in vesicles and subsequently released upon stimulation by a secretagogue. In the constitutive pathway, newly synthesized protein is not stored but is transported directly to the cell surface and secreted even in the absence of any extracellular signal. Catecholamine storage vesicles within chromaffin cells are prototype examples of regulated secretory vesicles (14, 27, 39, 46, 47).

Therefore, to investigate whether chromaffin cell t-PA is sorted into catecholamine storage vesicles, we examined the subcellular localization of chromaffin cell t-PA, using both sucrose gradient fractionation and functional secretagogue-mediated release studies.

Sucrose gradient studies demonstrated co-localization of PC-12 cell t-PA and catecholamines to the same subcellular fraction at 1.4 M sucrose. The peak at 1.4 M is consistent with the buoyant density which we and others (27, 30, 31) have demonstrated previously for chromaffin granules isolated from PC-12 cells. In addition, lysates of human chromaffin granules, isolated from pheochromocytoma tissue on a sucrose gradient, were enriched in t-PA antigen compared with the whole tumor homogenate. This enrichment in t-PA antigen paralleled the enrichment in catecholamine content, suggesting co-localization of t-PA and catecholamines to the same subcellular compartment, the catecholamine storage vesicle.

Interpretation of sucrose gradient results may be limited by the potential for co-purification of heterogeneous organelles to the same fraction (30-32). Hence, we cannot exclude the possibility that part of the chromaffin cell t-PA may reside in a different subcellular compartment of similar buoyant density to chromaffin granules. Nonetheless, previous sucrose gradient fractionation studies by our group have yielded relatively clear separation of markers for chromaffin granules, lysosomes, and mitochondria, with buoyant densities for these organelles as follows: chromaffin granules > lysosomes > mitochondria (48). Furthermore, the sucrose gradient results coupled with results of functional secretagogue release studies suggest co-localization of t-PA and catecholamines to the catecholamine storage vesicle.

We used several well-established chromaffin cell exocytotic secretagogues to assess the subcellular location of t-PA in PC-12 cells as well as in bovine adrenal chromaffin cells, as a source of nontransformed chromaffin cells. t-PA was released from both cell types in response to either 60 µM nicotine (which acts through nicotinic cholinergic receptors), 55 mM KCl (a membrane depolarizing agent), or 2 mM BaCl₂ (a calcium agonist (27)). Moreover, release of t-PA from these chromaffin cell sources was in parallel with that of catecholamines, consistent with release from the same subcellular pool. The values for percent release during secretagogue-mediated stimulation were, in general, less for t-PA compared with those for [³H]norepinephrine. Previously, we noted differences between release of a chromaffin granule protein, chromogranin A, and release of [³H]norepinephrine in response to secretagogues (27). Release of endogenous chromaffin vesicle proteins after their synthesis requires trafficking through the endoplasmic reticulum and Golgi stacks before final localization in the dense core secretory vesicles. Thus, a percentage of the measured total cellular protein may still be in route to the secretory vesicle and, therefore, unavailable for exocytotic release. In contrast, since the [³H]norepinephrine is supplied to the cells as an exogenous label and does not require synthesis or routing through the endoplasmic reticulum or Golgi, a greater percentage of cell [³H]norepinephrine may be immediately available for release in response to secretagogue stimulation. In addition, recent studies have shown that both the chromogranins (49) and t-PA (50) bind to chromaffin cell membranes in a saturable and specific manner, so that some of the chromogranins and t-PA initially secreted upon stimulation into the release medium may be rapidly bound to the chromaffin cell membrane and therefore not measured in the secreted fraction. Another possible explanation is the existence (in addition to the regulated vesicles containing both [³H]norepinephrine and t-PA) of a population of recycled vesicles (51) which, having already released their t-PA, contain only (re-loaded) [³H]norepinephrine. Nonetheless, our secretion studies with a variety of secretagogues (Figs. 4 and 5) showed parallel secretion of [³H]norepinephrine and t-PA such that each increment in [³H]norepinephrine release was associated with a corresponding increase in t-PA release, consistent with co-localization of [³H]norepinephrine and t-PA to the same subcellular pool.

Using fibrin zymography, we confirmed the authenticity of t-PA within the three chromaffin cell sources. Plasminogen activator activity was present and was of the appropriate size for authentic t-PA. In addition, t-PA activity was present in the releasable pool from all three chromaffin cell sources studied. Although a band (M_r (app) ~110,000) consistent with t-PA complexed with inhibitor was detected following immunoprecipitation of [³⁵S]methionine-labeled PC-12 cells, such a complex was not observed initially in cell releasates by fibrin zymography (Fig. 6). This could have been due to either greater sensitivity of the immunoprecipitation method compared with fibrin zymography in detecting these complexes or the absence of t-PA inhibitors from the storage granules. To further investigate this question and to determine whether t-PA·inhibitor complexes could be detected in the storage granules, we subjected human chromaffin vesicle lysates to fibrin zymography using a prolonged exposure (Fig. 7). Under these conditions, a small lytic zone consistent with the M_r (app) of a t-PA·inhibitor complex could be detected in the chromaffin vesicle lysate. Thus, these results suggest the presence of t-PA·inhibitor complexes within the storage granule. The detection of t-PA·inhibitor complexes by fibrin zymography is presumably based on dissociation of some of the complexes during the processing of the zymograms (33) and, therefore, the absolute concentrations of t-PA·inhibitor complexes relative to free t-PA could not be estimated. Nonetheless, the detection of the free t-PA band in the storage granules as well as in releasates from both bovine and PC-12 cells (Fig. 6) indicates that free t-PA is present within the vesicles and is released upon stimulation. Interestingly, transfection of AtT-20 mouse pituitary cells with a cDNA for PAI-1 results in PAI-1 expression and targeting of this inhibitor to the regulated secretory pathway in this cell type (52). Further studies will be necessary to resolve whether t-PA inhibitors are targeted to the regulated pathway of secretion in the chromaffin cell.

Catecholamines (primarily epinephrine) stimulate t-PA release into the circulation, presumably by their actions on vascular endothelium (10). However, during exercise only ~50% of the increase in plasma t-PA concentrations can be attributed to this mechanism (exercise-induced increases in catecholamines causing endothelial cell t-PA release (53)). Moreover, the exercise-induced release of t-PA is not fully abolished by -adrenergic receptor antagonists (54-56). Therefore, release of catecholamine storage vesicle contents during sympathoadrenal activation may enhance the profibrinolytic capability of plasma by the direct release of t-PA and, secondarily, by catecholamine secretion. Further studies will be required to assess the overall importance of chromaffin cell t-PA as a source of t-PA release during stress in vivo.

It has been generally considered that t-PA follows the constitutive secretory pathway (5). The present study identifies the catecholamine storage vesicle as a specific subcellular compartment from which t-PA is released in a regulated fashion. Intriguing data in this regard also are emerging from studies of other cell types. t-PA and parathyroid hormone exhibit parallel calcium-regulated release from 24-h cultures of human parathyroid cells, raising the possibility that t-PA and parathyroid hormone may follow the regulated secretory pathway in parathyroid cells, an additional neuroendocrine cell type (16). Also, a rapid release of t-PA is induced upon stimulation of endothelial cell cultures with products formed during coagulation, thrombin, factor Xa, fibrin, bradykinin, and platelet-activating factor (57), but the identity of a specific subcellular granule that contains t-PA in these cells has not been established (58).

Two possible mechanisms have been proposed for targeting proteins into the regulated pathway of secretion and for sorting regulated secretory proteins away from constitutively secreted proteins at the trans-Golgi cisternae: 1) receptor-mediated protein targeting and 2) selective protein aggregation and condensation. In the first mechanism, correct targeting to the regulated pathway results from specific binding to carrier proteins or receptors in the Golgi membrane (59) which recognize sorting signals on targeted proteins. Alternatively, in the second mechanism, the sorting of regulated secretory products from other soluble (constitutive) proteins in the pathway results from formation of molecular aggregates triggered by conditions in the trans-Golgi region (acidic pH and millimolar concentrations of calcium ions) (60, 61). The targeting of t-PA to catecholamine storage vesicles could perhaps be mediated or facilitated through either of these mechanisms by a specific region or regions within the primary structure of t-PA. Mutagenesis studies have identified domains within the sequences of secretory proteins, for example, for chromogranin A (27) and P-selectin (62), which may mediate targeting to the regulated pathway of secretion. These types of mutagenesis studies may be useful to elucidate the mechanisms responsible for the targeting of t-PA to catecholamine storage vesicles and perhaps to other regulated secretory granules.

In summary, these results demonstrate that t-PA is expressed in chromaffin cells where it is sorted into the regulated secretory pathway (into catecholamine storage vesicles) and is co-released with catecholamines by chromaffin cell stimulation. These results suggest that catecholamine storage vesicles may serve as a reservoir, and sympathoadrenal activation may be an important physiologic mechanism, for the rapid release of t-PA. In addition to providing a potential source of circulating vascular levels of t-PA, the presence of t-PA within catecholamine storage vesicles has potential implications for neuroendocrine function. t-PA is synthesized by neurons in most areas of the brain (18) and may be important in neuronal migration and regeneration (50, 63-66). t-PA appears to play a role in excitotoxin-induced injury to the hippocampus as assessed in mice with t-PA gene inactivation (67). Also, t-PA-deficient mice exhibit differences in synaptic transmission and in long term potentiation (68). Moreover, consistent with our results, administration of KCl into the murine brain in vivo increases t-PA activity within the hippocampus (44). In addition, the widespread distribution of t-PA in neuroendocrine tissue has suggested that t-PA may contribute to prohormone processing (20). Recently, we have obtained evidence that chromaffin cell t-PA can participate in plasmin-dependent processing of chromogranin A (69), the major soluble protein present in chromaffin granules as well as in regulated secretory granules throughout the neuroendocrine system (70), and a prohormone precursor of pancreastatin (71-73) and other bioactive peptides (74), including peptides which modulate catecholamine release (75). The present study, therefore, supports a link between the neuroendocrine and fibrinolytic systems, a link with multiple functional consequences.