Degradation of muscle basement membrane zone by locally generated plasmin

Regulation of proteinases during mouse peri-implantation development: urokinase-type plasminogen activator expression and cross talk with matrix metalloproteinase 9

Trophoblast cells express urokinase-type plasminogen activator (PLAU) and may depend on its activity for endometrial invasion and tissue remodeling during peri-implantation development. However, the developmental regulation, tissue distribution, and function of PLAU are not completely understood. In this study, the expression of PLAU and its regulation by extracellular matrix proteins was examined by RT-PCR, immunocytochemistry, and plasminogen-casein zymography in cultured mouse embryos. There was a progressive increase in Plau mRNA expression in blastocysts cultured on gestation days 4-8. Tissue-type plasminogen activator (55 kDa) and PLAU (a triplet of 40, 37, and 31 kDa) were present in conditioned medium and embryo lysates, and were adsorbed to the culture plate surface. The temporal expression pattern of PLAU, according to semi-quantitative gel zymography, was similar in non-adhering embryos and embryos cultured on fibronectin, laminin, or type IV collagen, although type IV collagen and laminin upregulated Plau mRNA expression. Immunofluorescence revealed PLAU on the surface of the mural trophectoderm and in non-spreading giant trophoblast cells. Exogenous human plasminogen was transformed to plasmin by cultured embryos and activated endogenous matrix metalloproteinase 9 (MMP9). Indeed, the developmental expression profile of MMP9 was similar to that of PLAU. Our data suggest that the intrinsic developmental program predominantly regulates PLAU expression during implantation, and that PLAU could be responsible for activation of MMP9, leading to localized matrix proteolysis as trophoblast invasion commences.

Characterization of matrix metalloproteinases in denervated muscle

In a nerve crush model of denervation, we examined muscle matrix metalloproteinase (MMP) expression, localization and activity. In normal muscle, MMP mRNA levels were low, and immunohistochemically MMPs were distributed around the muscle fibre with MMPs-3, -7 and -9 also staining at the neuromuscular junction. Seven days after nerve crush, muscle MMP immunoreactivity, especially MMP-12 and MMP-14, became irregularly distributed. At 20 days reinnervation of the muscle was observed, and some restitution of the normal pattern of immunoreactivity was noted concomitant with a higher level of MMP mRNA expression. In situ zymography showed that MMP activity was very weak in normal muscle whereas it was increased up to 40 days following denervation. Our results suggest that MMPs in muscle are involved in the tissue changes following denervation. Further experiments are required to test the hypothesis that MMP inhibition may be beneficial in protecting muscle from excessive remodelling following denervation and therefore improve reinnervation.

Plasticity and stabilization of neuromuscular and CNS synapses: interactions between thrombin protease signaling pathways and tissue transglutaminase

The first association of the synapse as a potential site of neurodegenerative disease burden was suggested for Alzheimer's disease (AD) almost 30 years ago. Since then protease:protease inhibitor (P:PI) systems were first linked to functional regulation of synaptogenesis and synapse withdrawal at the neuromuscular junction (NMJ) more than 20 years ago. Confirmatory evidence for the involvement of the synapse, the rate-limiting or key unit in neural function, in AD did not become clear until the beginning of the 1990s. However, over the past 15 years evidence for participation of thrombin, related serine proteases and neural PIs, homologous and even identical to those of the plasma clot cascade, has been mounting. Throughout development a balance between stabilization forces, on the one hand, and breakdown influences, on the other, becomes established at synaptic junctions, just as it does in plasma clot proteins. The formation of protease-resistant cross-links by the transglutaminase (TGase) family of enzymes may add to the stability for this balance. The TGase family includes coagulation factor XIIIA and 8 other different genes, some of which may also influence the persistence of neural connections. Synaptic location of protease-activated, G-protein-coupled receptors (PARs) for thrombin and related proteases, their serpin and Kunitz-type PIs such as protease nexin I (PNI), alpha1-antichymotrypsin (alpha-ACT), and the Kunitz protease inhibitor (KPI)-containing secreted forms of beta-amyloid protein precursor (beta-APP), along with the TGases and their putative substrates, have all been amply documented. These findings strongly add to the conclusion that these molecules participate in the eventual structural stability of synaptic connections, as they do in coagulation cascades, and focus trophic activity on surviving terminals during periods of selective contact elimination. In disease states, this imbalance is likely to be shifted in favor of destabilizing forces: increased and/or altered protease activity, enhanced PAR influence, decreased and/or altered protease inhibitor function, reduction and/or alteration in tTG expression and activity, and alteration in its substrate profile. This imbalance further initiates a cascade of events leading to inappropriate programmed cell death and may well be considered evidence of synaptic apoptosis.

Matrix metalloproteinases MMP-2 and MMP-9 in denervated muscle and injured nerve

Nerve crush or axotomy results in a transient or longterm denervation accompanied by remodelling in nerve, muscle and neuromuscular junctions. These changes include an increased turnover of several extracellular matrix molecules and proliferation of Schwann cells in injured nerves. Given the role of matrix degrading metalloproteinases MMP-2 and MMP-9 (gelatinases-type IV collagenases) in extracellular matrix remodelling, we investigated their regulation and activation in denervated muscles and injured nerves in mice. For this, immunofluorescence using MMP-2 and MMP-9 antibodies was carried concomitantly with gelatin zymography and quantification of gelatinase activity using [3H]-gelatin substrate. Results show that in normal mouse muscles MMP-2 and MMP-9 are localized at the neuromuscular junctions, in Schwann cells and the perineurium of the intramuscular nerves. In denervated mouse muscles, MMP-2 immunolabelling persists at the neuromuscular junctions but decreases in the nerves whereas MMP-9 immunolabelling persists at the neuromuscular junctions but is enhanced in degenerated intramuscular nerves. Denervated muscles did not show any significant change of gelatinolytic activity or expression pattern, while injured nerves exhibited a transient increase of MMP-9 and activation of MMP-2. In conclusion, this study demonstrates that MMP-2 and MMP-9 are expressed at mouse neuromuscular junctions and that their localization and expression pattern appear not to be modified by denervation. Their modulation in injured nerves suggests they are involved in axonal degeneration and regeneration.

Detection of fibronectin expression by human endothelial cells using a enzyme-linked immunosorbent assay (ELISA): enzymatic degradation by activated plasminogen

An enzyme-linked immunosorbent assay (ELISA) has been developed to measure cellular fibronectin (cFN) in association with human umbilical vein endothelial cells (HUVEC) in culture. The expression of a number of functional domains on the cFN molecule was demonstrated using three specific murine monoclonal antibodies. This system was found to be sensitive, detecting as little as 0.156 microg/ml of cFN, and required only 1.3 x 10(5) cells per well confluent cells per experimental condition. This allowed multiple experiments to be performed on one batch of endothelial cells. cFN was detected on both viable and methanol fixed endothelial cells without significant non-specific antibody binding. The utility of this experimental model was studied by exploring the effect of urokinase activated plasminogen, a potent protease, on the expression of cFN and its functional domains.

Alterations in endothelial cell proteinase and inhibitor polarized secretion following treatment with interleukin-1, phorbol ester, and human melanoma cell conditioned medium

Polarized secretion of matrix metalloproteinases and plasminogen activators by monkey aortic endothelial cells was studied in vitro, using transwell inserts. The endothelial cells constitutively expressed matrix metalloproteinase-2, tissue inhibitors of metalloproteinases 1 and 2, urokinase, and tissue plasminogen activator, all with basal preference. Matrix metalloproteinase-9 activity was induced by phorbol 12-myristate 13-acetate (apical), interleukin-1 alpha (basal), and by conditioned medium from DX3 human melanoma cells (basal). The DX3 melanoma conditioned medium also stimulated basal secretion of matrix metalloproteinase-2, urokinase, tissue plasminogen activator, and tissue inhibitors of metalloproteinases. The rise in proteolytic activity in the basal direction was reflected by increased capacity to degrade subendothelial basement membrane type IV collagen, shown immunohistologically, using monkey kidney tissue sections and basement membrane deposited by endothelial cells into the transwell membrane. Thus, IL-1 alpha and DX3 melanoma conditioned medium can stimulate endothelial cells in vitro to concentrate secretion of proteinases spatially onto the underlying basement membrane. We suggest that the stimulation of endothelial cell proteinase activity by tumor cells may facilitate tumor cell extravasation.

Secretion of plasminogen activator and lysosomal enzymes from mouse skeletal muscle: effect of denervation

Levels of hydrolytic enzymes increase in skeletal muscle after denervation and their activities in the extracellular matrix appear to be important for interaction between muscle and nerve. Using enzymatic assays for beta-glucuronidase, beta-galactosidase, and plasminogen activator, we show that secretion of these enzymes from mouse skeletal muscle increases after denervation and that drugs interfering with the secretory pathway or the reuptake of enzymes modulate this release. Thus, brefeldin A inhibited secretion of plasminogen activator activity and mannan increased secreted amounts of beta-glucuronidase, but not of beta-galactosidase, in denervated muscle. In innervated muscle, brefeldin A decreased secreted activity of plasminogen activator, but mannan had no effect on secretion of either beta-glucuronidase or beta-galactosidase. Furthermore, secretion of plasminogen activator was temperature dependent. These observations, together with previous studies, suggest that secretion of hydrolytic enzymes from adult skeletal muscle may be of physiological significance in nerve-muscle communication.

Synaptic transmission blockade increases plasminogen activator activity in mouse skeletal muscle poisoned with botulinum toxin type A

Experimental denervation, either by nerve crush or axotomy, leads to a dramatic increase in muscle plasminogen activator (PA) activity, suggesting a regulation of muscle PA levels by some neural influence (Festoff et al., 1986, J. Cell Biol., 103:1415-1421; Hantaï et al., 1990, Proc. Natl. Acad. Sci. U.S.A., 87:2926-2930). The Botulinum toxin (BoTx) type A is known to selectively interrupt the release of acetylcholine without structurally altering synaptic morphology. In the present study we have used acute BoTx poisoning of hind limb muscles to further explore the neural regulation of muscle PA activities directly after poisoning and during the process of collateral reinnervation. Electromyographic recording and study of ultraterminal sprouting after zinc iodideosmium and silver-cholinesterase staining were used to monitor "denervation" and reinnervation. Muscle choline acetyltransferase activity did not decrease, as is observed after experimental denervation, but in contrast increased and, therefore, reflected the functional integrity of intramuscular nerve endings. Within 2 days of BoTx poisoning, muscle urokinase-PA, and to a lesser extent, tissue-PA activities, rose in muscle extracts as shown by an amidolytic assay and fibrin zymography. When reinnervation occurred, muscle urokinase-PA activity decreased but did not return to baseline levels within the 80 days of our study. These results suggest that cholinergic transmission-regulated events determine activity of muscle PAs and that PAs likely have a role in neuromuscular formation and plasticity.

The insulin-like growth factor signaling system and ALS neurotrophic factor treatment strategies

Because of its multi-faceted potential as a neurotrophic factor, insulin-like growth factor I (IGF-I) has been given to hundreds of ALS patients world-wide. Unlike some patients with post-polio syndrome and fragile elderly males, it is unclear whether any of these patients possess disturbances in IGF signaling. We found that about 25% of ALS patients in a controlled trial of human growth hormone (hGH) had lower or higher than normal IGF-I serum levels. Many ALS patients do have some of the characteristics of type II diabetes mellitus, where IGF-I therapy is also under way. In addition, in type I diabetes significant increase in a circulating molecule that binds IGF-I, IGF-I binding protein 1 (IGFBP-1), occurs along with reduced IGF-I, when neuropathic complications are prominent. We have studied the response of IGFBPs in ALS patients to subcutaneous rhIGF-I and found transient induction of IGFBP-1. Studies related to the IGFBPs have not been done in familial ALS (FALS) patients. However, the gene for another IGFBP, BP-2, co-localizes with the gene for juvenile ALS (ALSJ) on chromosome 2. IGF-I has been given to several models of motor neuron degeneration in the mouse, including motor neuron disease and wobbler, with beneficial effects. However, it is also not known whether any accepted genetic mouse model of motor neuron degeneration possesses any disturbance in the IGF signaling system.

Synthesis and secretion of matrix-degrading metalloproteases by human skeletal muscle satellite cells

The expression of matrix-degrading metalloproteases (MMPs) by human skeletal muscle satellite cells was investigated by zymography of cell culture media and by Northern blot analysis of mRNA prepared from satellite cells. Zymography in gelatin substrate gels revealed that satellite cells constitutively synthesize and secrete 72 kDa gelatinase (MMP-2). In addition, treatment of satellite cell cultures with phorbol ester resulted in an induction of 92 kDa gelatinase (MMP-9) activity. On casein substrate gels, little or no proteolytic activity was detectable in control or phorbol ester treated satellite cell cultures, suggesting that compared to fibroblasts, satellite cells secrete little or no interstitial collagenase (MMP-1) or stromelysin (MMP-3) activity. Northern blotting, however, revealed that there is detectable expression of mRNA transcripts encoding MMP-1 in satellite cell cultures, and that increased accumulation of MMP-1 mRNA transcripts occurs upon treatment of these cells with phorbol ester. In contrast, no constitutive, or induced expression of transcripts encoding MMP-3 was detectable in satellite cells. These findings show that satellite cells can synthesize and secrete selected members of the MMP family and suggest that skeletal muscle cells may participate directly in remodelling of the extracellular matrix during myogenesis and the regeneration of skeletal muscle.

Activation of serpins and their cognate proteases in muscle after crush injury

Direct muscle injury was induced in rats in order to evaluate alterations in the balance of serine proteases and inhibitors (serpins) as a response to tissue damage. It was previously found that certain proteases, specifically urokinase-like plasminogen activator (uPA) and others, required activation in order to effect regeneration. We hypothesized that the magnitude and temporal sequence of serpin activation would follow, pari passu, activation of their cognate proteases. In addition to uPA, tissue PA (tPA) and tissue kallikrein were the proteases studied. The serpins we analyzed were protease nexin I (PNI), PA inhibitor 1 (PAI-1, and the kallikrein-binding protein (KBP). uPA nearly doubled 48 h after injury, while there was no change in amidolytic activity after addition of fibrin monomer as an estimation of tPA activity. Tissue kallikrein activity, barely detectable in normal muscle, slowly increased, nearly tripling at 7 days after injury. Greater magnitude and more rapid changes in muscle serpins occurred over the same post-injury time course. By 24 h PNI increased threefold, while PAI-1 increased more slowly, reaching double the control values by 5 days after injury. Surprisingly, KBP, the serpin-class inhibitor of tissue kallikrein, had the most robust response, increasing tenfold over control 48 h after crush injury of muscle. These results further implicate the serpin:protease balance in tissue injury. Participation of complex receptors, such as the alpha 2-macroglobulin receptor/low density lipoprotein receptor-related protein (LRP), various growth factors, cytokines, and other molecules, in regulating this balance is implicated by these data.

Increase in the number of intrafusal muscle fibres in rat muscles after neonatal motor denervation

Rat muscle spindles disintegrate after total neonatal muscle denervation in which both sensory and motor axons are cut, but develop almost normally during the first three weeks after neonatal de-efferentation, attaining the usual complement of four intrafusal fibres, two bag and two chain fibres. Thereafter additional intrafusal fibres differentiate in most of the de-efferented spindles. We have examined the development of supernumerary fibres in muscle spindles of the rat hind limb muscles one to four months after neonatal de-efferentation using transmission electron microscopy. Additional intrafusal fibres originate both from activated intrafusal myosatellites and by different forms of fibre division. In a sample of 27 muscle spindles examined along the A zone five weeks after de-efferentation, the mean number of intrafusal fibre profiles increased almost two-fold to 7.9 +/- 3.3 (S.D.). Up to 20 intrafusal fibre profiles per spindle cross-section were found in muscles de-efferented for four months. The supernumerary fibres were apparently derived from all three intrafusal fibre types, but predominantly from the chain fibres. It is noteworthy that de-efferentation which causes wasting of extrafusal muscle initiates myogenesis and maintains additional intrafusal fibres within the spindle capsules.

Plasminogen activators in the neuromuscular system of the wobbler mutant mouse

Wobbler, the neurological mutant mouse, carries an autosomal recessive gene (wr) and has been characterized as a model of lower motoneuron disorders with associated muscle atrophy, denervation and reinnervation. During normal murine neuromuscular development a decrease in muscle plasminogen activator (PA) activity accompanies synapse maturation. In contrast, experimental denervation in adult mice leads to an increase in muscle PA activity. The purpose of the present study was to determine the possible involvement of PAs in the denervation/reinnervation phenomena and motoneuron degeneration that characterize the wobbler mutant mouse. We determined the degree of innervation and its characteristics in wobbler mice by measuring choline acetyltransferase (ChAT) activity. We measured ChAT in the spinal cord as well as in two different muscles known to be differentially affected, biceps brachii and gastrocnemius. We found a sharp decrease of ChAT activity in both muscles but not in spinal cord extracts. We estimated the extent of sprouting by the silver/cholinesterase stain. Motoneuron terminal sprouting, not detected in normal animals, was present in 40% of the neuromuscular junctions in wobbler mice. We estimated specific PA activities in biceps brachii and gastrocnemius muscle extracts, as well as spinal cord extracts, using both an amidolytic assay and fibrin zymography. Increased PA, predominantly urokinase-PA (uPA), was observed in wobbler mouse muscle. A greater uPA was detected in biceps brachii muscle than in gastrocnemius muscle, which is less impaired by the mutation. There was no change in spinal cord PA, although tissue type PA (tPA) is the predominant PA type there.(ABSTRACT TRUNCATED AT 250 WORDS)

Plasminogen activators and inhibitors in the neuromuscular system: III. The serpin protease nexin I is synthesized by muscle and localized at neuromuscular synapses

Recent studies suggest that the nature of events leading to the formation, maintenance, and elimination of synapses may be regulated by cascade-type, locally expressed proteases and protease inhibitors acting on adhesive extracellular matrix components. We have identified a molecule in conditioned medium of murine skeletal muscle cells that in molecular weight, target protease inhibition, heparin-binding and cross-reactivity with authenic antisera is similar to the human serine proteinase inhibitor, protease nexin I. Protease nexin I is a 43-50 kDa glycoprotein of the serpin superfamily (arg-serpin class). Purified anti-protease nexin I antibody (anti-47 kDa) stains adult mouse skeletal muscle in discrete foci that precisely superimpose on synaptic neuromuscular junctions. Protease nexin I appears in patches on surfaces of cultured mouse skeletal myotubes, but not on myoblasts. These patches co-localize with acetylcholine receptor clusters and acetylcholinesterase staining during cellular maturation in culture. Evidence that protease nexin I is a synaptic, extracellular antigen is particularly intriguing since it has been shown to be identical, in structure and activity, with a factor released by glial cells, called glia-derived nexin that stimulates mouse neuroblastoma cell neurite outgrowth and inhibits granule cell migration. Protease nexin I inhibits both tumor cell and myoblast plasminogen activator-mediated destruction of extracellular matrix. Thus, such observations as presented in this report provide further evidence for involvement of cascade proteolytic systems, and their post-translational regulation by specific serpins, in the remodeling that occurs in synapse formation and elimination.

Plasminogen activators and their inhibitors in the neuromuscular system: I. Developmental regulation of plasminogen activator isoforms during in vitro myogenesis in two cell lines

Plasminogen activators (PAs), were estimated qualitatively and quantitatively in two different clonal murine skeletal muscle cell lines. Both cell lines produced the two major types of PAs found in mammalian cells, urokinase-type (uPA) and tissue type (tPA). These two lines are models for the study of myogenesis in vitro, but differ in several growth and differentiation characteristics. Because of their possible involvement in these characteristics we assayed the expression of PAs in both cell systems during development in culture. Utilizing fibrin zymography two isoforms of tPA were detected. One co-migrated with human tPA at 75 kd and another may represent a tPA:inhibitor complex at 105 Kd. Several isoenzymes of uPA were detected and these changed depending on whether cell homogenates or conditioned medium was analyzed and whether myogenic cells were at single-cell myoblast or multi-nucleated myotube stage. Species-specific antisera to mouse uPA identified 4 uPA bands in muscle cell medium and 5 in cell layers. Antigenic uPA bands also varied depending on stage of myogenesis. Quantitative amidolytic studies using chromogenic substrates showed that maximal PA activity, both uPA and tPA, occurred at the time of myoblast fusion. Furthermore, uPA activity in membranes increased during myogenesis, while both uPA and tPA in medium decreased after fusion. These studies indicate that muscle PA expression is developmentally regulated and may correlate with growth and differentiation in skeletal muscle.

Plasminogen activators and their inhibitors in the neuromuscular system: II. Serpins and serpin: protease complex receptors increase during in vitro myogenesis

In the course of studies on the regulation of plasminogen activator-mediated extracellular matrix degradation in muscle we found the presence of a factor, a cellular inhibitor of serine proteases having features similar to the serpin protease nexin I (PNI). This factor was present in the medium and at maximum concentration following fusion of skeletal muscle cells in culture. The ability of the PNI homologue in mouse muscle to inhibit ECM degradation by urokinase in myoblast medium was compared to that of human PNI purified from human fibroblasts. Stable (to SDS) 1:1 molar ratio complex formation between PNI and proteases, the proposed means by which these enzymes are regulated and removed, was also detected. Cell surface receptors for protease:PNI complexes, the specific binding sites for inactive complex internalization, were found on multinucleated myotubes, while little or no receptor activity was detected on myoblasts. These data suggest that developmental regulation of a) increased PNI proteolytic inhibitory activity expression and b) the appearance of protease:inhibitor complex receptors on muscle cell surfaces during myogenesis may constitute important regulatory features of muscle surface proteolytic activity. They complement previous studies of proteoglycan metabolism in muscle, which itself contains molecules capable of regulating the activity of myotube surface proteases.

Rapid neural regulation of muscle urokinase-like plasminogen activator as defined by nerve crush

Muscle plasminogen activators (PAs), such as urokinase-like PA and, to a lesser extent, tissue PA, increase dramatically after denervation induced by axotomy. The PA/plasmin system has also been implicated in degradation of specific components of the muscle fiber basement membrane after local activation of plasminogen. These results suggest that neural regulation of muscle extracellular matrix metabolism accompanies or precedes regeneration after injury and is mediated by activation of PAs. In the present study, we have used nerve crush to explore the neural regulation of muscle PA activities directly after subtotal axon interruption and during the process of reinnervation. Muscle contraction after nerve stimulation and estimation of choline acetyltransferase activity were used to monitor reinnervation. Within 24 hr of nerve crush, muscle urokinase (but not tissue PA) activity rose in soluble and membrane-bound muscle fractions, as shown by an amidolytic assay and a fibrin zymography. Membrane-bound activity was 5-fold higher than cytosol activity, but there was no shift between cellular compartments during the time course of denervation. Coincident with the return of choline acetyltransferase activity and muscle contractility, muscle urokinase returned almost to baseline levels. These results show tight regulation of muscle urokinase levels by some neural influence.

Protease nexin I, a serpin, inhibits plasminogen-dependent degradation of muscle extracellular matrix

Clonal, fusing, mouse skeletal muscle cells (C2) were grown to the myotube stage (90% confluence) before they were subjected to isotope-containing serum-free media (3H-proline or 35S-methionine). C2 myotubes secrete and organize a biosynthetically labeled matrix which adheres to the plastic after removal of myotubes with detergent and ammonium hydroxide. When these homotypic-labeled myotube matrices were incubated with myoblast-conditioned media containing high specific activity urokinase-type plasminogen activator, slow, but clearly detectable, release of label occurred. However, degradation of matrix, with solubilization of label, was accelerated sixfold by addition of human plasminogen to diluted myoblast-conditioned media. If protease nexin I, a cellular serine protease inhibitor purified from human fibroblast-conditioned media, was added (0.2 microgram/ml) with plasminogen, inhibition of matrix hydrolysis by 52% occurred. Higher concentrations (0.8 microgram/ml or above) of protease nexin 1 completely inhibited the degradation of extracellular matrix components. A similar protease inhibitor was purified from C2 myotube-conditioned media, and this molecule also inhibited the plasminogen-dependent release of extracellular matrix. We propose that protease nexin 1 inhibits the destruction of myotube matrix by inactivating the plasmin/plasminogen activation system and may be the physiologic regulator of this system during muscle development in vivo.

Decrease in plasminogen activator correlates with synapse elimination during neonatal development of mouse skeletal muscle

Previous studies have implicated proteases, acting extracellularly, in the mechanism of polyneuronal synapse elimination. Most studies have focused on mammalian, especially rodent, skeletal muscle, where retraction of subordinate nerve terminals occurs during a narrow time window 2-3 weeks after birth. To date no specific protease(s) has been detected that (i) coincides in time with maximal synapse elimination and (ii) is known to act extracellularly on specific extracellular matrix proteins. In previous studies of denervation in adult mouse muscle, rapid activation of urokinase-type plasminogen activator, a neutral serine protease, was detected. This enzyme, by activation of plasminogen to plasmin, specifically degrades matrix components such as fibronectin, type IV collagen, and laminin in muscle. We now present evidence for an initial increase and subsequent decrease in soluble urokinase-type PA--and, to a lesser extent, tissue PA--in developing muscle, suggesting postnatal developmental regulation of these enzymes during the period of maximal synapse elimination. Although considerably higher in specific activity, membrane-bound PA activity followed the wave of synapse elimination, possibly indicating a longer half-life of membrane-bound enzyme(s).