Antiproteinase activity of alpha 2-macroglobulin

Blood serum separation by the method of gel filtration on Sephadex G-200 with the subsequent immunochemical determination of the quantitative content of basic proteolysis inhibitors permitted isolating the alpha 2-macroglobulin fraction while alpha 1-antitrypsin and alpha 1-antichymotrypsin separation was a failure. The immunochemical analysis of the antienzymic activity of the isolated inhibitors showed that 32.3 +/- 3.5% of the introduced kallikrein, 18.7 +/- 0.6% of trypsin and 14.4 +/- 4.1% of chymotrypsin were bound in the zone of alpha 2-macroglobulin. The rest of antienzymic activity was localized in the zone of alpha 1-antitrypsin and alpha 1-antichymotrypsin. After a preliminary saturation of blood serum with trypsin in the amount equivalent to its antitryptic capacity (200 micrograms/ml) the ability of alpha 2-macroglobulin to bind kallikrein and chymotrypsin lowers considerably (by 69 and 72%, respectively). In the zone of alpha 1-antitrypsin and alpha 1-antichymotrypsin a decrease in the ability to bind kallikrein and chymotrypsin amounted to 44 and 12% respectively. Thus, alpha 2-macroglobulin being bound with trypsin looses considerably its ability to bind other enzymes.     

Biochemical and enzymatic characterization of human kallikrein 5 (hK5), a novel serine protease potentially involved in cancer progression

Human kallikrein 5 (KLK5) is a member of the human kallikrein gene family of serine proteases. Preliminary results indicate that the protein, hK5, may be a potential serological marker for breast and ovarian cancer. Other studies implicate hK5 with skin desquamation and skin diseases. To gain further insights on hK5 physiological functions, we studied its substrate specificity, the regulation of its activity by various inhibitors, and identified candidate physiological substrates. After producing and purifying recombinant hK5 in yeast, we determined the k(cat)/K(m) ratio of the fluorogenic substrates Gly-Pro-Arg-AMC and Gly-Pro-Lys-AMC, and showed that it has trypsin-like activity with strong preference for Arg over Lys in the P1 position. The serpins alpha(2)-antiplasmin and antithrombin were able to inhibit hK5 with an inhibition constant (k(+2)/K(i)) of 1.0 x 10(-) (2)and 4.2 x 10(-4) m(-1) min(-1), respectively. No inhibition was observed with the serpins alpha(1)-antitrypsin and alpha(1)-antichymotrypsin, although alpha(2)-macroglobulin partially inhibited hK5 at high concentrations. We also demonstrated that hK5 can efficiently digest the extracellular matrix components, collagens type I, II, III, and IV, fibronectin, and laminin. Furthermore, our results suggest that hK5 can potentially release (a) angiostatin 4.5 from plasminogen, (b) "cystatin-like domain 3" from low molecular weight kininogen, and (c) fibrinopeptide B and peptide beta15-42 from the Bbeta chain of fibrinogen. hK5 could also play a role in the regulation of the binding of plasminogen activator inhibitor 1 to vitronectin. Our findings suggest that hK5 may be implicated in tumor progression, particularly in invasion and angiogenesis, and may represent a novel therapeutic target.     

Human kallikrein 4: enzymatic activity, inhibition, and degradation of extracellular matrix proteins

Human kallikrein 4 (hK4) is a member of the expanded family of human kallikreins, a group of 15 secreted proteases. While this protein has been associated with ovarian and prostate cancer prognosis, only limited functional information exists. Therefore, we have undertaken an investigation of its enzymatic properties regarding substrate preference, degradation of extracellular matrix proteins, and its inhibition by various inhibitors. We successfully expressed and purified active recombinant hK4 from supernatants of the Pichia pastoris expression system. This enzyme seems to cleave more efficiently after Arg compared to Lys at the P1 position and exhibits modest specificity for amino acids at positions P2 and P3. hK4 forms complexes with alpha1-antitrypsin, alpha2-antiplasmin and alpha2-macroglobulin. The protease mediates limited degradation of extracellular matrix proteins such as collagen I and IV, and more efficient degradation of the alpha-chain of fibrinogen. The cleavage of extracellular matrix proteins by hK4 suggests that this enzyme may play a role in tissue remodeling and cancer metastasis.     

Interactions in vitro and in vivo between rat serum protease inhibitors and anodal and cathodal rat trypsin and chymotrypsin.

Reaction mixtures of increasing amounts of the pancreatic homologous proteases, anodal and cathodal chymotrypsin and trypsin, respectively, and normal rat serum were analyzed by immunoelectrophoretic methods in order to determine their distribution on serum protease inhibitors. This paper concerns three proteins occurring in normal serum and capable of binding protease viz. alpha1-macroglobulin, alpha1-antitrypsin and alpha1-inhibitor 3. The distribution of the enzymes among these protease inhibitors differed significantly from one protease to another. The distribution of the proteases among the serum protease inhibitors following intravenous injection of 125I-labelled proteases corresponded to that in vitro. Complexes formed with alpha1-macroglobulin and alpha1-inhibitor 3 were quickly eliminated irrespective of the enzyme species used, whereas those formed with alpha1-antitrypsin persisted much longer in the circulation.     

Studies on the influence of Trasylol on the partition of trypsin between the human plasma protease inhibitors in vitro.

The distribution of trypsin between the protease inhibitors of human serum with and without Trasylol was studied in vitro. 1) Trypsin was preferentially bound by alpha2-macroglobulin on addition of small amounts of the enzyme to normal serum in both the presence and absence of Trasylol in a molar concentration equal to that of alpha2-macroglobulin. 2) On saturation of alpha2-macroglobulin, a considerable amount of trypsin was bound by Trasylol even when most of the serum alpha1-antitrypsin was in a free form. 3) In reaction mixtures containing small amounts of trypsin, Trasylol was identified in a free form as well as in complex with trypsin-alpha2-macroglobulin complex and to a limited extent with trypsin. 4) With larger amounts of trypsin, sufficient to saturate alpha2-macroglobulin, increasing amounts of Trasylol were bound to trypsin. The relative amount of Trasylol bound to trypsin-alpha2-macroglobulin complexes was now smaller. This was explained by a higher affinity (or binding rate) of Trasylol for trypsin than for trypsin-alpha2-macroglobulin complexes. 5) Trypsin-Trasylol complexes showed no signs of dissociation after 5 h incubation at 37 degrees C in serum.     

Inhibition profiles of human tissue kallikreins by serine protease inhibitors

Accumulated evidence has shown that human tissue kallikreins (hKs), a group of 15 homologous secreted serine proteases, are novel cancer biomarkers. We report here the inhibition profiles of selected hKs, including hK5, hK7, hK8, hK11, hK12, hK13, and hK14, by several common serine protease inhibitors (serpins) found in plasma. The association constants for the binding of serpins to kallikreins were determined and compared. Protein C inhibitor was found to be the fastest-binding serpin for most of these hKs. alpha2-Antiplasmin, alpha1-antichymotrypsin, and alpha1-antitrypsin also showed rapid inhibition of certain hKs. Kallistatin exhibited fast inhibition only with hK7. Our data demonstrate that these hKs are specifically regulated by certain serpins and their distinct inhibition profiles will be valuable aids in various aspects of kallikrein research.     

The identification of distinctive forms of human alpha 2-macroglobulin by using the numerical relationship between trypsin binding in alpha- and beta-modes.

Interactions between the serine proteinase trypsin and the protein proteinase inhibitors in human blood were expressed in terms of a coupled set of non-linear differential equations, which has been solved for each of 110 samples of serum obtained from colleagues and from a variety of hospital sources. Optimization of nine unknown theoretical parameters and 21 experimental rate measurements of the hydrolytic activity of trypsin in free and bound states after admixture with various amounts of a given serum was achieved by an iterative procedure using initial estimates of the parameters derived from the "four-straight-line" model described in the preceding paper [Topping & Seilman (1979) Biochem. J. 177, 493--499.] Such a procedure yielded the following information for each sample of serum examined: (a) the concentrations of alpha 1-antitrypsin and alpha 2-macroglobulin; (b) the unequivocal assignment of alpha 2-macroglobulin into one of seven categories on the basis of trypsin binding in two kinetically differentiated modes (alpha and beta); (c) the hydrolytic activities of trypsin (versus Bz-Arg-OEt) when bound to alpha 1-antitrypsin, and to alpha 2-macroglobulin in the alpha- and beta-modes. Molecular interpretations of the binding of trypsin to alpha 2-macroglobulin are discussed and the potential clinical value of recognizing the nature of such binding is reported.     

Platelet alpha2-macroglobulin and alpha1-antitrypsin.

Subcellular membrane and granule fractions derived from human platelets contain immunologically identifiable alpha2-macroglobulin and alpha1-antitrypsin. These platelet-derived inhibitors show a reaction of immunologic identity when compared to alpha2-macroglobulin and alpha1-antitrypsin purified from human plasma. Further, the platelet protease inhibitors possessed a similar subunit polypeptide chain structure to their plasma counterparts as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoretic analysis. Studies of the binding of radiolabeled trypsin to the various solubilized platelet subcellular fractions suggest that the granule-associated alpha2-macroglobulin and alpha1-antitrypsin, as well as membrane-associated alpha2-macroglobulin were functionally active. Quantitatively, circulating platelets contain relatively small concentrations of these inhibitors as compared to platelet-associated fibrinogen and factor VIIIAGN. Platelet protease inhibitors may modulate the protease-mediated events involved in the formation of hemostatic plugs and thrombi.     

Suppression of high-affinity ligand binding to the major glutathione S-transferase from Galleria mellonella by physiological concentrations of glutathione.

We have identified a tissue-kallikrein-binding protein in human serum and in the serum-free culture media from human lung fibroblasts (WI-38) and rodent neuroblastoma X glioma hybrid cells (NG108-15). Purified and 125I-labelled tissue kallikrein and human serum form an approximately 92,000-Mr SDS-stable complex. The relative quantity of this complex-formation is measured by densitometric scanning of autoradiograms. Complex-formation between tissue kallikrein and the serum binding protein was time-dependent and detectable after 5 min incubation at 37 degrees C, with half-maximal binding at 28 min. Binding of 125I-kallikrein to kallikrein-binding protein is temperature-dependent and can be inhibited by heparin or excess unlabelled tissue kallikrein but not by plasma kallikrein, collagenase, thrombin, urokinase, alpha 1-antitrypsin or kininogens. The kallikrein-binding protein is acid- and heat-labile, as pretreatment of sera at pH 3.0 or at 60 degrees C for 30 min diminishes complex-formation. However, the formed complexes are stable to acid or 1 M-hydroxylamine treatment and can only be partially dissociated with 10 mM-NaOH. When kallikrein was inhibited by the active-site-labelling reagents phenylmethanesulphonyl fluoride or D-Phe-D-Phe-L-Arg-CH2Cl no complex-formation was observed. An endogenous approximately 92,000-Mr kallikrein-kallikrein-binding protein complex was isolated from normal human serum by using a human tissue kallikrein-agarose affinity column. These complexes were recognized by anti-(human tissue kallikrein) antibodies, but not by anti-alpha 1-antitrypsin serum, in Western-blot analyses. The results show that the kallikrein-binding protein is distinct from alpha 1-antitrypsin and is not identifiable with any of the well-characterized plasma proteinase inhibitors such as alpha 2-macroglobulin, inter-alpha-trypsin inhibitor, C1-inactivator or antithrombin III. The functional role of this kallikrein-binding protein and its impact on kallikrein activity or metabolism in vivo remain to be investigated.     

In vitro interaction of porcine serum and colostrum protease inhibitors with pancreatic trypsin, chymotrypsin and elastase

The partition of 125I-labelled pancreatic trypsin, chymotrypsin and elastase between the inhibitors, alpha 2-macroglobulin f and s, alpha 1-protease inhibitor, alpha 2-antitrypsin, inter-alpha-trypsin inhibitor and the specific sow colostrum protease inhibitor, was studied in vitro by gradually increasing the concentration of these proteases in blood serum from adult and newborn pigs. As revealed by immunoelectrophoresis in combination with autoradiography, differences were noted in the abilities of the various protease inhibitors to interact with and to form complexes with the three proteases, resulting in changes in location, height and numbers of precipitates. Among the serum inhibitors, alpha 2-macroglobulins showed the highest relative affinity to all three proteases, while alpha 1-protease inhibitor showed a high relative affinity only for chymotrypsin. Serum alpha 2-antitrypsin complexed only with trypsin, with a low relative affinity. alpha 2-Antitrypsin also interacted with chymotrypsin and elastase, but without forming complexes. When complexes of sow colostrum protease inhibitor and trypsin were added to the serum from neonatal pigs, these complexes remained stable. The results obtained from these in vitro studies, indicating differences in the relative affinities of the inhibitors to the various proteases, give some information about the role of the inhibitors in vivo, both in adult and in neonatal pigs.     

On the role of the pancreatic secretory trypsin inhibitor as an inactivator of trypsin-alpha 2-macroglobulin complexes in acute pancreatitis

High levels of immunoreactive pancreatic secretory trypsin inhibitor (PSTI) were demonstrated in the serum and peritoneal exudates of patients suffering from acute pancreatitis. Trypsin-like immunoreactivity in these fluids was found in complex with alpha 1-antitrypsin and in complex with alpha 2-macroglobulin and also as a free peak correlating to free trypsin(ogen). No trypsin-PSTI complexes or PSTI were demonstrated in the macroglobulin fraction of the peritoneal exudates. Saturated and partially saturated trypsin-alpha 2-macroglobulin complexes were prepared in vitro. PSTI was able to partially inhibit the BzArgNan-cleaving activity of both types of complexes in a slow dose-dependent non-linear reaction. Equilibrium was reached in each case within 1 h, but total inhibition was not reached even with large amounts of PSTI. Partially saturated trypsin-alpha 2-macroglobulin complexes were inhibited more readily than saturated complexes. The results support the concept of PSTI acting as a strictly local inhibitor of trypsin in compartments lacking plasma protease inhibitors.     

Identification and characterization of trypsin, chymotrypsin and elastase inhibitors in porcine serum.

Characterization of the trypsin-, chymotrypsin- and elastase-inhibiting properties of porcine serum was carried out by gel filtration on Ultrogel, AcA 44, and agarose gel electrophoresis with subsequent processing for protease-inhibiting activity. Moreover, by allowing the fractions obtained from gel filtration to react with antibodies to porcine serum protease inhibitors, the specific inhibiting properties of these inhibitor molecules were identified. At least six protease inhibitors were identified and partially characterized in porcine serum. Two alpha 2 -macroglobulins (alpha 2 Mf and alpha 2 Ms), homologues to human alpha 2 -macroglobulin, with slightly different electrophoretic mobilities, were both found to exhibit trypsin, chymotrypsin and elastase inhibiting activity. Alpha 1 -Protease inhibitor (Mr 51 000), a homologue to human alpha 1 -protease inhibitor (alpha 1 -antitrypsin), also showed trypsin-, chymotrypsin- and elastase-inhibiting properties. Inter-alpha-trypsin inhibitor (Mr 162 000 and 129000), a porcine serum counterpart to human inter-alpha-trypsin inhibitor, showed trypsin- and chymo-trypsin-inhibiting properties. In addition, a specific trypsin inhibitor, alpha 2 -antigrypsin (Mr 58 000), and a specific elastase inhibitor, beta-elastase inhibitor, were characterized in porcine serum, and these seem to have no counterparts in human serum.     

The in vitro interactions of rat pancreatic elastase and normal and inflammatory ray serum.

The partition of labelled rat pancreatic elastase (EC 3.4.21.11) between the different protease inhibitors of rat plasma was studied at different levels of saturation of the inhibitors of rat plasma was studied at different levels of saturation of the inhibitor capacity of plasma with the enzyme. The reaction mixtures were analysed by immunoelectrophoretic methods utilizing specific antisera against the different inhibitors and by gel filtration on Sephadex G-200. Rat serum was shown to contain four elastase binding proteins. alpha 1-antitrypsin, alpha 1-macroglobulin and alpha 2-acute phase protein and alpha 1-inhibitor 3 which exhibits immunologic cross-reaction with human inter-alpha-trypsin inhibitor and is of similar molecular weight. With minute amounts of labelled elastase the partition among the binding protein was alpha 1-macroglobulin 60%, alpha 1-antitrypsin 24% and alpha 1-I3 16%. The 60% value of alpha 1-M bound radioactivity in normal serum corresponds to the sum of alpha 1-M and alpha 2-AP labelling in inflammatory serum.     

Serum elastase 1 and immunoreactive trypsin in chronic pancreatic disease: is there any relationship with trypsin inhibitors?

In order to investigate modifications of serum levels of elastase 1, immunoreactive trypsin, alpha 1-antitrypsin and alpha 2-macroglobulin in chronic pancreatic disease, and to speculate on the possible relationships among these parameters, the enzymes and inhibitors were assayed in the sera of 33 control subjects, 34 pancreatic cancer, 28 chronic pancreatitis and 36 extra-pancreatic diseases. An increase of elastase 1, alpha 1-antitrypsin and alpha 2-macroglobulin was detected in pancreatic cancer, chronic pancreatitis and extra-pancreatic diseases; no changes were found for serum immunoreactive trypsin. Multiple regression analyses showed that only 7% of elastase 1 was explained by inhibitors with alpha 1-antitrypsin playing a major role. Inhibitors did not influence immunoreactive trypsin. Our data indicate that the variations of the serum levels of proteases and antiproteases in chronic pancreatic disease are probably independent of each other.     

New human colorectal carcinoma cell lines that secrete proteinase inhibitors in vitro

Two new human cell lines, RCM-1 and CoCM-1, have been established from primary colorectal adenocarcinomas. Both cell lines were unique in that the cultures secreted trypsin inhibitors in vitro. The activities of these inhibitors were accumulated in serum-free media of both cell lines over a period of several days. Two inhibitors (PI-1 and PI-2) were isolated from serum-free conditioned medium in which RCM-1 was grown by anion-exchange and gel filtration high-performance liquid chromatography. PI-1 inhibited trypsin and chymotrypsin strongly, and pancreatic elastase weakly. Its molecular weight was about 57 kilodaltons (Kd) as determined by gel filtration chromatography. It cross-reacted with the antiserum elicited against human alpha 1-antitrypsin in double immunodiffusion. PI-1 corresponding to alpha 1-antitrypsin was also demonstrated immunohistochemically in both cell lines. PI-2 inhibited trypsin strongly, and chymotrypsin, kallikrein and plasmin weakly. It had higher molecular weight (200-300 Kd) than that of PI-1, and did not cross-react with antisera against human alpha 1-antitrypsin, alpha 2-macroglobulin, alpha 1-antichymotrypsin, alpha 2-plasmin inhibitor, inter-alpha-trypsin inhibitor and urinary trypsin inhibitor. RCM-1 and CoCM-1 are the first colorectal adenocarcinoma cell lines that secrete functionally active trypsin inhibitors, including alpha 1-antitrypsin in vitro, and are useful for the study of tumor-cell derived proteinase inhibitors.     

Comparison of the inhibition of thrombin by three plasma protease inhibitors.

Human alpha-thrombin is inhibited by the circulating protease inhibitors alpha1-antitrypsin, antithrombin III, and alpha2-macroglobulin. Kinetic analyses of the inhibitor thrombin interactions were carried out utilizing either fibrinogen or the synthetic substrate Bz-Phe-Val-Arg-p-nitroanilide as substrates to determine residual thrombin activity. These studies demonstrated that the inhibition of thrombin by alpha1-antitrypsin, antithrombin III, and alpha2-macroglobulin followed second-order kinetics. The rate constants for the inhibition of thrombin by alpha1-antitrypsin, antithrombin III, and alpha2-macroglobulin are 6.51 +/- 0.38 x 10(3), 3.36 +/- 0.34 x 10(5), and 2.93 +/- 0.02 x 10(4) M-1 min-1, respectively. Comparison of the second-order rate constants and the normal plasma levels of the three inhibitors demonstrates that, under the in vitro conditions utilized, antithrombin III is five times and alpha2-macroglobulin is one-third as effective as alpha1-antitrypsin in the inhibition of thrombin.     

Photoreduction and incorporation of iron into ferritins.

The characteristics of a new kallikrein-binding protein in human serum and its activities were studied. Both the kallikrein-binding protein and alpha 1-antitrypsin form 92 kDa SDS-stable and heat-stable complexes with human tissue kallikrein. In non-SDS/PAGE, the mobility of these complexes differ. Complex-formation between kallikrein and the binding protein is inhibited by heparin, whereas that between kallikrein and alpha 1-antitrypsin is heparin-resistant. In normal or alpha 1-antitrypsin-deficient-serum, the amount of 92 kDa SDS-stable complex formed upon addition of kallikrein is not related to serum alpha 1-antitrypsin levels. The rate of complex-formation between kallikrein and the binding protein is 12 times higher than that between kallikrein and alpha 1-antitrypsin. Purified alpha 1-antitrypsin, which exhibits normal elastase binding, has a kallikrein-binding activity less than 5% of that of serum. Binding of tissue kallikrein in serum is not inhibited by increasing elastase concentrations, and elastase binding in serum is not inhibited by excess tissue kallikrein. A specific monoclonal antibody to human alpha 1-antitrypsin does not bind to either 92 kDa endogenous or exogenous kallikrein complexes isolated from human serum. The studies demonstrate a new tissue kallikrein-binding protein, distinct from alpha 1-antitrypsin, is present in human serum.     

Levels of alpha 1-antitrypsin and alpha 2-macroglobulin in blood serum and its antitryptic capacity

Immunochemical analysis in combination with gel filtration and isoelectric focusing made it possible to state that in blood serum of healthy people 81.3 +/- 0.5% of administered trypsin is bound with alpha 1-antitrypsin and 18.7 +/- 0.6%--with alpha 2-macroglobulin. The latter is functionally heterogeneous, only 40% of it is bound with trypsin and in the formed complex the antigenic properties of trypsin and alpha 2-macroglobulin are lost. A great number of blood serum alpha 1-antitrypsin cannot fix trypsin. The content of such alpha 1-antitrypsin rises sharply with pathology available. In the immunochemical estimation of the organism inhibitory potential relative to proteolytic enzymes not only the amount of the inhibitor but also its functional activity should be taken into account. The data of immunochemical research of the blood serum isoelectrophoregrams show that the most considerable changes under conditions of pathology occur in alpha 2-macroglobulin.     

Mutant recombinant serpins as highly specific inhibitors of human kallikrein 14

The reactive center loop (RCL) of serpins plays an essential role in the inhibition mechanism acting as a substrate for their target proteases. Changes within the RCL sequence modulate the specificity and reactivity of the serpin molecule. Recently, we reported the construction of alpha1-antichymotrypsin (ACT) variants with high specificity towards human kallikrein 2 (hK2) [Cloutier SM, Kündig C, Felber LM, Fattah OM, Chagas JR, Gygi CM, Jichlinski P, Leisinger HJ & Deperthes D (2004) Eur J Biochem271, 607-613] by changing amino acids surrounding the scissile bond of the RCL and obtained specific inhibitors towards hK2. Based on this approach, we developed highly specific recombinant inhibitors of human kallikrein 14 (hK14), a protease correlated with increased aggressiveness of prostate and breast cancers. In addition to the RCL permutation with hK14 phage display-selected substrates E8 (LQRAI) and G9 (TVDYA) [Felber LM, Borgo?o CA, Cloutier SM, Kündig C, Kishi T, Chagas JR, Jichlinski P, Gygi CM, Leisinger HJ, Diamandis EP & Deperthes D (2005) Biol Chem386, 291-298], we studied the importance of the scaffold, serpins alpha1-antitrypsin (AAT) or ACT, to confer inhibitory specificity. All four resulting serpin variants ACT(E8), ACT(G9), AAT(E8) and AAT(G9) showed hK14 inhibitory activity and were able to form covalent complex with hK14. ACT inhibitors formed more stable complexes with hK14 than AAT variants. Whereas E8-based inhibitors demonstrated a rather relaxed specificity reacting with various proteases with trypsin-like activity including several human kallikreins, the two serpins variants containing the G9 sequence showed a very high selectivity for hK14. Such specific inhibitors might prove useful to elucidate the biological role of hK14 and/or its implication in cancer.     

Increased survival of rat hepatocytes in serum-free medium by inhibition of a trypsin-like protease associated with their plasma membranes

Bovine pancreatic trypsin-inhibitor (bPTI) is required for survival of adult rat hepatocytes for more than 2 days in primary cultures in serum-free medium. Of the various protease inhibitors tested, all trypsin inhibitors increased the survival of rat hepatocytes in serum-free medium, their potencies being in the order bPTI greater than alpha 2-plasmin inhibitor greater than leupeptin greater than soybean trypsin inhibitor greater than alpha 1-antitrypsin = alpha 2-macroglobulin. Elastatinal, a specific inhibitor of elastase, was also effective. bPTI did not inhibit the degradation of proteins with short or long lives, suggesting that it did not increase the survival of hepatocytes by inhibiting cellular protein degradation. alpha 2-Plasmin inhibitor immobilized on Sepharose 4B caused dose-dependent increase in survival. Plasma membranes purified from adult rat liver had significant protease activity, about 80% of which was sensitive to bPTI, alpha 2-plasmin inhibitor and leupeptin. From its specificity for substrates and sensitivity to inhibitors, the membrane-bound protease was characterized as a trypsin-like protease. The effects of various inhibitors on the membrane-bound protease correlated well with their abilities to increase survival of rat hepatocytes. Therefore, it seems that bPTI acts on the cell surface and increases hepatocyte survival in serum-free cultures by inhibiting a trypsin-like protease associated with the plasma membranes.