Polarized secretion of fibrinogen by lung epithelial cells.

The lung epithelium has recently been identified as a novel site of fibrinogen (FBG) biosynthesis. A coordinated upregulation of A alpha, B beta, and gamma chain FBG gene transcription occurs upon stimulation of A549 lung epithelial cells with dexamethasone (DEX) and the proinflammatory mediator interleukin-6 (IL-6). Subsequently, the cells synthesize and secrete fully assembled FBG. This study addresses the polarity of such FBG secretion by A549 cells cultured on polycarbonate membrane filters. After induction with IL-6 and DEX, cells were metabolically labeled, and FBG was immunopurified from the apical and basolateral chambers. Analysis by gel electrophoresis revealed that A549 cells secreted newly synthesized FBG in a polarized manner, with the majority (80%) of FBG secreted basolaterally. Consistent with this observation, immunoelectron microscopy using Protein A-gold labeling showed FBG within secretory vesicles in close proximity to the basolateral aspect of the A549 cell membrane. Polarized secretion was microtubule-dependent since depolymerization using colchicine significantly reduced the basolateral component of secretion, causing intracellular retention of FBG. These data provide evidence that FBG is secreted by lung alveolar epithelial cells vectorially toward the basement membrane, which may reflect in vivo processes associated with local injury, inflammation, and repair mechanisms.

Calcium modulates plasmin cleavage of the fibrinogen D fragment gamma chain N-terminus: mapping of monoclonal antibody J88B to a plasmin sensitive domain of the gamma chain.

Plasmin sensitive sites are found on the A alpha, B beta and gamma chains of fibrinogen at regions joining the two C-terminal D fragments with the central E fragment. We have developed a monoclonal antibody (MoAb) reactive with this plasmin sensitive region on the human fibrinogen gamma chain and mapped its epitope. MoAb J88B reacts with gamma chains of both native as well as with reduced and denatured fibrinogen and fibrin, the CNBr fragment of the fibrinogen central domain, plasmin cleaved fragments D, gamma-gamma dimers, but not with plasmic fragments E. These data indicate that J88B maps to the plasmin sensitive domain localized to gamma 63-78. MoAb J88B failed to react with synthetic peptide gamma 70-78, which suggests that the epitope includes the newly exposed N-terminal residues gamma 63-70 of the early plasmic fragment D1A. As calcium has a marked influence on plasmin cleavage of C-terminal sites on the gamma chain, the effects of calcium on modulating plasmin cleavage of D1A to D1 were assessed in the absence or presence of J88B. The results indicated that calcium delays and J88B (+/- calcium) protects the gamma chain from plasmin cleavage at the N-terminus of D1A, suggesting that this enzymatically labile site is calcium-sensitive. Thus, MoAb J88B should prove useful in studies examining the structure of plasmin cleaved fibrinogen and fibrin.

Thrombin cleavage enhances exposure of a heparin binding domain in the N-terminus of the fibrin beta chain.

Thrombin (IIa)-cleavage of fibrinogen (FBG) to form polymerized fibrin promotes endothelial cell spreading, proliferation, and von Willebrand factor release, requiring the exposure of the beta 15-42 domain. Studies reported here indicate that IIa-cleavage of fibrinopeptide B enhances exposure of a heparin binding domain at the beta 15-42 neo-N-terminus of fibrin. Crossed immunoelectrophoresis showed heparin-induced mobility shifts indicative of complexing with FBG and with N-terminal CNBr fragments of FBG (NDSK) and of fibrin (IIa-NDSK), but not evidence of heparin complexing with FBG lacking B beta 1-42 or with FBG fragments D and E was seen. Elution from heparin-agarose with a linear gradient of NaCl showed that bound portions of both intact FBG and D fragments eluted below physiologic salt concentrations, whereas E3 fragments lacking B beta 1-53 did not bind. NDSK bound with higher affinity than did intact FBG, whereas binding of IIa-NDSK was maximal in this system. Binding of fibrin(ogen) to heparin agarose was saturable as well as inhibitable in a dose-dependent manner with both FBG and heparin. Scatchard analysis indicated a single class of binding site, with dissociation constants (kd) of 0.3 mumol/L for IIa-NDSK, 0.8 mumol/L for NDSK, and 18 mumol/L for FBG. Immobilized fibrin had twofold more heparin binding sites than did immobilized FBG and required a 5.5-fold higher concentration of heparin to inhibit by 50% the binding of labeled heparin. Together, the results indicate that IIa-cleavage results in enhanced exposure of two heparin binding domains (HBDs) with approximately threefold higher affinity in fibrin than in FBG. Synthetic peptide beta 15-42 showed highest binding to heparin-agarose followed by B beta 1-42, whereas peptides beta 18-31, beta 18-27, and beta 24-42 did not bind. Thus, the primary structure of beta 15-42 is required for specificity of heparin binding. Basic residues within the beta 15-32 region segregate primarily to one side of an alpha-helix in a helical wheel diagram, as is typical for authentic HBDs. Desulfated heparin and heparan sulfate bound more fibrin(ogen) than did other proteoglycans; however, heparin bound sixfold more Ila-NDSK than NDSK. These results confirm that fibrin binds to heparin with higher affinity than does FBG and that fibrin binding is not solely dependent on charge interactions of beta 15-42 with the negatively charged glycosaminoglycan.

Purification and functional characterization of homodimeric gamma B-gamma B fibrinogen from rat plasma.

Heterogeneity in the human and rat plasma fibrinogen (FBG) gamma chains is due to differential splicing of the primary gamma chain mRNA transcript. The subunit composition of FBG containing the predominant form of the gamma chain is homodimeric (human: A alpha 2, B beta 2, gamma 50-gamma 50; rat: A alpha 2, B beta 2, gamma A-gamma A). The subunit composition of FBG containing the longer, minor form of the gamma chain is heterodimeric (human: A alpha 2, B beta 2, gamma 50-gamma 57.5; rat: A alpha 2, B beta 2, gamma A-gamma B). The variant gamma chain-FBGs comprise about 10% of the total plasma FBG in the human and 30% in the rat. Although the presence of homodimeric gamma B-gamma B FBG has been speculated, proof of its existence has been difficult to obtain experimentally. We now show that 5% to 6% of rat plasma FBG is found as homodimeric FBG with a subunit composition of A alpha 2, B beta 2, gamma B-gamma B. Commercially purified rat FBG was further purified by diethylaminoethyl-Sephacel chromatography with a combined pH and ionic strength gradient. The enriched gamma B-gamma B FBG fraction eluted at the lowest pH (6.3) and highest ionic strength (4.5 mmho) due to its increased net negative charge. To further purify gamma B-gamma B FBG, a Mono Q column with an NaCl gradient was used. The subunit composition of the purified gamma B-gamma B FBG was confirmed by its electrophoretic mobility under reducing and denaturing conditions both as FBG, and clotted and cross-linked fibrin. Homodimeric gamma B-gamma B FBG was unable to support adenosine diphosphate (ADP)-induced platelet aggregation, whereas heterodimeric gamma A-gamma B FBG was able to support ADP-induced platelet aggregation at 40% of that achieved with homodimeric gamma A-gamma A FBG, similar to previous observations with human FBGs. Taken together, these results support the hypothesis that the A alpha-RGD sites alone are not sufficient for dimeric FBG support of platelet aggregation. Furthermore, the data suggest that at least one intact platelet recognition site at the carboxyterminus of the gamma A or gamma 50 chain is required.

E-selectin-dependent neutrophil adhesion to Rickettsia rickettsii-infected endothelial cells.

Increased neutrophil or HL60 cell adhesion to Rickettsia rickettsii-infected endothelial cells (ECs) was observed at 6 to 8 hours after the initiation of infection, diminishing by 24 hours. Similar increases were observed using formaldehyde-fixed neutrophils. Cellular association and likely the intracellular presence of rickettsiae was required for enhanced neutrophil adhesion, because culture medium conditioned by infected cells or rickettsiae rendered noninfective by pretreatment with tetracycline were ineffective at inducing neutrophil adhesion. Increases in neutrophil adhesion caused by infection were blocked by pretreatment of ECs with cycloheximide, suggesting the involvement of new protein synthesis in the cells' response. Flow cytometric analysis of infected cells showed increases in cell surface expression of E-selectin compared with uninfected control cells. Furthermore, incubation of 6- to 8-hour infected cells with a blocking monoclonal antibody against E-selectin (BB11) inhibited neutrophil adhesion an average of 61%. These results suggest the involvement of E-selectin in neutrophil adhesion to infected ECs occurring early in the course of the infection process. EC-initiated recruitment of neutrophil adhesion during rickettsiae infection could contribute to the pathologic changes associated with Rocky Mountain Spotted Fever.

Rickettsia rickettsii infection of cultured endothelial cells induces release of large von Willebrand factor multimers from Weibel-Palade bodies.

The clinical manifestations of Rocky Mountain spotted fever (RMSF) result from Rickettsia rickettsii (R rickettsii) infection of endothelial cells and are mediated by pathologic changes localized to the vessel, including in situ thrombosis and tissue ischemia. This study uses in vitro infection of cultured human umbilical vein endothelial cells with R rickettsii to test the hypothesis that such infection induces von Willebrand factor (vWF) release from Weibel-Palade bodies, a process that could contribute to thrombotic changes. At 24 hours postinfection, there was an increase in metabolically prelabeled large multimers of vWF in the culture medium, with a concomitant decrease of these forms in the cell lysate samples. This release reaction was specific for the large multimer pool of vWF, localized to Weibel-Palade bodies, because no change in the distribution of dimeric forms between cells and culture medium was detected. Double-label immunofluorescence staining showed an inverse correlation between the number of R rickettsii and the number of Weibel-Palade bodies in infected cells. Cell lysis was minimal at 24 hours postinfection, as no detectable intracellular precursor forms (molecular weight 260,000) of vWF were released into the culture medium, there was no decrease in cell viability as measured by trypan blue exclusion, and no increase in 51Cr-release into the culture medium was observed when compared with uninfected controls. Release was likely a direct effect of the intracellular presence of the organism, rather than due to a noxious soluble factor such as endotoxin, because culture medium conditioned by infected endothelial cells was ineffective at inducing release in uninfected endothelial cell cultures. In summary, in vitro infection of endothelial cells by R rickettsii induces release of Weibel-Palade body contents, a process that may contribute to the pathogenesis of RMSF.

Divergent fates of von Willebrand factor and its propolypeptide (von Willebrand antigen II) after secretion from endothelial cells.

The intracellular site of cleavage of pro-von Willebrand factor subunit and the subsequent fate of the propolypeptide (von Willebrand antigen II) and of the mature von Willebrand factor (vWf) were investigated. Both the propolypeptide, which was found to be a homodimer of noncovalently linked subunits, and mature vWf were released from Weibel-Palade bodies of endothelial cells following stimulation with secretagogues. The stoichiometry of the two released proteins was essentially equimolar. This indicates that vWf and the propolypeptide were packaged into the Weibel-Palade bodies as one unit, pro-vWf, and that the proteolytic cleavage of pro-vWf is likely to be a post-Golgi event. The association of prosequences into dimers supports their hypothetical role in the multimerization process. After secretion, the two proteins were distributed differently, as based on the following observations. The propolypeptide did not associate with vWf in the culture medium, did not codistribute with vWf in the extracellular "patches of release" on stimulated endothelial cells, and was not detected in the endothelial cell extracellular matrix, which did contain vWf. Additionally, in contrast to vWf, the propolypeptide did not bind to the matrix of human foreskin fibroblasts. Since the propolypeptide does not associate with vWf and does not interact with extracellular matrices in vitro, it is highly unlikely that it would promote platelet adhesion to subendothelium in vivo.

Topology and order of formation of interchain disulfide bonds in von Willebrand factor.

Interchain disulfide bonds between the subunits in von Willebrand factor (vWf) dimers and in vWf multimers have been studied using some unique features of the cultured human umbilical vein endothelial cell system. Ammonium chloride inhibition of multimerization of vWf allowed selective examination of vWf dimeric molecules, and monoclonal antibody against the vWf propolypeptide was used to separate pro-vWf dimers from mature dimers. After cleavage of dimers and multimers with Staphylococcus aureus V-8 protease, the location of interchain disulfide bonds in amino (N)-terminal or carboxyl (C)-terminal fragments was determined by gel electrophoresis under reduced and nonreduced conditions. The first interchain disulfide bonds formed during dimerization are in the C-terminal region of the subunits, whereas interdimer disulfide bonds are located in the N-terminal portion. These data confirm recent electron microscopic projections of disulfide bond locations and provide support to the hypothetical role of the propolypeptide in the multimerization process.