Proteolysis of factor VIII heavy chain polypeptides in plasma and concentrates .sx G. Kemball-Cook , S. J. Edwards and T. W. Barrowcliffe .sx Summary .sx Factor VIII heavy chain ( FVIII HC ) polypeptides have been studied in both normal plasma and FVIII concentrates on exposure to three coagulation proteases .sx FVIII samples were incubated with labelled affinity-purified anti-FVIII Fab' fragments , immunocomplexes formed were visualized by autoradiography after sodium dodecyl sulphate polyacrylamide gel electrophoresis ( SDS-PAGE ) , and apparent relative molecular masses ( M r ) of each band assigned .sx FVIII HC polypeptides were detected in all types of samples , including plasma , without further purification .sx Normal plasma contained a range of polypeptides with the largest dominant band at a net apparent M r of 250-300 kD , and the smallest at 80-90 kD ; the bands visualized correspond to the 90-120 kD HC species seen on conventional analysis of purified FVIII .sx No bands were produced from samples of haemophilic plasma .sx Treatment of plasma or FVIII concentrate with low concentrations ( 1 IU/ml ) of thrombin removed the 250-300 kD and other intermediate bands , intensified then removed the 80-90 kD polypeptide and produced a band at 40-50 kD .sx Thrombin-associated rise and fall in FVIII clotting activity by one-stage assay correlated with intensity of the 80-90 kD polypeptide .sx A polypeptide of M r 40-50 kD was also produced after incubation with activated factor X :sx activated factor VII plus thromboplastin had no effect on HC structure .sx FVIII polypeptides were visualized in prothrombin complex concentrates , with a more degraded profile seen in a deliberately 'activated' product .sx In recent years the primary structure of factor VIII ( FVIII ) has been elucidated by purification of the protein ( Fulcher & Zimmerman , 1982 ; Rotblat et al , 1985 ) and cloning of the gene ( Toole et al , 1984 ; Vehar et al , 1984) .sx In addition , recombinant FVIII has been produced using genetic engineering techniques ( Wood et al , 1984 ; Truett et al , 1985) .sx Despite these great advances , the exact nature of the FVIII species active in coagulation is subject to conjecture .sx Proteolysis of highly-purified FVIII by coagulation pro - teases , especially thrombin , activated factor X ( FXa ) and activated protein C ( APC ) , closely define its biological activity ( Eaton et al , 1986) .sx Specific cleavages made by thrombin or FXa in the heavy chain ( HC ) and light chain ( LC ) are thought to be essential to produce the active form which takes part in the activation of FXa as a cofactor , while APC and FXa can make inactivating cleavages in the heavy chain .sx Also , the cleaved fragments of HC and LC must be in association [probably via divalent cations] in order to retain activity ( Fulcher et al , 1985 ; Andersson et al , 1986 ; Fay et al , 1986) .sx The structure of thrombin-activated human FVIII has been postulated to be a 51 kD/73 kD heterodimer derived from the HC and LC respectively , with a conformational relaxation possibly responsible for loss of activity on further incubation ( Fay , 1987) .sx However , a stable form of porcine activated FVIII consisting of a heterotrimer has also been reported ( Lollar & Parker , 1989) .sx All these studies on polypeptide cleavage of plasma-derived FVIII have been performed using very highly-purified FVIII .sx However , the FVIII molecule is very sensitive to proteolysis during isolation , for example during the manufacture of FVIII concentrates ( Barrowcliffe et al , 1986 ; Kemball-Cook et al , 1990) .sx Thus studies using purified FVIII may not represent the proteolysis of the molecule in , for example , plasma or low purity FVIII concentrates .sx In a recent paper ( Kemball-Cook et al , 1990 ) , we demonstrated a modified sodium dodecyl sulphate polyacrylamide gel electrophoresis ( SDS-PAGE ) method for visualization of factor VIII heavy chain ( FVIII HC ) polypeptides .sx This approach , based on that first described by Weinstein et al ( 1981 ) enables FVIII structure to be studied in a wide range of samples including plasma without further purification .sx We have therefore used this technique to study the proteolytic breakdown of FVIII HC in plasma and concentrates when exposed to a range of coagulation enzymes .sx In addition a correlation between HC structure and FVIII activity ( FVIII :sx C ) following thrombin treatment has been made .sx MATERIALS AND METHODS .sx 125 I-labelled specific anti-FVIII Fab' fragments were prepared from inhibitor plasma CC8000 as previously described ( Kemball-Cook et al , 1990) .sx Fresh human plasmas were obtained by clean venepuncture and immediately anticoagulated by mixing 9 :sx 1 with 0.109 M trisodium citrate .sx Severe haemophilic plasma was obtained frozen from Dr P. B. A. Kernoff at the Royal Free Hospital , London .sx FVIII samples were prepared for SDS-PAGE by diluting to between 0.5 and 5 IU/ml FVIII clotting activity ( FVIII :sx C ) , then 10 mu l were mixed with 10 mu l of 125 I-Fab' ( 3000-10 000 cpm ) and 4 mu l of 9% ( w/v ) PEG 4000 ( Sigma , Poole , Dorset ) in the presence of protease inhibitors ( see below ) , and incubated in a closed plastic tube at 37 degree C for 2 h to allow formation of immunocomplexes .sx 20 mu l of sample buffer ( 62 m M Tris-HCl-10% glycerol-2% SDS-0 .sx 001% bromophenol blue , pH 6.8 ) were added and incubation continued for a further 30 min .sx after which samples ( 30-35 mu l ) were loaded onto the gel .sx Inhibitors added to the 125 I-Fab' to terminate reactions with proteases were recombinant desulphatohirudin ( CGP 39393 , kind gift of Dr R. B. Wallis , Ciba-Geigy , Horsham , Sussex ) , 100 mu g/ml ; I-2581 ( Kabi , Uxbridge , Middlesex ) , 25 mu M ; and PMSF ( Sigma , Poole , Dorset ) , 10m M. Factor VIII concentrates were type 8Y ( Bio Products Laboratory , Elstree , Hertfordshire ) ; Profilate HT solvent - treated ( Alpha Therapeutics , Thetford , Norfolk ) and Hemofil M ( Baxter Travenol , Thetford , Norfolk) .sx The BPL and Alpha concentrates are low purity conventionally-produced materials ( specific activity approximately 5 IU FVIII :sx C/mg protein ) , while Hemofil M is made using immobilized monoclonal antibodies to FVIII and is of high specific activity ( >1000 IU/mg ) before addition of albumin stabilizer .sx Prothrombin complex concentrates :sx non-activated ( Pro - thromplex ) and activated ( Feiba ) materials were supplied by Immuno ( Sevenoaks , Kent) .sx Purfified human thrombin was NIBSC reagent 82/570 , specific activity 700 IU/mg protein .sx Purified human factor X ( FX ) was prepared by the method of Modi et al ( 1984 ) using factor IX concentrate ( Bio Products Laboratory , Elstree , Hertfordshire ) as starting material .sx Sulphated dextran was produced according to Miletich et al , ( 1980) .sx RVV-X , the FX-activating protein from Russell's Viper venom ( RVV ) , was purified from crude RVV ( Diagen ) by ion exchange chromatography on QAE-Sephadex ( Pharmacia) .sx Human activated FX ( FXa ) was prepared in situ by mixing purified FX with RVV-X protein at a ratio of 10 :sx 1 in the presence of 5 m M CaCl 2 for 15 min :sx under these conditions FX was fully activated .sx Recombinant FVIIa was kindly supplied by Novo Nordisk , Bagsvaerd , Denmark .sx Rabbit brain thromboplastin was a product of Manchester Comparative Reagents Ltd , Stockport , Cheshire .sx Thrombin treatment of FVIII samples :sx samples were mixed with thrombin and incubated at the temperature indicated .sx Subsamples removed for SDS-PAGE were added immediately to 125 I-Fab'/PEG in the presence of protease inhibitors .sx One-stage assays on subsamples from thrombin activation of FVIII were carried out as follows :sx 100 mu l of severe haemophilic plasma were mixed with 100 mu l kaolin ( 5 mg/ml ) and incubated for 9 min :sx 100 mu l of phospholipid ( NIBSC 86/516 bovine brain extract , 10 mu g/ml ) and 100 mu l of FVIII sample were added , then at 10 min the mixture was recalcified with 100 mu l 33 m M CaCl 2 and the clotting time recorded .sx The percentage activation was calculated by reference to a standard curve constructed from the sample prior to thrombin addition .sx FXa proteolysis of FVIII was studied by incubating very high purity ( VHP ) FVIII concentrate ( final concentration 5 IU/ml FVIII :sx C ) with a mixture of purified human FX ( 10 mu g/ml ) , RVV-X protein ( 1 mu g/ml ) and CaCl 2 ( 5 m M ) for 15 min at 37 degree C before addition of labelled antibody :sx the ratio of FX to RVV-X was selected to give full activation of FX in 10-15 min .sx Recombinant hirudin ( 100 mu g/ml ) was present throughout to exclude the action of any trace thrombin .sx Factor VIIa ( FVIIa)/thromboplastin ( TP) :sx VHP FVIII concentrate ( 2.5 IU/ml FVIII ) was incubated with recombinant FVIIa ( 27 U/ml FVII :sx C ) in the presence or absence of 10 m M CaCl 2 and TP ( Manchester , diluted 1 in 10 ) , before incubation at 37 degree C for 90 min .sx This FVIIa/TP/CaCl 2 mixture clotted FVII-deficient plasma in less than 1 min .sx Slab gel preparation , electrophoresis and autoradiography were carried out as previously described ( Kemball-Cook et al , 1990) .sx Molecular weight markers for use in non-reduced Laemmli SDS-PAGE were HMW kit ( Pharmacia , Milton Keynes , Buckinghamshire ) and SDS-6H ( Sigma , Poole , Dorset) .sx Assignment of apparent relative molecular masses ( M r ) :sx calculation of approximate M r was by measurement of polypeptide migration distance on the autoradiograph and comparison with molecular weight markers on the dried Coomassie blue-stained gels , with subtraction of 50 kD ( =M r of one Fab' molecule ) to give the apparent net mass of the polypeptide .sx Quantitative densitometry was carried out using an LKB Ultroscan Laser Densitometer .sx RESULTS .sx Visualization of FVIII in plasma and FVIII concentrates .sx When 125 I-labelled A/FVIII Fab' was incubated with either fresh normal plasma or FVIII concentrate and electrophoresed , autoradiography revealed a range of bands with apparent M r between 80-90 kD and 250-300 kD :sx no bands were seen when severe haemophilic plasma was used ( Kemball-Cook et al , 1990) .sx Effect of thrombin .sx Fig 1 shows the effect of incubating either frozen normal pooled plasma or , in a separate experiment , a sample of one type of conventional low purity FVIII concentrate , with human thrombin ( final concentration of 1 IU/ml ) for 20 min at 37 degree C. In the plasma sample , the band at M r 250-300 kD disappeared and a degraded band became visible below M r 80-90 kD at M r 40-50 kD .sx The concentrate showed a loss of virtually all the original pattern with production of one major degradation band at M r approximately 40-50 kD .sx Diffuse staining seen at M r above 300 kD is due to limited self - association of Fab' monomers .sx As previously reported ( Barrowcliffe et al , 1986 ; Kemball-Cook et al , 1990 ) , a low purity FVIII concentrate from one manufacturer consistently showed a polypeptide at 40-50 kD even in the absence of added thrombin .sx In order to determine whether this polypeptide could be a degradation product formed by trace thrombin present during concentrate manufacture , samples of a batch of this product ( here coded B ) and another low purity FVIII concentrate ( A ) were incubated with thrombin ( final concentration 1 IU/ml ) for 20 min at 37 degree C to ensure complete thrombin-induced degradation .sx The SDS-PAGE results obtained before and after thrombin treatment are seen in Fig 2 .sx Although the FVIII profile in the two concentrates was quite different before thrombin treatment , afterwards very similar patterns were seen in both materials , with essentially one band only at M r 40-50 kD .sx This band co-migrated with the lowest M r band present in product B before proteolysis , indicating the possibility of partial thrombin proteolysis of FVIII in this product during manufacture .sx figure&caption .sx Thrombin cleavage and biological activity .sx In addition , the changes in polypeptide pattern in the two concentrates were correlated with the thrombin-induced rise and subsequent decay of one-stage FVIII :sx C. Fig 3 presents the polypeptide patterns obtained from subsamples taken at various times from thrombin-FVIII activation mixtures :sx the conditions were chosen to produce a relatively prolonged cleavage of FVIII .sx Both concentrates showed similar degradation patterns , with immediate disappearance of the M r 250-300 kD band ( initially visible only in product A ) , intensification of the M r 80-90 kD band and then its reduction simultaneously with appearance or intensification of the M r 40-50 kD band .sx Similar patterns of FVIII degradation over time were seen with thrombin treatment of fresh plasma ( not shown) .sx figure&caption .sx The strengths of the M r 250-300 kD , 80-90 kD and 40-50 kD HC bands seen in product A's pattern were quantitated by densitometry and are compared with the one-stage FVIII :sx C in Fig 4 .sx The rise ( to 350% of pre-incubation activity ) and fall ( to below 10% ) in FVIII :sx C were closely paralleled by the appearance and disappearance of the M r 250-300 kD band , whereas neither the M r 40-50 kD nor the M r 250-300 kD band correlated with clotting activity .sx A similar correlation was seen with thrombin treatment of product B ( not shown) .sx figure&caption .sx Effect of FXa .sx VHP FVIII concentrate ( final concentration 5 IU/ml ) was incubated with a mixture of FX ( 10 mu g/ml ) , RVV-X ( 1 mu g/ml ) and CaCl 2 ( 5 m M ) for 15 min at 37 degree C in the presence of excess recombinant hirudin , before addition of labelled antibody .sx Fig 5 presents the polypeptide patterns obtained .sx Firstly , it is apparent that before exposure to thrombin this product was largely deficient in the larger M r HC species , with most staining in the M r 80-90 kD position :sx secondly , neither RVV-X alone nor FX alone had any effect , but FXa generated in situ caused generation of a band at M r 40-50 kD .sx