Resistance of γA/γ′ Fibrin Clots to Fibrinolysis*

Elevated plasma fibrinogen levels are a major risk factor for thrombosis. This report shows two mechanisms by which fibrinogen can affect the fibrinolysis rate in vitro and thus may lead to thrombosis. First, the lysis rate of fibrin decreases as the initial concentration of fibrinogen increases. Second, a minor variant form of fibrinogen decreases the rate of fibrinolysis. This variant, γA/γ′ fibrinogen, has one altered γ chain and is known to bind to factor XIII zymogen. In a fibrinolysis assay containing purified thrombin, fibrinogen, tissue-type plasminogen activator, and plasminogen, clots from γA/γA and γA/γ′ fibrinogen lysed at similar rates. However, when factor XIII was added, slower lysis was seen in γA/γ′ fibrin clots when compared with γA/γA fibrin clots. A D-dimer agglutination assay showed that the γA/γ′ clots were more highly cross-linked than the γA/γA clots. The lysis rates of γA/γ′ clots were similar to γA/γA clots in the presence ofN-ethylmaleimide, a specific inhibitor of factor XIIIa. The γA/γ′ fibrin clots made in the presence of factor XIII showed increased proteolytic resistance to both plasmin and trypsin. Clots made from afibrinogenemic plasma reconstituted with γA/γ′ fibrinogen also showed significant resistance to lysis compared with γA/γA fibrinogen. These data demonstrate γA/γ′ fibrin is resistant to fibrinolysis, possibly as a result of concentrating factor XIII on the clot. The total fibrinogen concentration and the amount of γA/γ′ fibrinogen increase clot stability in vitro and thus may contribute independently to the risk of thrombosis in humans.

circulates as an inactive proenzyme until it is activated by thrombin cleavage of a 4000-Da activation peptide from each a subunit, which is followed by the dissociation of the b subunits. Activated factor XIII, or XIIIa, catalyzes the formation of ␥-glutamyl-⑀-lysine bonds between polypeptide chains in fibrin (5). These cross-links strengthen the fibrin clot (6) and increase its resistance to lysis (7)(8)(9).
Recent evidence has revealed an association between a variant form of fibrinogen and factor XIII that may play a role in modulating the stability of a fibrin clot. This fibrinogen variant, referred to as ␥A/␥Ј fibrinogen, peak II fibrinogen (10), ␥A-␥B fibrinogen (11), or F␥ 50, 57.5 (12), contains an altered ␥ chain termed ␥Ј (13), ␥B (14), or ␥ 57.5 (12), and comprises approximately 7-15% of the total fibrinogen found in plasma (10). The ␥Ј chain arises from alternative processing of the ␥ chain mRNA (15,16) that leads to the translation of a polypeptide with a 20-amino acid sequence substituted for the carboxyl-terminal four amino acids of the ␥ chain. The physiological function of the ␥Ј chain remains unclear; however, recent studies have shown that the ␥Ј chain binds directly to the zymogen form of factor XIII (17). This suggests that the ␥Ј chain of fibrinogen may serve as a carrier for factor XIII to increase the local concentration of factor XIII at the fibrin clot.
Elevated fibrinogen levels are a major risk factor for thrombosis. Several mechanisms have been proposed to explain the correlation between fibrinogen levels and thrombosis (18,19). Fibrinogen may contribute to thrombosis by its role in atherosclerotic plaque formation. Fibrinogen binds to platelet membranes through glycoprotein IIb-IIIa (integrin ␣ IIb ␤ 3 ) and thereby mediates platelet aggregation (20,21). Fibrin degradation products stimulate the proliferation of vascular smooth muscle cells (22,23), which may contribute to narrowing of blood vessels during atherosclerosis. In addition, once fibrinogen is incorporated into the atherosclerotic plaque, it binds low density lipoprotein and sequesters more fibrinogen (18), thus enhancing plaque formation. Another mechanism by which fibrinogen may lead to thrombotic disorders is by increasing blood viscosity. Increased blood viscosity favors the hypercoagulable state (24). Furthermore, increased fibrinogen levels lead to increased fibrin deposition (25)(26)(27). However, the causal role of these potential factors in thrombosis is still unclear.
The present report demonstrates that increases in total fibrinogen concentration result in increased clot stability toward fibrinolysis in vitro, suggesting a mechanistic explanation of how elevated fibrinogen levels per se may cause an increased risk of thrombosis. In addition, clots made from ␥A/␥Ј fibrinogen are more stable to fibrinolysis in vitro than clots made from ␥A/␥A fibrinogen. Taken together, these findings suggest that both the total amount of fibrinogen and the amount of ␥A/␥Ј fibrinogen in plasma may be independent risk factors for thrombosis. pH 8.6, containing 5 mM ⑀-aminocaproic acid (EACA) 1 and 0.2 mM phenylmethylsulfonyl fluoride (PMSF) and dialyzed exhaustively into the same buffer at 4°C. Insoluble residue was removed by centrifugation at 10,000 ϫ g for 30 min at 4°C. The ␥A/␥A and ␥A/␥Ј forms of fibrinogen were separated using DEAE-cellulose as described previously (28). Briefly, the fibrinogen solution was adsorbed to a column of DEAE-cellulose (6 ϫ 20 cm) and eluted with a 1200-ml exponential gradient generated in a 600-ml constant volume mixing chamber from the starting buffer (39 mM Tris-PO 4 , pH 8.6, 5 mM EACA, 0.2 mM PMSF) to the final buffer (193 mM Tris-PO 4 , pH 4.6, 5 mM EACA, 0.2 mM PMSF). The absorbance was monitored at 280 nm, and 11-ml fractions were collected. The elution profile showed two peaks; ␥A/␥A fibrinogen composed the first peak and ␥A/␥Ј fibrinogen composed the second smaller peak (data not shown).
In some experiments, ␥A/␥A and ␥A/␥Ј fibrinogen were further purified using a glycine-L-proline-L-arginine-L-proline-L-cysteine (GPRPC)agarose affinity resin (29). Briefly, the resin was prepared by reacting 10 mg of glycine-L-proline-L-arginine-L-proline-L-cysteine peptide (Howard Hughes Medical Institute Biopolymer Laboratory, Seattle, WA) with 10 ml of 5-thio-2-nitrobenzoate-agarose (Pierce) according to the manufacturer's protocol. The dialyzed ␥A/␥A or ␥A/␥Ј fibrinogen pool from DEAE-cellulose was adsorbed to a column (3 ml) of GPRPCagarose, washed with 100 mM NaCl, 50 mM Tris-PO 4 , pH 7. Fibrinolysis Assay Using Purified Components-Microtiter plate fibrinolysis assays were carried out as described previously by Jones and Meunier (30) using 96-well assay plates (Corning 25-880-96). Fibrinogen and Lys-plasminogen (Calbiochem) were added to an interim mixing plate containing assay buffer (0.1 M NaCl, 30 mM NaHCO 3 , 4 mM KCl, 1 mM CaCl 2 , 1 mM Na 2 HPO 4 , 0.3 mM MgCl 2 , 0.4 mM MgSO 4 , 10 mM HEPES, pH 7.4, 0.01% Polysorbate 80). A separate assay plate contained ␣-thrombin (a generous gift from Dr. Walter J. Kisiel, University of New Mexico) and tissue plasminogen activator (Calbiochem) in assay buffer. The fibrinogen/plasminogen solution was then dispensed from the interim plate into the assay plate wells containing thrombin and tissue plasminogen activator. The final concentrations of reagents were 1.25 mg/ml fibrinogen (unless stated otherwise), 30 g/ml Lys-plasminogen, 16 ng/ml tissue plasminogen activator, and 13.2 NIH units/ml thrombin in a total volume of 100 l. For assays containing factor XIII (a generous gift from Drs. Michael W. Mosesson and David A. Meh, University of Wisconsin Medical School, Milwaukee Clinical Campus), factor XIII was added to a final concentration of either 10 or 100 g/ml to the wells in the interim plate containing the plasminogen/fibrinogen mixture. In some assays 1 mM N-ethylmaleimide was also added to the interim plate. The turbidity of the clot was measured at room temperature every 6 min at 405 nm. The optical density was converted to percent lysis as follows, and plotted versus time; the data were then fit to a sigmoidal curvefitting routine. For clarity, a maximum of 15 data points per set are shown in each graph. D-Dimer Assay-A D-dimer agglutination assay kit (Sigma) was used to measure D-dimer content in lysed fibrin clots. Clots were formed from ␥A/␥A or ␥A/␥Ј fibrinogen using 1.0 mg/ml fibrinogen, 13.2 NIH units/ml ␣-thrombin, 16 ng/ml tissue-type plasminogen activator, 30 g/ml Lys-plasminogen, and 10 g/ml factor XIII and allowed to lyse completely. After lysis, 50 l of the lysed clots were serially diluted in the manufacturer's proprietary buffered saline solution, pH 7.3, containing 0.02% sodium azide. 20 l of the diluted sample were then mixed with 20 l of a suspension of latex beads coated with mouse monoclonal antibody (MA-8D3) specific for the fibrin D-dimer on the test card. The test card was rocked back and forth, and agglutination was scored after 3 min on duplicate samples.
Trypsin Digest of Fibrin Clots-␥A/␥A or ␥A/␥Ј fibrinogen and factor XIII were added to a 96-well assay plate containing assay buffer (0.1 M NaCl, 30 mM NaHCO 3 , 4 mM KCl, 1 mM CaCl 2 , 1 mM Na 2 HPO 4 , 0.3 mM MgCl 2 , 0.4 mM MgSO 4 , 10 mM HEPES, pH 7.4, 0.01% Polysorbate 80), and ␣-thrombin was added to the wells to initiate clotting. The final concentrations of the reagents were 1.0 mg/ml fibrinogen, 10 g/ml factor XIII, and 13.2 NIH units/ml ␣-thrombin in a total volume of 100 l. The clots were incubated at room temperature until maximum turbidity was reached. When clotting was complete, 50 l of 11.7 g/ml trypsin were added to the clots. The turbidity of the clots was measured every 5 min at 405 nm. The optical density was converted to percent lysis as described above.
Whole Plasma Fibrinolysis Assay-The ␥A/␥A and ␥A/␥Ј forms of fibrinogen were iodinated with Na 125 I (DuPont NEN) according to the method of Fraker and Speck (31) to a specific activity of 85,000 -120,000 cpm/ng as described previously (29). The plasma fibrinolysis assays were carried out using the method of Wun and Capuano (32). Briefly, the iodinated ␥A/␥A or ␥A/␥Ј fibrinogen in 137 mM NaCl, 2.7 mM KCl, 1 mM CaCl 2 , 10 mM HEPES, pH 7.4, was added 1:1 (v/v) to fibrinogendeficient human plasma (George King Biomedical) at a final concentration of 1.25 mg/ml on ice. (The plasma contained less than 0.1 mg/ml residual fibrinogen.) Clotting of the samples was initiated with 1 NIH unit/ml ␣-thrombin and the clots were incubated at 37°C in the presence of 0.02% NaN 3 . Samples were taken daily for 7 days and centrifuged for 5 min in a microcentrifuge to pellet the insoluble clot. 20-l aliquots of the supernatant were counted in a gamma counter to measure the soluble 125 I-fibrin degradation products. The percentage of clot lysis was calculated using the following equation and plotted versus time.

RESULTS
Fibrinolysis Rates Vary Inversely with Fibrinogen Concentration-Since high plasma fibrinogen levels predispose individuals to a greater risk of thrombosis, we hypothesized that clots made from high concentrations of fibrinogen may be more resistant to fibrinolysis. To test this hypothesis, fibrinolysis assays were performed using purified components, including fibrinogen, thrombin, tissue-type plasminogen activator, and plasminogen (30). Concentrations of fibrinogen from 1 to 4 mg/ml were used, and clotting was initiated by the addition of thrombin. The calculated lysis t1 ⁄2 for clots formed from 1, 2, 3, and 4 mg/ml total fibrinogen concentrations was 18.7 Ϯ 0.1, 30.4 Ϯ 0.7, 55.7 Ϯ 0.6, and 70.3 Ϯ 0.3 min, respectively (Fig. 1). The fibrinolysis half-times were essentially linear with respect to the initial fibrinogen concentrations (Fig. 1, inset). Similar Fibrinolysis assays using 13.2 NIH units/ml ␣-thrombin, 16 ng/ml tissue-type plasminogen activator, 30 g/ml Lys-plasminogen, and 10 g/ml factor XIII were carried out at room temperature using the following concentrations of fibrinogen: 1 mg/ml (q), 2 mg/ml (E), 3 mg/ml (ϫ), and 4 mg/ml (f). Clot turbidity was monitored by absorbance at 405 nm and expressed as percent lysis. Each point is the average of duplicate determinations. Inset, the clot lysis t1 ⁄2 was determined from each curve and graphed versus fibrinogen concentration. results were seen in the absence of factor XIII (data not shown). These results are consistent with other published data (30) and demonstrate that fibrin clots made at high fibrinogen concentrations are more resistant to lysis, suggesting a mechanistic explanation for the role of fibrinogen as a risk factor for thrombosis (33).
␥A/␥Ј Fibrin Clots are Resistant to Fibrinolysis-A second potential risk factor for thrombosis is the amount of ␥A/␥Ј fibrinogen in plasma. This form of fibrinogen has been shown to act as a carrier protein for factor XIII by binding directly to factor XIII via the ␥Ј chain (17). We hypothesized that this binding may serve to concentrate factor XIII locally at the surface of the growing fibrin clot, thereby resulting in a more highly cross-linked and stabilized clot. Fibrinolysis assays were therefore carried out using purified ␥A/␥A or ␥A/␥Ј fibrinogen. As seen in Fig. 2, both the rate of clotting and the rate of lysis were significantly decreased in ␥A/␥Ј fibrin clots. Clots made from ␥A/␥Ј fibrinogen in the presence of factor XIII clotted more slowly and subsequently lysed more slowly (Fig. 2, inset). Clots formed from ␥A/␥A fibrinogen in the presence of physiological factor XIII concentrations (10 g/ml) showed a lysis t1 ⁄2 of 72 Ϯ 0 min, whereas the clots formed from ␥A/␥Ј fibrinogen had a t1 ⁄2 of 179 Ϯ 4 min. Clot stability was enhanced further in the presence of supraphysiological concentrations of factor XIII. In experiments where the factor XIII concentration was increased to 100 g/ml, the ␥A/␥A fibrinogen clots showed a lysis t1 ⁄2 of 99 Ϯ 1 min, while the clots formed from ␥A/␥Ј fibrinogen had a t1 ⁄2 of 207 Ϯ 8 min. These data suggest that factor XIII increases the clotting time as well as decreases the fibrinolysis rates of ␥A/␥Ј fibrin. In the absence of added factor XIII, fibrin clots made from ␥A/␥A and ␥A/␥Ј fibrinogen clotted and lysed at similar rates, with a lysis t1 ⁄2 of 50 Ϯ 1 min for ␥A/␥A fibrinogen and 54 Ϯ 1 min for ␥A/␥Ј fibrinogen (Fig. 3), demonstrating that the resistance of ␥A/␥Ј fibrin to fibrinolysis is dependent on the presence of factor XIII. Furthermore, when factor XIIIa was inhibited as it was formed during clotting by the prior addition of N-ethylmaleimide, the fibrinolysis rates of ␥A/␥Ј fibrin clots were similar to clots made without factor XIII (data not shown). These data show that active factor XIII is required for the decreased fibrinolysis rate of the ␥A/␥Ј fibrin.
A D-dimer agglutination assay was performed to determine if the ␥A/␥Ј fibrin clots were more highly cross-linked than the ␥A/␥A fibrin clots. As seen in Table I, the lysed ␥A/␥Ј fibrin clots showed positive agglutination down to a dilution of 1:300, whereas the ␥A/␥A fibrin clots showed positive agglutination down to a dilution of only 1:128. These data show that the ␥A/␥Ј fibrin clots are more highly cross-linked, possibly because of high local concentrations of factor XIII at the clot surface due to binding to the ␥Ј chain. It should be noted that these assays were carried out with ␥A/␥A and ␥A/␥Ј fibrinogen that had been separated on DEAEcellulose and further purified by GPRPC-agarose chromatography. Factor XIII co-purifies with ␥A/␥Ј fibrinogen on DEAEcellulose (17), presumably by binding directly to the ␥Ј chain in ␥A/␥Ј fibrinogen. We have verified that factor XIII co-purifies with ␥A/␥Ј fibrinogen on DEAE-cellulose, but is depleted from ␥A/␥Ј fibrinogen purified further on GPRPC-agarose (Fig. 4). In experiments using fibrinogen purified only on DEAE-cellulose, the ␥A/␥Ј fibrin clots were resistant to lysis in the absence of added factor XIII (Fig. 5), presumably due to the contaminating factor XIII.
To ensure that the increased lysis resistance was not due to factor XIII affecting the fibrinolytic system directly, ␥A/␥A and ␥A/␥Ј fibrin clots were formed with thrombin and then digested with trypsin. The ␥A/␥A fibrin clots had a lysis t1 ⁄2 of 5.7 Ϯ 0.2 min, while the ␥A/␥Ј clots had a t1 ⁄2 of 39.1 Ϯ 1.8 min (Fig. 6). These data suggest that the increased fibrinolytic resistance of the ␥A/␥Ј fibrin clots is due primarily to proteolytic resistance, although inhibition of the fibrinolytic system cannot be ruled out entirely.
␥A/␥Ј Fibrin Clots in Whole Plasma Are Resistant to Fibrinolysis-Fibrinolysis assays were also carried out in whole plasma (32) to ensure that the effect of ␥A/␥Ј fibrinogen on clot lysis was not an artifact of the purified fibrinolysis assay. ␥A/␥A and ␥A/␥Ј fibrinogens were iodinated and added to fibrinogen-deficient plasma obtained from an afibrinogenemic individual. Clotting was initiated by adding thrombin, and samples were assayed daily for 7 days to quantitate the amount of soluble fibrin degradation products (Fig. 7). After 7 days, the clots formed from ␥A/␥A fibrinogen showed complete lysis, with a lysis half-time of 4.3 Ϯ 0.1 days, whereas clots formed from ␥A/␥Ј fibrinogen exhibited no detectable lysis. These results demonstrate that ␥A/␥Ј fibrin clots generated in whole plasma, in which a number of potentially confounding factors are pres-FIG. 2. Fibrinolysis rates of ␥A/␥ fibrin clots are slower than ␥A/␥A fibrin clots. Fibrinolysis assays using 1.25 mg/ml purified fibrinogen, 13.2 NIH units/ml ␣-thrombin, 16 ng/ml tissue-type plasminogen activator, 30 g/ml Lys-plasminogen, and factor XIII were carried out using ␥A/␥A fibrinogen with 10 g/ml factor XIII (q), ␥A/␥A fibrinogen with 100 g/ml factor XIII (f), ␥A/␥Ј fibrinogen with 10 g/ml factor XIII (E), and ␥A/␥Ј fibrinogen with 100 g/ml factor XIII (Ⅺ). Clot turbidity was monitored by absorbance at 405 nm and expressed as percent lysis. Each point is the average of duplicate determinations. Inset, the raw absorbance data at 405 nm, showing both the rate of clotting and rate of fibrinolysis, was graphed versus time.
FIG. 3. Fibrinolysis rates of ␥A/␥A and ␥A/␥ fibrin clots are similar in the absence of factor XIII. Fibrinolysis assays using 1.25 mg/ml purified fibrinogen, 13.2 NIH units/ml ␣-thrombin, 16 ng/ml tissue-type plasminogen activator, and 30 g/ml Lys-plasminogen were carried out using ␥A/␥A fibrinogen (q) or ␥A/␥Ј fibrinogen (E) in the absence of factor XIII. Clot turbidity was monitored by absorbance at 405 nm and expressed as percent lysis. Each point is the average of triplicate determinations. ent (particularly protease inhibitors), are also resistant to fibrinolysis.

DISCUSSION
Fibrinolytic therapy is a widely used tool in the treatment of thrombosis. Fibrinolytic agents such as streptokinase or tissuetype plasminogen activator are used to activate the fibrinolytic system by converting the protease zymogen plasminogen to active plasmin, which then cleaves the fibrin clot into soluble components. Unfortunately, the administration of these agents often produces secondary hemorrhagic events that may be lifethreatening. The development of new fibrinolytic agents with greater clot specificity is an area of intense study, but less attention has been focused on the factors that determine clot stability. If clot stability could be reduced, then lower amounts of these fibrinolytic agents could be used, thus reducing the chance of adverse side effects.
Increased fibrinogen levels are a major risk factor for cardiovascular disease (18,19). There are many potential mechanisms to explain how high fibrinogen levels may contribute to the disease process. One possible role is that high fibrinogen levels produce a hypercoagulable state that favors thrombosis. It is known that the plasma levels of fibrinogen have a direct influence on the amount of fibrin formed in clots (25)(26)(27). The present study suggests two additional mechanisms through which high fibrinogen levels may lead to thrombosis. In agree-  4. DEAE-cellulose-purified ␥A/␥ fibrinogen contains factor XIII. Two g/lane ␥A/␥A or ␥A/␥Ј fibrinogen purified on DEAEcellulose or DEAE-cellulose/GPRPC-agarose was reduced and separated on a 10% polyacrylamide gel. The proteins were transferred to nitrocellulose, the blot was probed with a goat anti-human factor XIII antibody (1:1000) and stained with a rabbit anti-goat IgG/horseradish peroxidase conjugate using diaminobenzidine.  5. DEAE-cellulose-purified ␥A/␥ fibrinogen is resistant to fibrinolysis without added factor XIII. Fibrinolysis assays using 1.25 mg/ml DEAE-cellulose-purified fibrinogen and 13.2 NIH units/ml ␣-thrombin, 16 ng/ml tissue-type plasminogen activator, and 30 g/ml Lys-plasminogen were carried out using ␥A/␥A fibrinogen (q) or ␥A/␥Ј fibrinogen (E). The fibrinogens were purified only on DEAE-cellulose without the use of GPRPC-agarose to remove endogenous factor XIII. Clot turbidity was monitored by absorbance at 405 nm and expressed as percent lysis. Each point is the average of triplicate determinations.
FIG. 6. Cross-linked ␥A/␥ fibrin is resistant to trypsin proteolysis. Clot lysis assays using 1 mg/ml purified fibrinogen, 13.2 NIH units/ml ␣-thrombin, 10 g/ml factor XIII, and 3.9 g/ml trypsin were carried out at room temperature using ␥A/␥A fibrinogen (q) or ␥A/␥Ј fibrinogen (E). Clot turbidity was monitored by absorbance at 405 nm and expressed as percent lysis. Each point is the average of duplicate determinations. ment with previously published data, the results show that fibrinolysis rates in vitro vary inversely with fibrinogen concentration (33). This suggests that in hyperfibrinogenemia, there may not only be a greater propensity for a hypercoagulable state, but the clots that result will also be more resistant to lysis. In addition, the present study shows that ␥A/␥Ј fibrinogen affects the stability of fibrin clots formed in vitro, possibly by increasing the local concentration of factor XIII at the clot. Fibrinolysis assays using purified components revealed that clots formed with ␥A/␥Ј fibrinogen show significantly decreased fibrinolysis compared with ␥A/␥A fibrinogen, but only in the presence of active factor XIII. ␥A/␥Ј fibrin clots are also more highly cross-linked than the ␥A/␥A fibrin clots, as assessed with a D-dimer agglutination assay. The resistance of ␥A/␥Ј fibrin clots to fibrinolysis was verified in reconstituted whole plasma. These results suggest that ␥A/␥Ј fibrinogen levels in plasma may modulate clot stability by affecting fibrinolysis rates.
The physiological function of ␥A/␥Ј fibrinogen has been elusive for many years. It is clear that the ␥Ј chain displays much less binding to the platelet fibrinogen receptor, glycoprotein IIb-IIIa (integrin ␣ IIb ␤ 3 ) (11,29,34), but this negative role does not explain the raison d'être for ␥A/␥Ј fibrinogen. The present studies suggest that the physiological role of ␥A/␥Ј fibrinogen may be to increase clot stability by localizing factor XIII at the fibrin clot (Fig. 8). In this scenario, unactivated factor XIII is delivered to the growing fibrin clot by ␥A/␥Ј fibrinogen. Through the action of thrombin, fibrinogen (including ␥A/␥Ј fibrinogen) is converted to fibrin. Factor XIII that is bound to the newly formed ␥A/␥Ј fibrin is subsequently activated to XIIIa, since fibrin acts as a positive modulator in factor XIII activation by thrombin (3). Thrombin cleavage dissociates the active a 2 dimer from the b subunits of factor XIII, thereby allowing the free a 2 dimer to catalyze the cross-linking of the fibrin clot. It is not clear at the present time whether the b subunits remain bound to the ␥Ј chain or dissociate with the a 2 subunits after thrombin activation. Thus, the delivery of factor XIII to the fibrin clot by ␥A/␥Ј fibrinogen could serve to increase the local concentration of factor XIIIa and result in a more highly stabilized clot.
The concentration of ␥A/␥Ј fibrinogen in plasma may therefore be a risk factor for thrombosis, independent of the total plasma fibrinogen concentration. The molecular mechanisms that affect the ratio of ␥A versus ␥Ј mRNA transcription are unknown, although the liver expresses both mRNAs, while other tissues express only the ␥A mRNA (35). It is clear that the two mRNAs arise by alternative processing of the original transcript (15,16); it is possible that the relative amounts of the mRNA pools or their relative translation rates in liver vary among individuals, giving rise to different amounts of ␥A/␥A and ␥A/␥Ј fibrinogen. There are presently no data available regarding the variability of ␥A/␥Ј fibrinogen levels in human populations. Epidemiologic studies to address this issue are therefore currently underway. FIG. 7. ␥A/␥ fibrin is resistant to fibrinolysis in whole plasma. ␥A/␥A (q) and ␥A/␥Ј (E) fibrinogen were iodinated with Na 125 I and added to fibrinogen-deficient (Ͻ0.1 mg/ml) human plasma at 1.25 mg/ ml. 50 l samples were clotted using 1 NIH unit/ml ␣-thrombin and incubated at 37°C. Samples were taken daily for 7 days and centrifuged 5 min to pellet the clots. Aliquots of the fibrin degradation products in the supernatant were counted with a gamma counter. The percent lysis was calculated as described under "Experimental Procedures." FIG. 8. A model of factor XIII delivery by ␥A/␥ fibrinogen during clotting. Unactivated factor XIII is delivered to the growing fibrin clot carried by ␥A/␥Ј fibrinogen. After thrombin generation, ␥A/␥Ј fibrinogen is converted to fibrin, and factor XIII is activated to XIIIa. Thrombin cleavage dissociates the active a 2 dimer from the b subunits of factor XIII, allowing the free a 2 dimer to catalyze the cross-linking of the fibrin clot. The delivery of factor XIII to the fibrin clot by ␥A/␥Ј fibrinogen serves to increase the local concentration of factor XIIIa and results in a more highly stabilized clot. It is not clear at the present time whether the b subunits remain bound to the ␥Ј chain or dissociate upon thrombin activation of the a 2 dimer.