With interest we read the article by Abha Sahni and coworkers; “The VE-cadherin binding domain of fibrinogen induces endothelial barrier permeability and enhances transendothelial migration of malignant breast epithelial cells” in the edition of August 1st, 2009.1 Sahni et al. showed that addition of intact fibrinogen induces an increase in endothelial permeability. This increase was shown to be mediated via binding of fibrinogen domain β15-42 to VE-cadherin on endothelial cells. Transendothelial migration of malignant breast cancer cells was increased by fibrinogen. The data provide novel insight in transendothelial migration of malignant cells, which require fibrinogen domain β15-42 and VE-cadherin binding.
Sahni et al. showed a significantly increased permeability of human endothelial cells using 300 nM (0.1 g/L) fibrinogen. This is a surprising finding, since this is below the fibrinogen concentrations in the human circulation. In our own experiments, 50 and 200 nM intact fibrinogen showed no increase in permeability of the endothelial monolayer, using highly purified intact fibrinogen fractions (unpublished). Published data on higher fibrinogen concentrations do not show a consensus on whether the intact protein increases vascular permeability. Tyagi et al. showed that at pathologically high levels, fibrinogen (≥12 μM) can increase permeability of the endothelial monolayer, whereas physiological concentrations (6–12 μM) had no effect.2 Here, Sahni et al. showed an induction of endothelial permeability using 300 nM fibrinogen. Referring to the human circulation, it would be surprising if fibrinogen concentrations beneath physiological levels would be able to increase permeability. Therefore, we carefully compared the experiments and proposed mechanism.
The fibrinogen-induced permeability was investigated using peptides and blocking antibodies. Sahni et al. showed an impaired fibrinogen-induced permeability after addition of antibody T2G1 excess. This suggests that the antibody T2G1 bound intact fibrinogen and thereby impaired endothelial permeability. However, Kudryk et al. showed that the monoclonal antibody T2G1 only binds β15-21 on fibrin and its degradation products, whereas the binding to intact fibrinogen was shown to be ∼250 times lower.3 Moreover, peptide β15-42 was shown to mimic fibrinogen-induced permeability, but the control peptide had no effect.1 As stated in the introduction, domain β15-42 is cryptic in fibrinogen and exposed after cleavage of fibrinopeptide B or proteolytic degradation.1, 4 Domain β15-42 is also being exposed after thrombin-independent fibrinogen matrix assembly.5 In the methods of the article, it was described that intact soluble fibrinogen was added to the apical side of the endothelial monolayer. The question rises whether fibrinogen domain β15-42 was exposed and bound to endothelial VE-cadherin to increase permeability. The role of VE-cadherin regulating permeability in the endothelial cell junctions has been extensively studied.6 Moreover, as shown by Ge et al., relatively low concentrations fibrin degradation products increase endothelial permeability. It was shown that 0.5–2.0 μM (0.05–0.2 g/L) fragment D caused a doubling of the transendothelial 125I-albumin clearance in 2 hr.7 Healthy individuals have D-dimer concentrations with a maximum of 0.0005 g/L. Taken together, these findings strongly suggest the presence of fibrin monomers or fibrin degradation products in the fibrinogen sample used by Sahni et al. Alternatively, the exposure of cryptic domains within the fibrinogen molecule by conformational changes may contribute to the exposure of fibrin specific epitopes, such as is also observed after fibrinogen binding to plastic culture plates.
The mechanism proposed by Sahni et al. is that tumor-associated fibrinogen can facilitate transendothelial migration of malignant cells. Intact fibrinogen would increase vascular permeability via binding of fibrinogen domain β15-42 to endothelial VE-cadherin. However, domain β15-42 was shown to be cryptic in intact fibrinogen.3 In our opinion, not intact fibrinogen but fibrin or its degradation products are responsible for the observed effects. This sheds a new light on the data. Palumbo et al. showed that fibrin(ogen) is an important determinant of the metastatic potential of tumor cells. In wild-type mice, a 20-fold reduction of tumor metastasis was seen after addition of the thrombin-inhibitor hirudin.8 In addition to a contribution of fibrinogen to platelet-microthrombi, which may provide protection to tumor cells against innate immune surveillance systems,9 Sahni et al. proposed that tumor-associated fibrinogen can facilitate transendothelial migration of malignant cells. Our addition to the latter proposed mechanism would be that tumor-associated fibrinogen is converted to fibrin monomers or fibrin degradation products. Thereafter, it acts as a bridging molecule between tumor cells and endothelial cells, after which transendothelial migration can take place. The bridging of fibrin to VE-cadherin on endothelial cells is mediated via domain β15-42 on the molecule.
In summary, findings demonstrating that relatively low concentrations of intact fibrinogen increase permeability should be reconsidered. Domain β15-42 is exposed after thrombin cleavage and degradation of fibrin, thereafter the fibrin specific antibody T2G1 can bind. The results shown by Sahni et al. strongly suggest the presence of fibrin specific epitopes, which bear impact on the interpretation of the data. In this view, not intact fibrinogen but fibrin monomers or fibrin degradation products form the bridging molecule between tumor and endothelial cells. Not withstanding, the results showing malignant cell migration mediated by domain β15-42 fibrin are novel and add to the existing models on extravasation of metastatic tumor cells. It would be wonderful when such new insight would result in improvement of therapeutic interventions.