Fibrin formation on fast forward

Fibrin gel, a 3-dimensional network of fibers constituting the scaffold of the hemostatic plug, is rapidly formed following fibrinogen activation during blood coagulation. In this issue of Blood, Chernysh and Weisel provide the first direct glimpse of the fibrin polymerization process under near-physiological conditions using deconvolution microscopy.

Fibrin formation has been intensely studied for more than 60 years using a variety of experimental techniques,^1,2  but the mechanisms of some fundamental processes, such as branching and lateral aggregation, are poorly characterized, and a detailed description of many steps in this very complex process is still lacking. The relatively recent crystal structures of many fibrin(ogen) fragments have contributed to the elucidation of basic intermolecular interactions,³  and snapshots of the process have been provided by high-resolution electron microscopy, which has shown many architectural details (reviewed in Weisel² ). However, electron micrographs are taken on dehydrated samples, and hardly cover the full sequence of events. On the other hand, time-resolved bulk techniques such as light scattering and turbidity have generated and constrained useful models of polymerization (see De Spirito et al⁴ ; reviewed in Weisel² ), but they are intrinsically limited by their averaging nature. This situation may change thanks to recent advances in light microscopy,⁵  which can combine the power of directly counting objects in their physiological medium with adequate spatial and time resolution.

In the meantime, in this issue of Blood, Chernysh and Weisel present an innovative study using an established but lesser-known technique, deconvolution microscopy, to image plasma clotting after spiking it with fluorescently labeled fibrinogen. Relying on a mathematical treatment of images to remove out-of-focus light from other parts of the sample so that a clear picture of a single focal plane can be obtained, this technique affords real-time monitoring of the process while avoiding the bleaching problems and the relative slowness that affect confocal microscopy. Quantitative analysis of the evolving gel parameters, like fiber lengths and branch points, can then be carried out. Furthermore, although the technique's spatial resolution does not allow the direct measurement of fiber diameters, by measuring the fluorescent intensity changes along the forming fibers, their mass can also be monitored and compared with the results of parallel turbidimetry experiments.

Examining the paper, it is difficult to resist first having a peek at the supplemental videos (available on the Blood website). At low magnification (Video S1, 10× objective), nothing apparently goes on for the first 10 seconds (corresponding to ∼12.5 minutes reaction time), and then the 3D network rapidly blooms in flower-like fashion, pervading the entire field in the last 6 seconds (∼7.5 minutes). The fibers seem to radiate from a few apparently unevenly distributed nodes until they come in contact, but it must be borne in mind that, in this later phase, we are mainly observing their thickening, the underlying scaffold being almost invisible at this magnification. Indeed, at higher magnification (Video S2, 63× objective), tantalizing details emerge. After 10 to 12 seconds (corresponding to ∼3-4 minutes, in excellent agreement with the clotting time and the start of turbidity rise) fibers begin to be visible, some already fixed in space, some wandering around until later on they make contact with others, but even then whole sections of the network are mobile for some time before they become frozen. This dynamic behavior is a new, important finding, and should be properly taken into account in future models of fibrin formation.

Incidentally, a rough calculation of the diffusion coefficient expected for rigid (proto)fibrils 10 to 15 μm long and 10 to 20 nm thick yields approximately 10 to 20 μm² per minute, in the same ballpark as the observed movements. Intriguing quantitative analyses of the data are then presented by the authors, revealing details that could not be inferred from bulk techniques alone. In particular, while this study definitively confirms that the abrupt transition in the turbidity profiles is directly linked to lateral thickening, there is still a fraction of fibers that continue to grow in length and branch after the gel point, contradicting some models of fibrin polymerization predicting that the entire scaffold is laid down by the clotting time. Whether these features are directly linked to the system used to initiate the reaction (citrated plasma plus CaCl₂ and tissue factor), mimicking the physiological situation, or are an intrinsic property of fibrin formation, must be elucidated in further studies. Waiting for the sequel, we are left with plenty of new evidence that fibrin assembly is a truly dynamic process requiring continued research and more detailed modeling than previously thought.