henhaobio
Sign in

Acta BiomaterialiaBiophysicsPeer-reviewed journal article

Nonlinear Relationships of Fibrin Network Structure as a Function of Fibrinogen and Thrombin Concentrations for Purified Fibrinogen and Plasma Clots

Can CaiiD, Zezhong Zhang, Arezoo NamenyiD, Daniel Nelli, Keith BoniniD, Brittany E. BannishiD, Nathan E. HudsoniD, Glen S. MarrsiD, Stephen R. BakeriD, Martin GutholdiD

PDFDOIPreprint
In-depth full-text digest available · Jump to digest ↓

Abstract

BackgroundFibrinogen levels are associated with bleeding disorders and thrombotic disease. Thrombin converts fibrinogen to fibrin, producing the load-bearing fibrin scaffold that governs clot mechanics and transport. ObjectiveQuantitatively map how initial fibrinogen and thrombin concentrations, [Fgn]0 and [T[h]r]0, determine fibrin architecture in human plasma and a purified fibrinogen system. MethodsScanning electron microscopy was used to quantify single-fiber morphology--fiber diameter and branch-to-branch segment length from a standardized sample-preparation protocol. Confocal microscopy was used to quantify network architecture--projected fiber density and pore/bubble size. Results and ConclusionsAcross plasma and purified systems, fibrin structural parameters were quantitatively captured by compact multiplicative scaling laws of the form Y = k[Fgn]0[T[h]r]0{beta}. Unlike prior studies, which examined narrower condition ranges without establishing predictive equations across a systematic fibrinogen-thrombin concentration matrix, this framework defines distinct, quantitative roles for fibrinogen and thrombin in fibrin assembly. The magnitudes and signs of the exponents and {beta} indicate that thrombin primarily controls individual fiber growth kinetics, strongly shortening branch-to-branch segment length and modestly thinning fibers, whereas fibrinogen primarily controls space filling, strongly increasing fiber density, reducing pore/bubble size, and thickening fibers. For matched [Fgn]0 and [T[h]r]0, compared to plasma clots, purified fibrinogen formed denser networks with thinner and shorter fibers, suggesting that the plasma biochemical environment partially inhibits thrombin activity. Fiber length analysis further suggests that each thrombin molecule nucleates one fiber segment. Together, these parameterized scaling relations provide a predictive quantitative framework linking clot composition to fibrin microstructure in plasma and purified fibrinogen clots.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 3 (29 Jul 2026) · CC BY 4.0

Fibrin fiber diameter, segment length, density and pore size in plasma and purified clots follow fibrinogen-thrombin power laws; thrombin sets segment length, fibrinogen sets space filling.

Published fibrin structure-concentration relationships are hard to compare because studies differ in imaging resolution, preparation and functional forms and often vary one factor at a time, leaving no transferable composition-to-structure mapping for testing assembly or calibrating simulations. Here a multiplicative power law is fitted to four structural metrics across fibrinogen-thrombin matrices in plasma and purified clots, assigning distinct roles to each protein.

Key results

  1. ShownOne multiplicative power law, Y = k[Fgn]0^α[Thr]0^β, fitted in log space, describes SEM fiber diameter and segment length plus confocal fiber density and bubble diameter in both plasma and purified fibrinogen clots. The fitted parameters stay stable when the data are resampled.

    Plasma 4×4 matrix (0.36–2.9 mg/mL fibrinogen, 0.05–1.0 U/mL thrombin) and purified 3×3 matrix; 16 SEM images and 36 confocal z-stacks per condition. 2000-iteration bootstrap 95% CIs sat close to the point estimates, and leave-one-condition-out refits shifted α and β only modestly (Tables S5, S6). · Curve fitting and statistical analysis; Power law fits quantify how fibrinogen and…; Scope and robustness of the scaling analysis

  2. ShownThrombin mainly sets single-fiber growth. Raising thrombin strongly shortens branch-to-branch segments and thins fibers, while fibrinogen barely changes segment length.

    Doubling thrombin shortens segments 15% in plasma (2^-0.24 = 0.85) and 18% in purified clots (2^-0.29 = 0.82). The fibrinogen exponent for length is α < 0.1 (5–7% lengthening per doubling). Thrombin β for diameter is −0.23 (plasma) and −0.21 (purified), or −14–15% per doubling, versus +11–14% per fibrinogen doubling. · Fiber diameter (SEM); Fiber length (SEM); Kinetic control (thrombin) vs space-filling…

  3. ShownFibrinogen mainly controls space filling. Raising it increases projected fiber density and shrinks pores, outweighing thrombin's effects on both. The bubble and density exponents are linked by a near-geometric packing relation.

    Doubling fibrinogen raises fiber density 37% in plasma (α = 0.45) and 66% in purified clots (α = 0.73), and cuts bubble diameter 15% and 25%. Thrombin changes bubble diameter 9% per doubling in both systems, and its density exponent is β = 0.30 (plasma) and 0.43 (purified). Bubble/density fibrinogen exponent ratios are −0.53 and −0.56. · Fiber density and pore size (confocal); Kinetic control (thrombin) vs space-filling…

  4. SuggestedSingle-fiber exponents are similar in plasma and purified clots, suggesting shared assembly; density and bubble exponents agree when fit on identical 3×3 matrices. Prefactor k differs: purified clots form denser networks of thinner, shorter fibers, plausibly from lower effective thrombin activity in plasma.

    Plasma vs purified prefactor differences: ∼1.28× (diameter), ∼1.49× (length), ∼1.30× (bubble diameter), and ∼2.26× in the opposite direction for fiber density. Applying a rescaling T_eff = sT0 to SEM diameter and length gives s ≈ 0.19–0.35, i.e. ∼65–81% lower effective thrombin in plasma. · Purified and plasma clots show distinct…; Conserved fibrinogen and thrombin concentration…; Kinetic control (thrombin) vs space-filling…

  5. SuggestedMeasured segment lengths match the spacing between thrombin molecules. This suggests each thrombin molecule (or a small cluster) nucleates about one fiber segment, which would explain why length depends on thrombin rather than fibrinogen.

    With 1 NIH U/mL = 9 nM, thrombin density is 0.271–5.42 μm^−3 (0.05–1 U/mL). Pooled Lseg ranges fall in the same micrometer range as the cube side-to-diagonal spacing and shrink with thrombin in both systems (Table 3). At 0.05–0.1 U/mL each per-thrombin volume holds 10³–10⁴ fibrinogen molecules, which can be activated within minutes at ∼10⁷ M⁻¹ s⁻¹. · Why is fiber segment length mainly dependent on…; Tables

Methods, data and reproducibility

Plasma (4×4) and purified fibrinogen (3×3) clots spanning fibrinogen and thrombin matrices were imaged: standardized SEM gave fiber diameter and branch-to-branch length, confocal z-stacks gave projected fiber density and bubble pore size, and each metric was fit as Y = k[Fgn]0^α[Thr]0^β by log-space ordinary least squares with bootstrap intervals.

  • Pooled citrated normal human plasma (George King Bio-Medical #0010) clotted over a 4×4 fibrinogen (0.36–2.9 mg/mL) by thrombin (0.05–1.0 U/mL) matrix · human · n = 4×4 fibrinogen-thrombin conditions
  • Peak 1 purified human fibrinogen (Enzyme Research Laboratories P1 FIB) clotted over a 3×3 fibrinogen (0.36–1.45 mg/mL) by thrombin (0.05–0.2 U/mL) matrix · human · n = 3×3 fibrinogen-thrombin conditions
  • SEM micrographs at 30,000× of fixed, hexamethyldisilazane-dried, gold-coated clots; at least 70 fibers/image for diameter and at least 30 for segment length · human · n = 16 images per condition (two biological replicates × eight)
  • Confocal z-stacks (212 μm × 212 μm × 15 μm, 33 sections at 0.47 μm steps) of hydrated clots containing 6.5% Alexa Fluor 488-fibrinogen · human · n = 36 z-stacks per condition (18 per slide, two slides)
  • Turbidity assays testing label perturbation in plasma and purified fibrinogen at 0.36, 0.72, 1.45 and 2.4 mg/mL with 0, 1.5, 3, 6 and 12% labeled fibrinogen · human
  • Unlabeled plasma clots imaged at identical 488 nm settings as a background fluorescence control (no network signal detected) · human
Methods & tools
Scanning electron microscopy (Zeiss Gemini 300, 30,000× magnification), Cressington Model 108 gold sputter coater, Laser-scanning confocal microscopy (Zeiss LSM 710, 40× 1.3 NA oil objective), ImageJ/Fiji (grid-sampled manual diameter and segment-length tracing; maximum projections), Line-probe intersection counting for projected fiber density, Custom MATLAB bubble (largest inscribed circle) pore analysis after Molteni et al., Log-space ordinary least squares power-law fitting, Image-level bootstrap resampling (2000 iterations, 95% CIs), Leave-one-condition-out refitting, Turbidimetry, Alexa Fluor 488-conjugated human fibrinogen (Thermo Fisher F13191), Michaelis–Menten activation-with-depletion kinetic rationale (Supplement S1)
Reproducibility gaps
The paper cites no code or data repository. The custom MATLAB bubble-analysis code and the image-level measurements are not linked; per-condition values and leave-one-out fits appear only in supplementary tables.
Limitations (6)
  • Confocal fiber density and bubble size are resolution-limited trends, not absolute 3D values. Fiber overlap in the densest networks can undercount density and merge pores; this forced the purified matrix down to 3×3 and may weaken the plasma density-fibrinogen scaling. · stated by the authors
  • Alexa Fluor 488-fibrinogen at 6.5% of total fibrinogen may subtly perturb assembly. Turbidity showed minimal effects up to 6% labeling but clear deviations at 12%, and star-like aggregates appeared above about 6.5%. · stated by the authors
  • The scaling laws are supported only within 0.36–2.9 mg/mL fibrinogen and 0.05–1 U/mL thrombin and are untested outside that range. Clots formed at room temperature with CaCl₂ fixed per system (20 mM plasma, 9.5 mM purified), so the exponents apply only to these conditions. · stated by the authors
  • No rheology was measured, so the authors do not predict bulk modulus from the scaling relations. SEM preparation (washing, dehydration, drying, coating) can also distort apparent fiber and network geometry. · stated by the authors
  • SEM and confocal clots used different buffers (50 mM Tris/100 mM NaCl vs 20 mM Tris/150 mM NaCl) and different vessels. The volume-fraction estimate still combines SEM diameters and lengths with confocal densities, so ionic-strength differences could confound αV and βV. · noted in this summary
  • The one-thrombin-per-segment idea rests on overlap between pooled length ranges and nominal thrombin spacing, not on a fitted exponent test. It also uses nominal plasma thrombin, although the authors themselves estimate 65–81% lower effective thrombin in plasma. · noted in this summary

Read this if you model fibrin assembly, clot mechanics or transport and need fibrinogen- and thrombin-dependent scaling laws for fiber diameter, segment length, network density and pore size.

Citations
1
FWCI
4.71
Field percentile
top 5%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISCaton '18Helms '24Zohravi '25Rieu '20Jansen '20Nameny '26
69% semantic matchBiophysics
F. Caton, X. Garcia, L. Seyve+4 more·bioRxiv·Oct 16, 2018