Fibrin network adaptation to cell-generated forces
- 753 Downloads
Fibrin promotes wound healing by serving as provisional extracellular matrix for fibroblasts that realign and degrade fibrin fibers, and sense and respond to surrounding substrate in a mechanical-feedback loop. We aimed to study mechanical adaptation of fibrin networks due to cell-generated forces at the micron-scale. Fibroblasts were elongated-shaped in networks with ≤ 2 mg/ml fibrinogen, or cobblestone-shaped with 3 mg/ml fibrinogen at 24 h. At frequencies f < 102 Hz, G′ of fibroblast-seeded fibrin networks with ≥ 1 mg/ml fibrinogen increased compared to that of fibrin networks. At frequencies f > 103 Hz, G″ of fibrin networks decreased with increasing concentration following the power-law in frequency with exponents ranging from 0.75 ± 0.03 to 0.43 ± 0.03 at 3 h, and of fibroblast-seeded fibrin networks with exponents ranging from 0.56 ± 0.08 to 0.28 ± 0.06. In conclusion, fibroblasts actively contributed to a change in viscoelastic properties of fibrin networks at the micron-scale, suggesting that the cells and fibrin network mechanically interact. This provides better understanding of, e.g., cellular migration in wound healing.
KeywordsOne-particle microrheology Microenvironment Fibroblast Fibrin Mechanosensing Viscoelasticity
Fibrin is a resilient naturally occurring biopolymer and the main structural protein in blood clots, which stops bleeding, and promotes wound healing by serving as a provisional extracellular matrix for cells such as fibroblasts (Laurens et al. 2006; Jansen et al. 2013). Upon vascular injury, the enzyme thrombin converts the plasma protein fibrinogen in fibrin that can withstand forces exerted by flowing blood and by embedded cells (Kroll et al. 1996; Shah and Janmey 1997). Fibrin displays non-linear mechanical properties such as reversible strain-stiffening, whereby the material becomes stiffer as the stress or strain of the material increases, which protects fibrin to further deformation (Roberts et al. 1973; Janmey et al. 1983; Shah and Janmey 1997; Yao et al. 2008; Kang et al. 2009; Wen et al. 2012). Fibrin utilizes irreversible structural adaptations at multiple scales to maintain reversible stress stiffening up to large strains. It adapts to moderate strains at the network scale, whereas it adapts to high strains at the fiber scale. This multiscale self-adaptation mechanism provides fibrin the ability to cope with recurring loads due to shear flows and wound stretching (Kurniawan et al. 2016). A variety of tissue-forming cells such as fibroblasts can sense and respond to mechanical properties of their surrounding substrate by actively realigning and degrading the fibers, locally pulling on the fibers, and at the same time actively changing the stiffness and tension of the surrounding substrate by applying localized forces, so-called traction forces, in a mechanical-feedback loop (Wolf and Friedl 2011; Jansen et al. 2013). These changes in the surrounding substrate enable cells to sense mechanical signals over a distance of several cell lengths and can serve as a signal over a long range to neighboring cells to detect each other (Winer et al. 2009).
Characterization of mechanical properties at the micron-scale is difficult since both network and cell mechanics have complex underlying mechanisms. Rheology is used to measure the viscoelastic properties of fibrin networks using different rheological techniques (Ferry et al. 1997; Jansen et al. 2013; Kurniawan et al. 2016; Piechocka et al. 2016). Ferry et al. (1997) were the first to study the linear viscoelastic behavior of fibrin clots using torsional rheometry (Ferry et al. 1997). Thereafter, significant advances led to new applications of rheological techniques. Rheology has two main techniques, i.e., microrheology and macrorheology. Microrheology has advantages over macrorheology such as much smaller sample volume, a high-frequency bandwidth (0 to 100 kHz), in situ measurement, and higher sensitivity to intracellular dynamics (Mizuno et al. 2008; Tassieri et al. 2012). Microrheology can be classified as passive microrheology, measuring spontaneous thermally driven fluctuations of beads, or active microrheology, driving the beads by an external force (oscillatory optical tweezers or magnetic tweezers (Jones et al. 2015; Wessel et al. 2015)). Passive microrheology by using optical tweezers is non-invasive and can access the high-frequency domain, and can be used to characterize polymer network behavior (Morse 1998). Passive and active microrheology are suitable to obtain information about the local micromechanical properties of extracellular matrix networks at the scale relevant to cells, as well as the network and fiber response (Jansen et al. 2013).
Fibroblasts affect the macroscopic mechanical properties of fibrin (Winer et al. 2009), via active global stiffening of fibrin networks by applying traction forces (Jansen et al. 2013). The strain-stiffening response protects fibrin networks against damage from traction forces exerted by cells (Piechocka et al. 2016). While cell contractile activity is known to occur at the macro-scale, it is also important to know how cell contractile activity affects the local and global mechanical properties of the cell microenvironment and vice versa, the so-called mechanical-feedback loop. Since it is essential to elucidate this mechanical-feedback loop at the micron-scale for a better understanding of three-dimensional cell-matrix mechanical interactions, local matrix stiffness and cell-generated contractile stress are important parameters to study whether strain-stiffening can be induced on a cellular level. Therefore, we aimed to study the mechanical adaptation of fibrin networks due to cell-generated forces at the micron-scale. We hypothesized that fibrin networks adapt to cell-generated forces. We expected that cell-generated forces will increase the micromechanical properties of fibrin network. Fibrin networks and fibroblast-seeded fibrin networks were prepared with 1, 2, or 3 mg/ml fibrinogen polymerized with 0.5 IU/ml thrombin, and 1.0-μm polystyrene beads were embedded. Viscoelastic properties of fibrin networks and fibroblast-seeded fibrin networks were assessed 3 h after polymerization by one-particle passive optical tweezers microrheology.
Materials and methods
Human CCL-224 fibroblasts were cultured in T-75 culture flasks (Nunc, Roskilde, Denmark) in Dulbecco’s modified Eagle medium (DMEM; Gibco, Paisley, UK) with 20% fetal bovine serum (FBS; Gibco, Paisley, UK) and 1% penicillin streptomycin (Gibco, Paisley, UK), and incubated at 37 °C in a humidified atmosphere with 5% CO2. After 24 h, cells were cultured in DMEM with 10% FBS and 1% penicillin streptomycin, and the medium was refreshed every 3–4 days. Upon confluency, cells were harvested by incubation with 0.25% trypsin-ethylenediaminetetraacetic acid (EDTA; Gibco, Paisley, UK). Phosphate-buffered saline supplemented with 20% FBS was added to 0.25% trypsin-EDTA and to stop the trypsin-EDTA reaction, the cells were sequentially washed with phosphate-buffered saline to further remove the trypsin-EDTA, and no EDTA remnant was expected to have any impact on the clotting factor XIII activity. Fibroblasts were replated and cultured until passage 11 to be used in fibroblast-seeded fibrin networks to visualize the dynamics of cell spreading in fibrin networks.
Mouse L292 fibroblasts were cultured to near confluency in T-75 flasks in α-minimum essential medium (α-MEM; Gibco, Life Technologies, Waltham, MA, USA) with 10% FBS and 1% penicillin streptomycin. Fibroblasts were incubated at 37 °C in a humidified atmosphere with 5% CO2, and subcultured until passage 3 to be used in fibroblast-seeded fibrin networks (see below) to perform one-particle passive microrheology.
Fibrin networks and fibroblast-seeded fibrin networks
Human fibrinogen powder (Enzyme Research, South Bend, IN, USA) was dissolved in a buffer of 20 mM sodium citrate-HCl, pH 7.4. For fibrin network preparation, fibrinogen solutions were mixed in fibrin assembly buffer containing 20 mM HEPES (Sigma-Aldrich, St. Louis, MO, USA), 150 mM NaCl, 5 mM CaCl2 (pH 7.4), and 10% culture medium (89% α-MEM, 10% FBS, 1% penicillin streptomycin) for 10 min at 37 °C in a dry bath (Benchmark Scientific, Inc., Edison, NJ, USA). Fibrin networks were prepared with 1, 2, or 3 mg/ml fibrinogen with randomly embedded 1.0-μm-diameter polystyrene beads (Phosphorex, Hopkinton, MA, USA) and polymerized with 0.5 IU/ml human α-thrombin (Enzyme Research, South Bend, IN, USA) in a cell culture dish (diameter 35 mm; SPL Life Sciences Co., Ltd., Gyeonggi-do, Korea). Thirty minutes after the polymerization was initiated, fibrin networks were covered with culture medium and incubated at 37 °C in a humidified atmosphere with 5% CO2 for 3 h.
Fibroblast-seeded fibrin networks were prepared by dissolving 1, 2, or 3 mg/ml human fibrinogen in fibrin assembly buffer (see above). Fibroblasts were resuspended in these three fibrinogen solutions with 1.0-μm-diameter polystyrene beads at a volume fraction of 10% and seeded at 500 cells/μl fibrinogen. Visual inspection by going from the bottom of the fibroblast-seeded fibrin networks to the top part by using a microscope revealed that the cell density was constant at different fibrinogen concentrations. Furthermore, the cell density was not suspected to vary dramatically within a few hours considering that the typical population doubling time of the human CCL-224 and mouse L292 fibroblasts takes 2 to 3 days. Fibroblast-seeded fibrinogen was polymerized with 0.5 IU/ml human α-thrombin in a cell culture dish. Thirty minutes after the polymerization was initiated, fibroblast-seeded fibrin networks were covered with culture medium and incubated at 37 °C in a humidified atmosphere with 5% CO2 for 3 h.
For visualizing the dynamics of cell spreading in fibrin networks, fibroblast-seeded fibrin networks were imaged using a TCM400 inverted bright field microscope (LaboMed America, Inc., Fremont, CA, USA) immediately after seeding, and at 1, 2, and 24 h after cell seeding.
One-particle optical tweezers passive microrheology
One-particle passive microrheology was performed in steady-state conditions 3 h after polymerization (Jansen et al. 2013) with a customized optical tweezers microrheology setup (Fig. 2a). The motions of the embedded 1.0-μm-diameter polystyrene beads within fibrin and fibroblast-seeded fibrin networks covered with CO2-independent medium (Gibco) in cell culture dishes were used for quantitative determination of the microscopic viscoelastic properties of fibrin and fibroblast-seeded fibrin networks with 1, 2, or 3 mg/ml fibrinogen. To demonstrate the cell-generated forces, one-particle passive microrheology was performed for 10-μm polystyrene beads embedded in 1, 2, or 3 mg/ml fibrin networks in cell culture dishes.
Polystyrene beads in the fibrin networks and fibroblast-seeded fibrin networks were weakly optically trapped using a × 100 objective (Zeiss, Thornwood, NY, USA) and a focused laser beam (MATRIX 1064-10-30, Coherent Inc., Santa Clara, CA, USA) with a wavelength of 1064 nm (Koenderink et al. 2006). The motions of the trapped beads were tracked by back-focal-plane interferometry. Thermally driven fluctuations of the micron-sized beads embedded in both types of fibrin networks were detected by a quadrant photodiode with a laser power of 15 mW at a sampling rate of 500 kHz and converted to the complex shear modulus G*, which is composed of the elastic modulus G′, and viscous modulus G″, to determine the viscoelastic properties by using the fluctuation-dissipation theorem (Gittes et al. 1997). The power spectral density of the monitored bead is proportional to the imaginary component of the complex compliance, and by linear response theory, the real component of the complex compliance was calculated via a Kramers-Kronig integration. The complex shear modulus was then calculated as proportional to the inverse of the complex compliance via the generalized Stokes-Einstein relation (Gittes et al. 1997). As a control, one-particle optical tweezers passive microrheology was performed in water using the same micron-sized polystyrene beads.
A customized microrheology optical tweezers setup was used to determine the viscoelastic properties of unseeded and fibroblast-seeded fibrin networks. Spontaneous diffusive motions (thermal fluctuations) of 10 selected 1.0-μm beads, randomly embedded in fibrin and fibroblast-seeded fibrin networks, were measured per condition in the x and y direction at a fixed distance of 10 μm above the glass surface using customized LabVIEW software. The complex response function was determined by using the fluctuation-dissipation theorem to relate the bead motions to the thermal forces on the beads in the frequency domain (Mizuno et al. 2008). The viscoelastic properties as a function of frequency were obtained from the complex response function via the generalized Stokes-Einstein relation (Gittes et al. 1997; Mizuno et al. 2008). The data shown is the mean of five samples per condition calculated from a pool of 50 measurements of bead fluctuations, i.e., from 10 bead fluctuations per sample.
Fibroblasts spread in fibrin networks
Mechanics of fibrin networks and fibroblast-seeded fibrin networks
We studied the mechanical adaptation of fibrin networks due to cell-generated forces at the micron-scale. We found that (i) fibroblasts were round in fibrin networks with 1, 2, and 3 mg/ml fibrinogen within 2 h after cell seeding, while 24 h after cell seeding, fibroblasts were elongated-shaped in fibrin networks with 1 and 2 mg/ml fibrinogen, and cobblestone-shaped in fibrin networks with 3 mg/ml fibrinogen; (ii) at frequencies f < 102 Hz, the elastic modulus G′ of fibroblast-seeded fibrin networks increased by 1.9-fold for 1 mg/ml fibrinogen, 2.0 fold for 2 mg/ml fibrinogen, and 1.6-fold for 3 mg/ml fibrinogen compared to fibrin networks; (iii) at frequencies f > 103 Hz, the viscous modulus G″ of fibrin networks decreased with increasing concentration following the power-law in frequency with exponents ranging from 0.75 ± 0.03 to 0.43 ± 0.03, while G″ of fibroblast-seeded fibrin networks decreased with increasing concentration following the power-law in frequency with exponents ranging from 0.56 ± 0.08 to 0.28 ± 0.06. Therefore, our results showed that fibroblasts generated forces causing microscopic matrix stiffening, suggesting that the cells and fibrin network mechanically interact. Fibroblasts were initially round in morphology, began to adhere to the fibrin fibers, and developed extensions as they changed from a round to an elongated-shaped or cobblestone-shaped phenotype 2 h after cell seeding. Cells embedded in fibrin networks attach and exert traction forces on the fibers of the surrounding network (Notbohm et al. 2015). These forces lead to reorganization of the matrix fibers around them (Notbohm et al. 2015). Reorganization is the first stage of the matrix remodeling process and takes place over a few hours to a few days (Piechocka et al. 2010). Structural remodeling of fibrin networks due to cellular traction forces is a complex process, in part because fibrin networks are not linearly elastic, exhibiting strain-stiffening when subjected to large deformations (Piechocka et al. 2010). The differences found in cellular shape between the dilute fibrin networks with 1 or 2 mg/ml fibrinogen and the denser fibrin network with 3 mg/ml fibrinogen 1 day after cell seeding indicate differences in reorganization. Fibroblasts change cellular shape most pronounced in dilute fibrin networks with fibrinogen concentrations c ≤ 2 mg/ml compared to denser fibrin networks with fibrinogen concentrations 3 ≤ c ≤ 6 mg/ml (Jansen et al. 2013). Cells in dilute fibrin networks are elongated and extend several thin protrusions into the network, while cells in denser fibrin networks remain round and extend short and thin protrusions into the network (Jansen et al. 2013). Since mesh size decreases with the square-root of fibrinogen concentration (Piechocka et al. 2010), the smaller mesh size in dense fibrin networks can constrain the cell body and can impede the extensions of protrusions. Fibrin’s plasticity can direct and control cell morphology and migration (Notbohm et al. 2015). Plastic deformation of fibrin networks leaves tube-like matrix gaps with available space for subsequent growth of cell extension (Notbohm et al. 2015). Another explanation for the differences in cellular shape can be that the dilute fibrin networks are more plastically deformed than the denser fibrin networks, suggesting more available space for cell protrusions. Thus, the structure of the microenvironment around cells in dilute fibrin networks differs from dense fibrin networks, which might lead to altered mechanical properties of the microenvironment. In previous unpublished work, fibroblasts destroyed the fibrin network, possibly by proteolytic activity, after a period of about 16 h in culture. This was visually confirmed by observing large “hole” structures in the fibrin network close to the cells. Since the current study addresses the micromechanical properties of fibrin networks and fibroblast-seeded fibrin networks 3 h after polymerization, further studies are needed to correlate the onset of cell spreading and the onset of gel-stiffening and plasticity. Furthermore, rheology measurements should be performed within a time period much lower than 24 h at which time it is suspected that the fibrin network would not resemble homogeneous and intact structure anymore.
The elastic modulus in the elastic plateau region G0 of semiflexible polymers follows a power-law in concentration with an exponent of 11/5 (MacKintosh et al. 1995; Gardel et al. 2004). In our study, G0 of fibrin networks followed a power-law in concentration with an exponent of 1.5. This exponent is in good agreement with prior measurements on fibrin networks where G0 followed a power-law in concentration with an exponent of nearly 2.0 and 2.3 measured by macrorheology (Jansen et al. 2013; Piechocka et al. 2010). The magnitude of the moduli measured by macrorheology is greater than our measurement. Fibrin networks were prepared under similar fibrin conditions, but in our study, the fibrinogen concentration range was small, i.e., 1 to 3 mg/ml, compared to other studies where the range was 0.1 to 6 mg/ml (Jansen et al. 2013) and 0.1 to 8 mg/ml (Piechocka et al. 2010). The difference in magnitude of the moduli can be explained by the smaller range of fibrinogen concentration used in our study.
The elastic modulus of fibroblast-seeded fibrin networks was 1.6–2.0-fold higher than that of fibrin networks at frequencies f < 102 Hz for all fibrinogen concentration. Prior macrorheology measurements on fibrin networks with and without fibroblasts show a threefold higher G′ for fibroblast-seeded fibrin networks for c ≤ 2 mg/ml, but for fibrinogen concentrations c > 2 mg/ml, the cells do not significantly influence the elastic modulus (Jansen et al. 2013). The difference we observed in G′ between fibroblast-seeded fibrin networks and fibrin networks might be due to higher local traction forces at higher fibrin concentrations corresponding to an increase in elastic modulus at the micron-scale, but not at the macroscopic scale. Microrheology measurements showed that cells stiffened the fibrin networks starting 3 h after cell seeding for higher fibrinogen concentrations.
The viscous modulus G″ of semiflexible networks follows a power-law in frequency with exponent ¾ (Gittes and MacKintosh 1998). In our study, the probable exponential decrease in the exponents of the G″ could be related to the stiffening of fibrin as well as the decrease in mesh size with increasing fibrinogen concentration. The same result was found for fibroblast-seeded fibrin networks where the exponents decreased with increasing fibrinogen concentration. This might indicate that cells do not constrain the formation of a uniform fibrin network, which results in a lower viscous network.
To unravel the mechanical-feedback loop between cell contractile activity and the local and global mechanical properties of their microenvironment, it is also important to take the ability of cells to push against the matrix as well as the nucleus into account. The nucleus also exhibits viscoelastic properties (Guilak et al. 2000). It is tightly coupled to the surrounding cytoskeleton, which in turn is physically connected to the extracellular matrix. Deformations of the nucleus can have important consequences for cellular functions, such as gene expression, resulting in defective nuclear cytoskeletal coupling and defects in cytoskeletal organization and stiffness, thereby affecting tissue function. Cells align fibrin fibers by pulling on the surrounding matrix, and compact fibrin fibers by pushing against the matrix, to invade the matrix and create permanent deformation of the matrix (Notbohm et al. 2015). The magnitude of pulling and pushing forces is roughly of the same order, but the location of these forces in the matrix changes over time (Notbohm et al. 2015). Pulling forces appear along the leading edges of cellular protrusions, whereas pushing forces often appear near the cell body and vary with time near the tip of a growing cell protrusion (Notbohm et al. 2015). This indicates that compressive forces are present within the cell which might be the result of nuclear deformation. Future studies are needed to investigate the mechanism underlying how cells push against the matrix.
In summary, cellular shape is both fibrinogen concentration- and time-dependent. The incorporation of fibroblasts into fibrin networks increases the elastic moduli and decreases the viscous moduli of fibrin networks, which might drive the fibrin networks into a non-linear stress-stiffened state. Therefore, our results show that the presence of fibroblasts in fibrin networks actively contribute to a change in viscoelastic properties of fibrin networks at the micron-scale, suggesting that the cells and fibrin network mechanically interact. This contributes to a better understanding of biologically relevant processes, such as cellular migration in infection, inflammation, immunology, and wound healing.
The authors thank Johannes Boonstra (Utrecht, The Netherlands) for providing the fibroblasts.
This work was supported by the University of Amsterdam, Amsterdam, The Netherlands (grant for the stimulation of the research priority area “Oral Regenerative Medicine”). Rommel G. Bacabac was funded by the Philippine Council for Industry, Energy and Emerging Technology Research and Development – Department of Science and Technology (PCIEERD-DOST), Taguig City, Philippines (grant-in-aid funding for the “Coils, Cells & Gels” program), 2015 and the “BINARY GELS” PCIEERD Project No. 0431-, 2017), and received logistic support from the Research Office of the University of San Carlos, Cebu City, Philippines.
- Ferry B, Antrobus P, Huzicka I, Farrell A, Lane A, Chapel H (1997) Intracellular cytokine expression in whole blood preparations from normals and patients with atopic dermatitis. Clin Exp Immunol 110:410–417. https://doi.org/10.1046/j.1365-2249.1997.4361452.x CrossRefGoogle Scholar
- Kroll MH, Hellums JD, McIntire LV, Schafer AI, Moake JL (1996) Platelets and shear stress. Blood 88:1525–1541Google Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.