The biophysics and mechanics of blood from a materials perspective

Abstract

Cells actively interact with their microenvironment, constantly sensing and modulating biochemical and biophysical signals. Blood comprises a variety of non-adherent cells that interact with each other and with endothelial and vascular smooth muscle cells of the blood vessel walls. Blood cells are further experiencing a range of external forces by the hemodynamic environment and they also exert forces to remodel their local environment. Therefore, the biophysics and material properties of blood cells and blood play an important role in determining blood behaviour in health and disease. In this Review, we discuss blood cells and tissues from a materials perspective, considering the mechanical properties and biophysics of individual blood cells and endothelial cells as well as blood cell collectives. We highlight how blood vessels provide a mechanosensitive barrier between blood and tissues and how changes in vessel stiffness and flow shear stress can be correlated to plaque formation and exploited for the design of vascular grafts. We discuss the effect of the properties of fibrin on blood clotting, and investigate how forces exerted by platelets are correlated to disease. Finally, we hypothesize that blood and vascular cells are constantly establishing a mechanical homeostasis, which, when imbalanced, can lead to hematologic and vascular diseases.

Introduction

Blood comprises trillions of cells that are pumped by the heart to circulate in the blood vessels throughout the body. Therefore, blood and vascular cells are constantly exposed to a hemodynamic microenvironment involving a range of external forces distinct from other tissue types. From a macroscale perspective, the mechanical properties of many blood-related components, such as blood pressure, fluid shear stress, blood viscosity, the stiffness of blood vessels as well as blood cells and clots, remain relatively stable in healthy tissue, suggesting that the key components of the circulatory system – vessels, blood and blood clots – are maintained in a state of mechanical equilibrium. Like many adherent cells, blood and vascular cells contain a pre-stressed cytoskeletal structure and a mechanotransductive machinery to sense, respond and modify the microenvironment¹. Coordinated mechanosensitive and mechanoresponsive behaviour enables cells to provide regulatory feedback to the blood system. Therefore, we hypothesize that in a healthy circulatory system, a mechanical homeostasis is maintained at the cellular and tissue level.

Here, mechanical homeostasis is defined as the process that maintains the mechanical equilibrium of a biological system using negative feedback mechanisms. This concept of mechanical homeostasis has been demonstrated in other types of adherent cells, such as fibroblasts and mammary epithelial cells^{2, 3}, and their associated tissues, such as skin and mammary gland^{3, 4}, which can maintain normal physiological conditions against intra- and extracellular forces and deformation. Alterations of the mechanical properties of blood cells and vascular tissues can be linked to the pathogenesis of numerous cardiovascular and hematologic diseases, including pro-inflammatory vascular conditions, such as sickle cell disease and bleeding disorders, and uncontrolled blood clotting in atherosclerosis and/or stroke ^4–9. Therefore, the specific mechanical properties of blood cells and tissues, such as blood vessel stiffness, blood cell contraction forces and vascular and blood cell stiffnesses, can be potentially used as biomarkers for diagnosing cardiovascular and hematologic diseases. Importantly, the mechanical disequilibrium associated with many of these diseases can be targeted as a treatment strategy ^{10, 11}.

However, a comprehensive understanding of the mechanical homeostasis of blood and vascular tissues remains elusive thus far. Particularly, a quantitative characterization of the mechanical dynamics in the hemodynamic microenvironment at the cellular and molecular level is difficult. Thus, how alterations of cell–cell and cell–extracellular matrix (ECM) interactions can result in pathophysiology at the tissue and organ level is not yet understood. Materials-based techniques offer the possibility to characterize the mechanical dynamics at micro- and nanoscales, allowing the identification of mechanical biomarkers and therapeutics for haematological and vascular diseases. Moreover, sophisticated and smart materials could be used as tools to engineer in vitro models that better recapitulate the in vivo mechanical microenvironment for studying mechanical equilibrium states.

In this Review, we discuss techniques for measuring the mechanical properties of blood tissues and materials that can be used to recreate their mechanical microenvironment. We then examine the mechanical homeostasis hypothesis in three distinct anatomical regions: blood vessels, blood and blood clots, highlighting key findings and important knowledge gaps that need to be filled to validate the hypothesis. Finally, we investigate exciting future areas of research at the crossroads between materials science and the study of blood tissues and disease.

Tools to measure mechanical properties

To study the mechanics of blood tissues across different length scales (FIG. 1a), tools and techniques are required that operate with measurement resolutions at broad length scales, ranging from nm to cm, and at broad force scales, ranging from pN to N^{8, 9, 12, 13}.

Figure 1: Tools to measure and recreate the mechanical properties of blood tissues.

a) The schematics compares the length scale of blood cells and tissues with the resolution of techniques for the investigation of the mechanical properties of blood cells and tissues. Typical measurements, values, and extracted values are also shown. b) Schematic of a 2D cell culture on a protein-conjugated hydrogel surface. c) Schematic of a 3D cell culture. Cells are encapsulated inside a hydrogel and can interact with the surrounding environment. d) Hydrogels can be used for the engineering of microvasculature by vasculogenesis (green: fibrin gels embedded with perivascular cells; purple: fluidic channels; blue: fibrin gels embedded with endothelial cells; red: microvascular network formed via vasculogenesis²⁰⁸. e) Hydrogels can also be applied for the engineering of microvasculature by angiogenesis, using vascular endothelial growth factor (VEGF)²⁰⁹.

Bulk mechanical properties

The characterization of bulk mechanical properties of tissues provides information on the correlation of tissue mechanics and the pathogenesis of diseases and aging processes. For example, stiffening of blood vessels has been identified as a mechanical biomarker of aging⁶ and diseases such as atherosclerosis^{7, 14}. The bulk mechanical properties of blood vessels can be measured using tensile testing, in which mN to N forces are directly applied to the vessel wall, generating a strain–stress curve to characterize the elastic modulus of the vessels^{12, 15}. Arterial stiffness is also clinically used by physicians and can be indirectly characterized by measuring the pulse wave velocity using a non-invasive approach. Thereby, the speed at which the elastic distortion of the vessel wall propagates is measured in response to the blood pressure pulse in the systole–diastole cycle^{16, 17}. In addition, the compliance of blood vessels can be measured by analyzing the percentage of vessel radius increase using ultrasound in the clinic, which corresponds to the increase in intramural pressure¹⁸. The alteration of other bulk mechanical properties, such as blood viscosity and clot viscoelasticity, could also be associated with disease states. These properties can be measured by viscometry¹⁹ and rheometry²⁰.

Some specialized tools have been developed specifically for measuring the bulk mechanical properties of blood. These tools have generally been optimized for use in a clinical setting. Thromboelastography bears some similarity to rheometry and measures clot mechanical properties over time, from initiation to lysis. In this tool, activated blood fills the space between two moveable surfaces. Initially, only one surface is moved back and forth in a steady function. As blood coagulates and solidifies, the motion of the moving surface is transferred to the other surface, which is free to move. By measuring the movement of the second plate and the timing of this movement, measurements about the clot initiation time (in s), clot maximum firmness (in mm), and clot lysis time can extracted ^{9, 21, 22}. However, note that the clot firmness is given in terms of mm of movement of the second plate, not Pa as seen in rheometres. Another system, created by Hemodyne, provided information about clot stiffness (Pa) and bulk platelet contraction forces (0 – 0.2N) ²³, although this tool is no longer available.

Mechanical properties at the micro- and nanoscale

The bulk mechanical properties of tissues can provide useful information for disease diagnosis and therapies; however, mechanical homeostasis is achieved through the interaction between cells and the extracellular environment at the micro- and nanoscale^{5, 24, 25}. Therefore, these scales need to be probed to understand how the mechanical properties at the cellular and molecular levels cause the alteration of bulk mechanics in disease. For example, the stiffness of cells and ECMs can be tested at the microscale using atomic force microscopy (AFM). An atomic force microscope can be used to measure the length of cells at the nm to μm scale and force at the pN to nN scale, and to characterize the mechanics of compliant objectives, including the stiffness of ECMs and cells^{12, 26}.

Cells generate forces at the single cell and the receptor level. The forces exerted by adherent cells can be assessed by traction force microscopy (TFM)²⁷ and micropost array detectors^{2, 28} by measuring the displacement of beads embedded in hydrogels with a known elastic modulus or of microposts with a pre-determined stiffness. We recently developed a technique called contraction cytometry, in which adhesive protein microdots are conjugated to a hydrogel of known stiffness. The contractile forces generated by single platelets can then be measured by characterizing the change of distance between a microdot pair²⁹. However, these techniques only provide nN sensitivity, which is not sufficient to determine the forces transmitted by individual receptors at the pN scale. AFM, magnetic or optical tweezers³⁰, and biomembrane force probes³¹ can be used to measure forces with pN sensitivity, but these techniques are limited by low-throughput. Molecular tension fluorescence microscopy (MTFM) has recently been developed to map the spatiotemporal forces of single receptors across the cell surface at the pN scale and with high throughput, by measuring the fluorescence emission of immobilized MTFM probes³². Using these techniques, cell and tissue mechanics can be quantified across a broad length scale to investigate the multiscale nature of mechanical homeostasis and the disturbance of mechanical equilibrium in hematologic and vascular disease.

Mimicking the microenvironment

To investigate cell behaviour in vitro, it is important to recreate a physiologically-relevant microenvironment that contains the biochemical and biomechanical cues which modulate cell behaviour in vivo³³. The incorporation of appropriate ligands into microenvironments is paramount, because by binding to specific receptors at distinct kinetics, ligands trigger specific downstream signalling pathways. For example, immobilized von Willebrand factor (vWF) and fibrinogen control platelet adhesion at high and low shear stress through binding to distinct platelet receptors^{34, 35}.

Biomechanical cues, such as stiffness, regulate cellular tensegrity, affect cell behaviours and can be recapitulated using materials (TABLE 1). For example, polyacrylamide (PA) gels can be used to study the effect of ECM stiffness on cell behaviour^36,37. The stiffness of this synthetic gel can be fine-tuned by controlling the crosslinking density to represent the stiffnesses of different tissues. PA hydrogels have been used for studying various types of cells, including endothelial cells³⁸ and platelets^{39, 40}, which play major roles in hematologic processes and disease. However, PA gels can only serve as 2D models owing to the incapability of biodegradation and cell encapsulation (FIG. 1b). By contrast, naturally-derived collagen and fibrin gels can be applied to investigate 3D cell behaviours (FIG. 1c); for example, white blood cell migration⁴¹, platelet contraction³⁹ and angiogenesis^42–44. However, biophysical properties, such as ligand density, ECM elasticity and porosity, are difficult to decouple in naturally-derived gels. Moreover, natural ECM-derived gels are too soft to be tuned to a broad stiffness range. Alternatively, alginate gels can be used to address these limitations. In alginate gels, ligand density can be easily controlled by titrating adhesive domains. In addition, the rigidity of the gel can be tuned by modulating the physical and chemical crosslinking density^{45, 46}. Fluid shear stress is also an important mechanical cue, which can be recapitulated using macro- and microfluidics (FIG. 1d-e). Traditional fluidics are made of polydimethylsiloxanes (PDMS)^47–50; however, fluidic systems can also be made of soft materials, such as collagen^{42–44, 51}, fibrin^52–54, alginate⁵⁵ and agarose⁵⁶, which allow the tuning of stiffness and permeability, making these systems more physiologically relevant.

TABLE 1.

Hydrogels for the investigation of blood tissues

Materials	Pros	Cons	Applications	Refs
Polyacrylamide	• Well characterized for a wide physiologically-relevant range of stiffness (hundreds of Pa – hundreds of kPa); • Conjugation of adhesive proteins on the surface	• Only suitable as 2D model • Not suitable for cell embedding • not degradable • Swelling ratio is not controllable	• Cell adhesion, spreading and differentiation • Traction force microscopy • Contraction cytometry	36–40,100
Collagen and fibrin	• Natural proteins that contain adhesive domains for cells • Suitable for cell encapsulation and 3D study • Degradable • Cell migration and angiogenesis studies • Easy to incorporate into microfluidic systems	• Soft and restricted stiffness range (< a few kPa) • Difficult to decouple stiffness and ligand density • Fibrin gel formation depends on thrombin and is thus difficult to control	• Microvasculature-on-a-chip by top-down and bottom-up approaches • Cell migration • Platelet contraction in 3D (fibrin gel)	39,41–44,51–54
Gelatin-methacrylate (MA)	• Denatured collagen contains adhesive domains for cells • Wide range of compressive modulus (hundreds of Pa to ~ 35 kPa) • Suitable for cell encapsulation and 3D study • Degradable • Easy to incorporate into microfluidic systems	• Difficult to decouple stiffness, ligand density and porosity • Degradability is compromised with increasing MA ratio • Swelling ratio is coupled to crosslinking and ligand density	• Microvasculature-on-a-chip by top-down and bottom up approaches	99–101
Agarose-gelatine interpenetrating network	• Wide stiffness range Gelatine contains adhesive domains for cells • Easy to incorporate into microfluidic systems• Swelling is negligible	• Not degradable • Not suitable for cell migration and angiogenesis studies • Requires post-crosslinking for the fabrication of microfluidics	• Microvasculature-on-a-chip by top-down approach with well-controlled microvascular size scale and flow dynamics • Long-term culture with physiologically-relevant endothelial permeability	106
Chemically modified alginate	• Covalent and/or physical crosslinking • Ligand density can be decoupled from stiffness, crosslinking density and porosity • Suitable for cell encapsulation and 3D studies • Can be incorporated into microfluidic systems	• Only degradable when modified with oxidization • Crosslinking may be affected by calcium in the medium	• Microfluidic channels • Cell encapsulation in 3D	45,46,55
Chemically modified polyethylene glycol	• Ligand density can be decoupled from stiffness, crosslinking density and porosity • Can be used for cell encapsulationin 3D and angiogenesis • Can be incorporated into microfluidic systems	• Depends on incorporation of degradable segments for degradability • Uncontrolled swelling	• Microvasculature-on-a-chip by top-down and bottom-up approaches	102–104
Desired hydrogel materials	• Contains appropriate ligands or ligands are easy to incorporate • Easy to adjust stiffness to a wide range • Decoupled stiffness, ligand density and porosity • Negligible swelling • Degradable • easy cell encapsulation • Easy to apply to microfluidic systems and to connect with tubing and pumps		• Microvasculature-on-a-chip by top-down and bottom-up approaches with well-controlled size and flow dynamics • Long-term culture with physiologically-relevant endothelial permeability • Cell encapsulation in 3D • Cell migration

Blood vessels

Mechanosensitive barriers between blood and tissues

Blood vessels form a tree-like, pressurized circuit that circulates blood between the heart and other organs. The blood vessel wall contains mechanically responsive cells, including endothelial cells and vascular smooth muscle cells that actively respond to the hemodynamic environment, making the vessel wall a mechanosensitive barrier. Exertion of mechanical forces between these cells and between cells and the ECM as well as feedback mechanisms provide mechanical homeostasis in blood vessels. In healthy tissue, vascular cells maintain a specific mechanical microenvironment and dysregulation of this mechanical homeostasis leads to cardiovascular pathology, highlighting the importance of mechanical homeostasis in vascular physiology ^{4, 57}. Therefore, understanding mechanical homeostasis and the key mechanical forces in blood vessels will greatly improve the engineering of blood vessels and vascular grafts.

The size, structure, cellular components and mechanical properties of blood vessels vary in the different tissues of the body depending on their function. For example, arteries carry pulsatile blood away from the heart as it contracts and generates pressure. As blood moves distally from the heart, the pressure drops along the vessel and the blood flow becomes less pulsatile. Eventually, the vessels branch into smaller arterioles and capillaries, which converge into larger venules and veins that return blood back into the atria (FIG. 2a)⁵⁸. With the exception of capillaries, all blood vessels possess an intima, which is the innermost layer comprising of a sheet of flattened endothelial cells that rest on a thin layer of connective tissue, a layer called the media, which consists of spindle-shaped vascular smooth muscle cells that are embedded in a matrix of elastin and collagen fibres, and an adventitia, which is the outermost layer composed mainly of fibroblasts and collagen matrix⁵⁹ (Fig. 2b). By contrast, capillaries only contain an intima that is covered with pericytes.

Figure 2: Mechanical forces and mechanical homeostasis in blood vessels.

a) The graphs show the distribution of blood pressure and blood velocity across a blood vessel. b) The tree-like vasculature has a heterogenic structure in arteries, microvasculature and veins²¹⁰. c) A negative feedback mechanism regulates mechanical homeostasis in blood vessels. Flow shear stress triggers signalling pathways in endothelial cells, which leads to cell proliferation or apoptosis and has an effect on vascular smooth muscle cells. d) Similarly, vascular smooth muscle cells respond to an increase in vessel wall stress by contracting or by producing collagen. e) Pulse wave velocity is slower in healthy arteries compared to stiff arteries (modified based on Ref. 210), which is related to the probability of cardiovascular diseases, including atherosclerosis and stroke. f) Disturbed blood flow occurs in the bifurcation region of arteries, making them atherogenic. g) Two vortical flows form in the venous valve sinus owing to its geometry, which increases the risk of thrombosis. h) In the early stage of plaque development, outward remodelling occurs to maintain a constant low flow shear stress and lumen size. i) Inward growth of a vessel wall at a late stage of plaque development results in a local change of the flow shear stress. Elevated shear stress appears upstream of the bulge, while disturbed flow with low oscillatory shear stress occurs downstream of the plaque. j) Computational modelling demonstrates that inward growth of the vessel wall can lead to higher wall stress on the shoulder of plaque, which is more vulnerable to rupture⁶¹.

Each blood vessel type is structurally and mechanically different, but they all are exposed to various hemodynamic forces, including flow shear stress, frictional forces parallel to the vessel wall caused by flowing blood, and circumferential stress perpendicular to the vessel wall, generated by transmural pressure^60–62. Vascular smooth muscle cells ⁶³ and endothelial cells⁶⁴ are sensitive to these physical forces; changes in force can trigger intracellular signalling pathways, which regulate cellular function and the structure of blood vessels.

Flow shear stress

The ability of endothelial cells to respond to physical perturbations and maintain the flow shear stress in a narrow physiological range by modulating the vascular lumen size was first reported decades ago^{57, 65}, leading to the hypothesis of mechanical homeostasis in blood vessels^{3, 57}, which can be regulated by vascular cells and endothelial cells through a negative feedback loop. For example, endothelial cells can sense the acute elevation of flow shear stress. They respond by producing nitric oxide (NO) and prostacyclin to relax the vascular smooth muscle cells⁶⁶, which causes an increase in vessel diameter and an decrease in flow shear stress. Thereby, endothelial cells can regulate the flow shear stress⁶⁷ (Fig. 2c). If the flow shear stress remains too high, endothelial cells induce the remodelling of the vessel wall, which can last for weeks or months, to widen the vessel lumen. Such changes in vessel wall and lumen diameter have been observed in ischemic injuries, to bypass the internal blockage of large arteries by arterialization of the collateral capillary and to carry more blood into the tissue⁶⁸. Conversely, as the flow shear stress decreases, endothelial cell signalling causes a decrease in blood vessel diameter and thus, an increase in flow shear stress⁶⁵.

Endothelial mechanotransduction and adaptive feedback signalling mechanisms^{5, 58, 69} are controlled by mechanosensitive protein machineries, including ion channels, G proteins, the apical glycocalyx, primary cilia, the platelet endothelial cell adhesion molecule (PECAM)-1/vascular endothelial (VE)-cadherin/vascular endothelial growth factor receptor (VEGFR) complex, integrins and NOTCH1, which are activated by apical flow shear stress. Downstream signalling pathways then lead to the production of NO and prostacyclin as well as endothelial cell proliferation and barrier function^{5, 70–72}. However, the physiological range of flow shear stress differs in arteries, the microvasculature and veins. Therefore, the questions remain how the mechanosensors in the different vessel types respond to and regulate different levels of shear stress and whether distinct blood vessel types contain distinct mechanosensors or whether they express different levels of mechanosensitive proteins^{72, 73}.

Vessel stiffness

Vascular smooth muscle cells also play a crucial role in maintaining a constant wall stress through a negative feedback mechanism. According to LaPlace’s law (S = P x r / w), the blood vessel wall stress (S) is proportional to the transmural pressure (P) and the radius of the blood vessel (r), and inversely proportional to the wall thickness (w)⁵. To keep the wall stress constant with increasing transmural pressure, vascular smooth muscle cells contract, which causes a decrease in vessel radius. By contrast, a chronic increase in transmural pressure, for example in patients with hypertension (high blood pressure), leads to the secretion of more collagen fibres by vascular smooth muscle cells, which causes thickening of the vessel wall and consequently, a decrease in wall stress^{6, 74} (FIG. 2d). However, the compliance of blood vessels is primarily established by elastin, which is a highly elastic protein of the connective tissue that is only minimally expressed after birth⁵⁹. Therefore, the elastin content does not change much throughout life, and thus, the deposition of collagen fibres by vascular smooth muscle cells is sufficient to cause a decline in vessel compliance and an increase in vessel stiffness.

Blood vessel stiffening is also observed during aging. Aging leads to the breaking of elastin fibrils, which become disordered and weak owing to repetitive strain injury⁷⁵. To compensate for the weaker vessel wall, vascular smooth muscle cells deposit more collagen fibres, which stiffens the vessel wall⁶. The stiffness of blood vessels can be clinically analyzed by measuring the pulse wave velocity, which is proportional to stiffness. For example, an inverse relationship between blood vessel stiffness and age has been demonstrated by reproducible pulse wave velocity analysis in humans¹². Children at the age of 10 have a pulse wave velocity of ~6.5 m/s, whereas the pulse wave velocity is ~11 m/s in adults at the age of 50. Moreover, the pulse wave velocity continuously increases with age. A threshold of carotid-femoral pulse wave velocity of >12 m/s has been suggested as an indicator of significant alterations of aortic functions and cardiovascular diseases¹² (FIG. 2e). In addition, the irreversible age- and hypertension-associated vessel stiffening leads to an increase in the resistance of the systemic circulation and thus, in the required cardiac work, which is associated with many cardiovascular diseases, for example, heart failure⁶.

Flow patterns and plaque formation

In addition to the magnitude of shear stress, endothelial cells are also sensitive to flow patterns, which play a crucial role in initiating dysregulation of mechanical homeostasis and in the development of atherosclerosis. Endothelial mechanosensors can respond to physiological laminar flow, which triggers the expression of transcription factor Kruppel-like factor (KLF2) and nuclear factor (erythroid-derived 2)-like (NRF2), as well as the downstream expression of anti-inflammatory, antithrombotic and antioxidative mediators⁷⁶. By contrast, disturbed flow patterns, including slow, oscillatory and turbulent flow around bifurcations, bends (FIG. 2f) and valves ^{77, 78}(FIG. 2g), inhibit the activation of KLF2 and NRF2 and continuously activate inflammatory pathways involving nuclear factor-κB (NF-κB)^{62, 79}. Of note, in such atheroprone regions, antioxidant, protective pathways are also upregulated, which has been shown by gene expression analysis of endothelial cells ^{80, 81}.

However, the upregulation of protective pathways is insufficient to balance the activated inflammatory pathways^{80, 81} resulting in a net chronic inflammatory phenotype in atheroprone regions. Although the chronic inflammation is mild, it increases vessel permeability and predisposes arteries to the risk of atherosclerosis at bifurcations and bends (BOX 1). Additional risk factors, such as high low-density lipoprotein (LDL) cholesterol and hyperglycemia⁵, promote the accumulation of apolipoprotein-B (apoB) containing lipoproteins (LPs) in these inflamed regions with higher permeability and slow shear stress. The retention of lipoproteins aggravates the inflammation and increases the expression of vascular cell adhesion protein (VCAM)-1 and intercellular adhesion molecule (ICAM)-1, which leads to the recruitment of monocytes that differentiate into macrophages. Lipid and macrophage accumulation eventually cause the formation of a plaque.

BOX 1. Mechanical imbalance and disease.

Atherosclerosis is a chronic disease that occurs in the focal regions of medium-sized arteries. The focal regions are either bends or bifurcation branches that are exposed to disturbed flow. This flow pattern can cause chronic inflammation of endothelial cells and thus, recruitment of monocytes and subendothelial retention of apolipoprotein B (apoB)-containing lipoproteins (LPs), resulting in plaque formation. Some atherosclerotic lesions undergo partial resolution by forming a fibrous cap, which prevents direct contact of the plaque with prothrombotic factors. Some atherosclerotic lesions develop a vulnerable plaque with a necrotic core and a thin fibrous cap, which results in the remodelling of the inner arterial wall, leading to a local change in flow shear stress. At regions with high flow shear stress, the size of the fibrous cap decreases owing to an increase in matrix metalloproteinase production, matrix degradation, intraplaque haemorrhage, accumulation of macrophages and microcalcification. The plaque ruptures once the wall stress exceeds the strength of the weak fibrous cap, causing life-threatening heart attacks and stroke.

Sickle cell disease stems from a single nucleotide polymorphism leading to the expression of haemoglobin that assembles and polymerizes in low oxygen conditions, resulting in a banana-like or sickle-like shape and the stiffening of red blood cells. A small fraction of red blood cells also become irreversibly sickled owing to repeated cycles of haemoglobin polymerization and depolymerization. Red blood cell membranes become fragile, which leads to cell lysis and the release of free haemoglobin, causing further inflammation. Red blood cells are also more adhesive than in healthy tissues, and thus, can adhere to endothelial and other cells, causing small occlusions. The increase in the number of platelets, increased cell adhesiveness and a proinflammatory environment can lead to stroke and undetectable microstrokes. However, defining the progression and frequency of these microstrokes remains elusive thus far.

Bleeding and thrombosis can be linked to the mechanical properties of blood clots. Plasma clots that are 2350% stiffer (3.1 vs 2.1 kPa) are associated with thrombotic disorders¹⁵⁰. Whole blood clots measured using rotational thromboelastography, which provides measurements analogous to clot elasticity, are much softer in patients with bleeding disorders (10 mm vs 60 mm). Fibrin formation and clot contraction contribute to the stiffness of blot clots. For example, bleeding is associated with low fibrin formation, which is a symptom of patients with haemophilia. In thrombosis, blood clots have highly branched thin fibrin fibres that are resistant to lysis¹⁵². Changes or impairment of clot contraction, volume or forces are also related to various diseases, including acute ischemic stroke ¹⁸⁹, acute uremia²⁰⁷, diabetes mellitus ²⁰⁷ and coronary artery disease ²³. Moreover, patients missing highly contractile platelets and with average contractile forces lower than 25nN correlate with symptomatic bleeding²⁹.

In most cases, early progression and remodelling of a plaque follows a negative feedback mechanism: outward remodelling of vessel matrix ensures maintenance of the lumen size and low shear stress^{82, 83}. However, persistent low shear stress accelerates plaque development through promoting inflammation and matrix degradation, eventually augmenting endothelial dysfunction^{84, 85} and leading to the formation of a large necrotic core (FIG. 2h). Once vessel matrix outward remodelling can no longer compensate plaque growth, the plaque begins to grow into the lumen⁶⁰. This inward lumen bulge substantially alters local shear stress and wall stress. Computational models have shown that high shear stress appears upstream of the bulge, low oscillatory shear stress occurs downstream of the plaque, and wall stress is higher at the shoulder regions of the plaque than the rest areas of the plaque^{60, 86} (FIG. 2i-j). In general, plaques are more vulnerable at the shoulder exhibiting high wall stress⁸⁷ and at the upstream region showing high shear stress⁸⁸ than at regions with low shear and wall stress. High shear stress increases matrix metalloproteinase production and thus, matrix degradation, initiates vascular smooth muscle cells apoptosis, induces intraplaque haemorrhage and promotes macrophage accumulation and microcalcification, causing the formation of thin and weak fibrous caps^{60, 61}. Ultimately, once the wall stress exceeds the strength of the fibrous cap, the plaque ruptures and can cause life-threatening heart attacks and stroke.

Vascular grafts

Mechanical homeostasis are not only important in vascular physiology, but also for determining the long-term success of engineered vascular grafts. Vascular grafts are vascular substitutes that are widely used in patients with peripheral arterial disease, coronary artery disease and end stage renal disease⁸⁹. Autografts are currently the gold standard in the clinic; however, they are often not available. Alternatively, synthetic grafts, such as expanded poly(ethylene terephthalate) (Dacron) or poly(tetrafluoethylene) (ePTFE), have been used since the mid-1970s⁹⁰. However, these synthetic grafts are difficult to re-endothelialize and therefore, have been limited to applications in arteries with diameters >6 mm and to high flow and low resistance conditions. Although these synthetic materials have a suitable strength, their lack of appropriate compliance can lead to their failure⁹¹, because incompatible compliance can result in extra stress and disturbed flow patterns at the suture lines, which subsequently causes endothelial dysfunction, thrombosis, neo-intimal hyperplasia and anastomotic aneurysms^{90, 91}. Moreover, synthetic grafts cannot undergo adaptive remodelling, and thus, cannot promote a state of mechanical homeostasis. Therefore, their stiff mechanical properties and low compliance lead to high resistance in the circulation, which can eventually cause various cardiovascular complications ⁹⁰.

Tissue-engineered vascular grafts can overcome these issues. These vascular grafts can be engineered by assembling sheets of autologous cells cultured in vitro ⁹² or by fabricating a tubular ECM mesh from allogeneic smooth vascular muscle cells which are first grown on a rapidly degrading tubular polyglycolic acid scaffold in a bioreactor⁸⁹ and then removed to render the grafts nonimmunogenic. Before implantation, these tissue-engineered vascular grafts have comparable strength and compliance (~3.5% per 100 mm Hg) to native vessels (human saphenous vein: 0.7–1.5 % per 100 mm Hg; human internal mammary artery: 11.5 ± 3.9 % per 100 mm Hg ⁸⁹). They further allow fast endothelialization, infiltration and migration of native vascular smooth muscle cells and endothelial cells, and adaptive ECM remodelling after implantation ^{89, 92}. Therefore, tissue-engineered vessels provide excellent patency and resist fatigue-induced aneurysms, dilatation and calcification. Tissue-engineered vascular grafts have been successfully applied in patients with end-stage renal disease^{93, 94} and they show great potential for coronary bypass procedures⁸⁹, highlighting the importance of matching the mechanical properties of grafts with those of native vessels and of re-endothelialization and remodelling by the host cells.

Microvasculature

Mechanical homeostasis has mainly been studied in arteries; however, the concept of mechanical homeostasis can also be considered for the microvasculature. The microvasculature is composed of a network of small vessels, including arterioles, capillaries and venules, and forms a barrier responsible for the exchange of fluid and plasma metabolites in organs. Increased shear stress can induce capillary growth in skeletal muscle, whereas reduced flow leads to capillary retraction ^{95, 96}. Capillary density is regulated by angiogenic factors, such as VEGF, which are induced by parenchymal hypoxia⁹⁷. Whether a mechanical negative feedback process cooperates with angiogenic factors to modulate capillary density is difficult to study in vivo owing to the challenges associated with decoupling flow forces from angiogenic and other factors, such as changes in nutrient exchange.

Microfabrication techniques can be used to create microvasculature models in vitro that recapitulate microvascular geometry and that can be endothelialized. These models can be applied to address questions related to negative feedback mechanisms, because flow and nutrient exchange can be independently controlled^{43, 49, 98}. Microfluidic devices with microvascular geometry can be fabricated using solid polymers, such as PDMS⁴⁹ and polyester⁹⁸, and more physiologically relevant hydrogels, such as collagen^42–44, fibrin^{52, 53}, alginate⁵⁵, gelatin^99–101 and polyethylene glycol (PEG)^102–104 (TABLE 1). Top-down fabrication approaches, including 3D printing¹⁰⁵, spatial laser degradation¹⁰² and soft-lithography-based layer-by-layer assembly¹⁰⁶, can be used to create a variety of microfluidic geometries, on which endothelial cells can easily grow and form a monolayer. Microvasculature models can further be engineered in hydrogel-based materials by bottom-up approaches, for example, in situ vasculogenesis^{52, 53} (FIG. 1d) and angiogenesis^42–44 (FIG. 1e). In these hydrogel-based microfluidic vascular models, physiological flow forces, such as interstitial shear stress and intraluminal flow shear stress, can be recreated.

Using microvasculature in vitro models, it has been demonstrated that interstitial flow directs endothelial morphogenesis and sprout formation⁹⁷. Moreover, a flow shear stress threshold for angiogenesis⁴⁴ has been proposed: a shear stress of > 10 dyn/cm² triggers matrix metallopeptidase 1-mediated endothelial sprouting towards the direction of draining interstitial fluid flow, and the subsequent higher microvessel density reduces flow shear stress, suggesting a mechanical negative feedback mechanisms governing angiogenesis. Conversely, halted flow causes microvessels to retract. In vitro microvasculature models also enabled the identification of a NOTCH1-mediated mechanosensitive pathway in endothelial cells ¹⁰⁷. Activation of NOTCH1 signalling by laminar shear stress leads to the formation of a multicomplex unit that regulates cytoskeletal remodelling and barrier functions.

In vitro models further allow the incorporation of substrates with physiological or pathological stiffnesses which can be accurately measured by performing AFM measurements^{6, 108}. Using such models, it has been shown that integrins mediate the response of endothelial cells to the stiffness of the abluminal substrate, and that the endothelial permeability increases as substrates stiffen via a Rho-associated signalling pathway¹⁰⁸. Stiff substrates induce excess activation of Rho, causing an increase in intracellular contractile forces. These forces pull cell–cell junctions apart, resulting in a leaky endothelium¹⁰⁸. Since leaky endothelium is a hallmark of atherogenesis, it suggests that age-associated blood vessel stiffening is a direct cause of atherosclerosis development. Using an agarose–gelatin interpenetrating network, tuned to a physiological stiffness of ~ 20 kPa, we showed that an engineered microvasculature can exhibit a physiological endothelial barrier function for months¹⁰⁶. However, most in vitro models only contain endothelial cells. Incorporating other cell types, such as pericytes and vascular smooth muscle cells, would enable the investigation of feedback loops and of crosstalk between the different cell types; for example, by using hydrogel materials, such as collagen, fibrin, and methacrylated gelatin that allow encapsulation of vascular smooth muscle cells and pericytes and that support long-term culture of endothelial monolayers. Importantly, the hydrogels need to be compatible with microfabrication techniques, provide appropriate mechanical properties and allow remodelling by cells to recreate mechanical homeostasis.

Blood

A flowing biomaterial

The concept of mechanical homeostasis can also be applied to blood, although the feedback mechanisms may be less complex and easier to perturb. The material properties and the mechanical equilibrium of single blood cells as well as of a collective of blood cells, contribute to the material properties and behaviour of blood, and can be related to blood pathologies.

Blood is a dynamic shear-thinning biofluid composed of a variety of proteins, small molecules, gases and mechanosensitive cells, including red blood cells, white blood cells and platelets, which are suspended in plasma (FIG. 3a). The size, shape and stiffness of blood cells remain uniform and are regulated by cytoskeletal proteins, ion pumps and the nucleus. Micropipette aspiration¹⁰⁹, microfluidic deformation assays¹¹⁰, optical tweezes¹¹¹ and atomic force microscopy¹¹² have been used to measure the stiffness of blood cells, which range from 100’s of Pa¹¹² to several kPa¹¹⁰. The cytoskeleton of blood cells influences the stiffness of the cells, which has been demonstrated by pharmacologically impairing or genetically modifying cytoskeleton proteins, such as spectrin, which is unique to red blood cells, actin¹¹³, microtubules and intermediate filaments¹¹⁴. Remarkably, each blood cell type has its own unique composition of cytoskeletal proteins influencing shape and stiffness; for example, discoid red blood cells uniquely rely on spectrin (FIG. 3a), spheroid white blood cells rely mainly on actin and intermediate filaments(FIG. 3b), and resting platelets contain an actin cytoskeleton with a prominent microtubule ring (FIG. 3c).

Figure 3: Material properties of individual blood cells.

Blood is a shear thinning fluid composed of cells suspended in a plasma. Each blood cell type is supported by specific cytoskeletal structures resulting in different shapes and material stiffness.^211–213 a) The electron micrograph shows a red blood cell with a cytoskeleton mainly composed of a spectrin mesh and actin, which significantly contributes to this cell’s shape and exceptional deformability. b) The electron micrograph shows a white blood cell with a cytoskeleton mainly composed of actin and intermediate filaments. c) The electron micrograph shows platelets, which possess microtubules, and F-actin that rapidly disassembles and reassembles upon platelet activation. d) Margination is triggered by physical interactions between cells of different stiffness and the vessel wall, and is a purely passive phenomenon¹¹². e) Margination leads to an increase in the concentration of white blood cells and platelets travelling along the periphery of the vessel. f) Blood viscosity as a function of shear stress is shown. The bulk material properties of blood, such as viscosity (shown), change in response to local shear stress rates²¹⁴. g) Haematocrit and applied shear stress change the viscosity as measured with a viscometer. Many parameters, such as the number of cells ²¹⁵, plasma volume and cell stiffness affect the bulk properties of blood. Changes to these parameters are often associated with pathological conditions.

Defining the mechanisms that regulate single cell stiffness is challenging owing to the complexity of cells in terms of protein composition, regulation and dynamics as well as cellular interactions. However, there is evidence that the dynamic rearrangement of structural components contributes to the regulation of the overall stiffness ¹¹⁵. For example, actin¹¹⁶ and spectrin¹¹⁷ are able to change behaviour in response to force^{116, 118}. Unfortunately, deciphering the mechanical feedback of cytoskeletal networks remains difficult, owing to a lack of tools capable of examining nanoscale network formation during mechanical perturbation, and a lack of studies on other cytoskeletal and nuclear proteins, especially intermediate filaments. The nucleus also plays a dominant role in defining the mechanical properties of single cells. For example, red blood cells, which do not possess a nucleus, have an average stiffness of 150 Pa¹¹², whereas neutrophils, which have lobed nuclei, have an average stiffness of 400 Pa¹¹². Correspondingly, lymphocytes, which have a large central nucleus, have a stiffness on the order of 2900 Pa¹¹⁰. The nuclear proteins lamin A and C are important regulators of nuclear mechanics¹¹⁹, and lamin A is involved in mechanosensitive differentiation as shown in hematopoietic stem cells ¹²⁰ and mesenchymal stem cells¹⁰¹. However, the underlying feedback mechanisms associated with nuclear stiffness¹²¹, especially in blood cells, remain elusive.

Mechanical properties of red blood cells

Healthy red blood cells have tightly controlled biophysical properties, and perturbations to the stiffness and shape of red blood cells are associated with impaired function, suggesting the existence of a state of mechanical equilibrium that is maintained in healthy tissue ¹²². Red blood cells have a discoid shape, maximizing their surface area to volume ratio, and a deformable soft cytoskeleton that is resistant to lysis and fracture, enabling them to pass through constricted spaces within the vasculature. Red blood cells do not possess a nucleus and have spectrin-based cytoskeletal architecture that provides less rigidity and is more deformable than other blood cells. Therefore, red cell mechanics influenced by the haemoglobin content membrane rigidity, hydration levels and ion pump function also impact cell stiffness and viscosity – parameters that are tightly regulated.

Spectrin is the major structural protein of red blood cells (FIG. 3a)¹²². Moreover, a variety of other cytoskeletal proteins, such as myosin IIA¹²³, are incorporated into the spectrin network, affecting cellular structure. Owing to their specific cytoskeletal architecture, red blood cells are substantially softer than other blood constituents and able to undergo elastic deformations. The ability to maintain membrane integrity while undergoing deformation is crucial for the transport of oxygen through reactive haemoglobin, which is encapsulated in red blood cells. Free haemoglobin cannot be transported in blood, because it would cause acute and chronic vascular disease, inflammation, thrombosis and renal impairment ¹²⁴. Therefore, pathological changes to the cytoskeletal structure that result in impaired membrane integrity are associated with multiple diseases that arise from different origins. For example, hereditary spherocytosis involves mutations to cytoskeletal proteins¹²² and sickle cell disease indirectly leads to changes in the cytoskeletal structure, yet both diseases disrupt the red cell membrane and are associated with haemolysis. The cargo encapsulated within red blood cells has an impact on the mechanical properties of red blood cells. Cargo type, concentration and structure affect the mechanical properties of the cell, which, if dysregulated, can cause pathological effects. For example, in patients with sickle cell disease¹²⁵ (BOX 1), mutated haemoglobin polymerizes in low oxygen conditions leading to an increase in cell stiffness and reduced deformability. In addition, as the red cell continuously travels between low and high oxygen concentrations, the fluctuating mechanical loads to the membrane applied by polymerizing haemoglobin contribute to the aforementioned haemolysis¹²⁶. Similarly, in the early stages of malaria infection, red blood cells that contain the parasite become stiffer ^{127, 128} and more adhesive, allowing them to stick to the wall of small blood vessels and evade detection by the immune system¹²⁹.

Hydration levels also substantially impact the mechanical properties of red blood cells. For example, red blood cells contain Piezo1, a mechanically sensitive ion channel, which connects mechanical force to red blood cell volume ¹³⁰. Piezo1 enables prolonged calcium entry in response to mechanical stretch and its dysfunction has been correlated to hereditary xerocytosis^131–133. In murine red blood cells, this gain-of-function leads to excess calcium import, which activates the KCa3.1 Gardos channel, leading to potassium export, and thus, dehydration of the cells ¹³⁰. Severe dehydration in adults and especially in neonates is associated with an increased risk of thrombosis^{134, 135}, which may be related to the corresponding increase in red blood cell stiffness and viscosity; however, the exact underlying mechanisms remain unclear thus far. Even small changes in the microenvironment can affect red blood cell transport in the microvasculature; for example, the tonicity of a clinical intravenous fluid affects sickle red blood cell adhesion to endothelial cells and transport through capillary size microchannels¹¹. Therefore, consistent material properties and thus, mechanical homeostasis, seem to be an important factor for healthy red blood cell function.

Emergent properties of the blood cell collective

Margination

Blood can also be considered as a collective of cells with emergent behaviours that arise from the mechanics of single cells and their interactions. For example, when traveling through a vessel, stiff leukocytes and small platelets marginate and predominantly travel near the periphery of the blood vessel. This effect is mediated by the differences in cell size, shape and stiffness of the different blood cell types¹³⁶. Indeed, margination is facilitated by an asymmetry in cellular collisions and by lift forces generated on deformable cells ¹¹². When a soft red blood cell and a stiff white blood cell collide, the soft cell primarily deforms and the stiff cell primarily displaces (FIG. 3d). Importantly, this displacement can push stiff cells across streamlines and towards the vessel wall (FIG. 3e). After many collisions, lift forces generated at the vessel wall have a smaller effect on stiffer white blood cells than on softer red blood cells¹¹², which helps to trap the cells at the periphery (FIG. 3f). Owing to their high stiffness and small size, platelets also travel near the vessel periphery similar to white blood cells ¹³⁷.

Therefore, the process of margination depends on the mechanical stiffness and size differences between single blood cells, which thus contribute to an emergent collective cell behaviour. There is further evidence that the mechanical properties of the blood cell collective are also regulated on a macroscopic level by the kidneys and spleen. The kidneys regulate plasma volume, red blood cell number¹³⁸ and electrolyte levels, which influence single cell stiffness and contribute to blood viscosity (FIG. 3g)¹³⁹. As the primary regulators of sodium and water excretion, the kidneys receive feedback on plasma volume by the hormone-based renin-angiotensin-aldosterone system and by the nervous system through baroreceptors, which can sense vessel pressures that are too high or too low¹³⁸. In addition, peritubular fibroblasts in the kidneys are transcriptionally regulated by oxygen levels, and respond to low oxygen by producing erythropoietin, which stimulates red blood cell production in the bone marrow¹⁴⁰, and therefore, changes the overall blood viscosity. Thus, the kidneys are involved in the regulation of the number of red blood cells and the plasma volume. The spleen further helps to control cell stiffness by mechanically filtering red blood cells that show reduced deformability^{141, 142}.

Effect on haemostasis and the immune system

A change in the collective cell number may also impair haemostatic function^{126, 132}. For example, abnormally high numbers of white blood cells can lead to leukostasis, a medical condition in which cells block blood flow, which can cause a transient ischemic attack and a stroke ¹⁴³. Similarly, high numbers of red blood cells can promote blood clotting by increasing platelet margination, which increases platelet accumulation at injury sites^{137, 144}. Although the mechanisms are unclear, thrombotic complications are a known concern for patients with polycythemia vera, which is a type of blood cancer, in which the bone marrow produces too many red blood cells¹⁴⁵. A low number of red blood cells can also be expected to influence the margination process and possibly lead to more bleeding; indeed, anemia is a predictor for bleeding in patients with atrial fibrillation and in patients taking anticoagulants ¹⁴⁶. However, studying the correlation between anemia and bleeding in a clinical setting is difficult, because excessive bleeding can also lead to low red blood cell counts, and therefore, confound measurements.

In addition to cell number, the material properties of cell collectives can impact haemostatic and/or immune function. For example, it has been demonstrated that an increase in red blood cell stiffness interferes with leukocyte adhesion¹⁴⁷. We further showed that endothelial cell permeability¹⁰⁶ increases with increasing red blood cell stiffness. Similarly, a systemic increase in white blood cell stiffness is associated with many pathological conditions, such as acute lung injury, sepsis, posttraumatic shock, diabetes and stroke¹⁴⁸. We demonstrated that a decrease in leukocyte stiffness can be correlated with a decrease in margination, a decrease in the time leukocytes reside in capillary beds, which increases the number of circulating leukocytes¹¹² and may have important implications for immune cell trafficking and susceptibility to infection.

Various blood pathologies can be associated with altered collective cell behaviour, which is modulated by the stiffness, deformability, number and interactions of the different blood cell types as well as by the plasma volume. Therefore, we hypothesize that an optimal mechanical equilibrium of the blood cell collective exists to ensure hemostasis and immune function, which is regulated by feedback mechanisms on the single cell level and macroscopically by the spleen and liver. However, more research is required to validate the concept of mechanical homeostasis and to decipher regulatory mechanisms in blood.

Blood clots

Blood clots are naturally occurring, self-assembled, active materials composed of contracting platelets, which pull on a nascent polymeric fibrin mesh that may contain entrapped red and white blood cells. Blood clots maintain vascular integrity by creating a plug capable of withstanding the forces applied by flowing blood. Soft clots are associated with bleeding¹⁴⁹ and stiff clots are associated with thrombotic disorders¹⁵⁰, suggesting that an optimal mechanical stiffness is required to facilitate healthy haemostasis (BOX 1). Although the mechanistic link between clot function and stiffness is still being elucidated, these findings allow for the hypothesis that blood clots must be stiff enough to prevent bleeding yet soft enough to not cause unhealthy blood clotting. In both conditions – pathological clotting and excessive bleeding – the mechanical equilibrium of platelets ^{23, 29} and/or of fibrin is perturbed ^{151, 152}. Moreover, the material properties of blood clots change in disease.

Fibrin structure and formation

Blood clots contain a network of fibrin (FIG. 4a), which is a fibrous biopolymer that reinforces the clot. Fibrin is enzymatically formed from the soluble plasma protein fibrinogen in response to injuries in the vasculature. The concentration of fibrinogen in blood (1.5 – 4 g L⁻¹) is an order of magnitude higher than that of other plasma proteins. Fibrin has interesting mechanical properties^153–160; single fibrin fibres are elastic¹⁵⁴, possess extraordinary extensibility¹⁶¹ and a stiffness that changes with diameter¹⁶², from ~1.5 MPa for a 1.4 nm fiber to ~250 kPa for a 2.2 nm fiber. Fibrin mechanics have been measured using numerous techniques including atomic force microscopy¹⁶³, optical tweezers¹⁵⁴, rheometers ¹⁶⁴ and universal testing machines¹⁵³. However, fibrin typically forms viscoelastic¹⁶⁴ networks that stiffen when stretched^{153, 165} and align in response to applied loads¹⁵³ or fluidic forces¹⁵⁸. In addition, fibrin can self-assemble into a sheet at the air–blood interface in vitro and in vivo, which can encapsulate blood clots and protect against microbial invasion¹⁶⁶. Fibrin formation and destruction are primarily controlled by thrombin and plasmin; however, biophysical cues also play a role in the regulation of fibrin network formation. For example, platelet contraction ¹⁶⁷, fibre diameter and fibre strain impact fibrin lysis by plasmin¹⁶⁸ and thus, blood clot lysis. Single molecule experiments have further demonstrated that the lifetime of fibrin–fibrin bonds increase, peak and decrease with increasingly high forces, suggesting that mechanical homeostasis may be established by a force-based regulatory mechanism^{169, 170}.

Figure 4: Blood clots.

Blood clots may contain red blood cells and white blood cells in addition to platelets and a fibrin mesh, which polymerizes from the plasma protein fibrinogen after tissue damage. a) The electron micrograph shows the fibrin matrix on the exterior of a blood clot formed in vitro. b) Platelets (red) can be thought of as actuators of clotting, contracting fibrin (green) triggered by local biochemical and biomechanical stimuli. Yellow indicates platelets and fibrin colocalization. c) Upon activation, platelets and nascent fibrin form into a gel structure. d) Contractile platelets can dramatically reduce the size of a clot. e) Blood clots can self-assemble in vivo into platelet-rich (CD41, red) structures with a layer of fibrin (green) anchoring the clot to the vessel¹⁸³. In clots, platelets have different activation levels, a core of platelets is defined by a marker of platelet activation, P-selectin (blue), and is surrounded by a shell of p-selectin negative platelets. Overlay of platelets and P-selectin is purple and overlay of all 3 fluorophores is white. f) Platelet activation levels affect the porosity of a clot, which affects the transport of solutes in the clot. An in vivo image of a platelet (blue) overlaid with an outline of a low transport region (LTR), which coincides with the activated platelet core. To define the LTR, fluorescent albumin was incorporated into the initial clot and diffused out over time. A pseudo colored image of the porosity was created by monitoring the loss of fluorescent signal over time, which correlates with loss of albumin and is dictated by the local porosity.¹⁹⁷. g) We hypothesize that mechanical homeostasis ensuring optimal forces, structure, porosity and stiffness enable correct clotting and that perturbations to this process can lead to either bleeding or excessive clotting.

The mechanisms of the regulation of fibrin structures are not yet fully understood; however, perturbations to fibrin formation can be associated with pathophysiological conditions and altered blood clot properties. For example, patients suffering from haemophilia, dysfibrinogenemia or afibrinogenemia, have impaired or no fibrin formation leading to the generation of unstable clots that are prone to rupture and that can break away. Structural changes to fibrin can further be related to multiple myeloma¹⁷¹, diabetes²³, nephrotic syndrome¹⁷², deep vein thrombosis¹⁷³, atrial fibrillation and stroke¹⁷⁴. In all these conditions, the fibrin meshes are stiffer, more densely packed with fibres and/or less prone to fibrinolysis than in healthy controls. Although the underlying mechanisms for this altered structure remain unclear, these diseases can be associated with increased thrombin generation¹⁷⁵, increased fibrinogen levels¹⁷⁶ and/or biochemical changes of fibrin itself²³, which all affect fibrin structure. Decoupling these different factors and identifying the dominant cause of altered fibrin structure remains challenging and requires the development of new tools and assays. Currently, fibrin formation and structure are mainly investigated by clotting platelet-poor plasma of patients, which contains a myriad of factors that can affect fibrin structure and which varies in fibrinogen, thrombin and coagulation factor concentrations. Therefore, using this approach, it is difficult if not impossible to isolate and test the effect of each condition on the final blood clot structure.

Platelets

Actuators of blood clotting

Nascent fibrin provides the structure of blood clots and reinforces the developing clot, whereas platelets can be thought of as the actuators of blood clotting, which can sense and apply force to equilibrate the local microenvironment (FIG. 4b). Sensing occurs biochemically on the platelet surface through a range of receptors, such as integrins, that can sense thrombin, ADP, epinephrine, collagen, vWF and fibrinogen. In addition, changing the biophysical arrangement of biochemical cues can affect the platelet response. For example, the extent of platelet spreading on glass surfaces depends on the concentration of fibrinogen on the surface¹⁷⁷. Paradoxically, platelets spread more on surfaces with low fibrinogen concentrations than on surfaces with high fibrinogen concentrations, with differences in the activation of signalling pathways.

Mechanical and biophysical stimuli can trigger a response in platelets. We have shown that platelets adhere, progressively spread less and have lower levels of activation as the stiffness fibrinogen or collagen coated polyacrylamide surfaces decreases from 100 kPa to 0.25 kPa ^{39, 40}. Interestingly, platelet mechanosensing is mediated by different pathways, depending on whether the platelet is undergoing initial adhesion or if it is actively spreading³⁹. Platelets attached to a surface are further capable of applying substantial forces (1.5 to 79nN), comparable to muscle cells if cell size is considered¹⁷⁸. Application of force by platelets is a mechanosensitive process, and the force increases with increasing substrate stiffness^{29, 178}; however, biochemical and mechanical conditions synergize to maximize platelet forces²⁹. In addition, there is an upper limit to the forces that a platelet can apply, which is defined by the total amount and the organization of actin and myosin. These studies suggest that platelets are actuators of blood clotting, which control clot stiffness and are regulated by signalling networks that respond to the biochemical and mechanical microenvironment.

Biophysical properties

We are only in the early stages of understanding platelet biophysics; however, it is known that perturbations to platelet forces are associated with pathophysiological conditions. In a blood clot, platelets facilitate a volumetric reduction ²⁰ and concomitant stiffening ¹⁷⁹, that can both be on the order of one magnitude. This process is referred to as the second phase of clot formation or clot contraction (FIG. 4c-d). Clot contraction which includes contributions from platelet biophysics, fibrin structure, and incorporated cells, are altered in many pathophysiological conditions, including severe coronary artery disease ¹⁸⁰ and systemic lupus erythematosus ¹⁸¹. Microenvironmental cues, including the mechanical properties of the underlying matrix substrate^{39, 178}, matrix geometry¹⁸², biochemical conditions¹⁸³ and shear stress^{184, 185}, mediate how platelets exert force on and contract other platelets and fibrin polymers.

Tools such as a rheometre¹⁷⁹, hemostasis analysis system¹⁸⁶, automated optical contraction analyzer²⁰ and platelet contraction cytometer²⁹ can be used to assess the biophysical properties of platelets. For example, platelets of patients with Glanzmann’s thrombasthenia²³, cannot aggregate and apply force owing to defective and low levels of fibrinogen-binding integrin α_IIbβ₃. However, a thorough understanding of single platelet biophysics in diseases related to bleeding or clotting remains elusive.

Bulk platelet forces can be measured using a customized whole-blood assay of platelet function, which can also measure the elastic modulus of blood clots²³. Using this assay, an increase in platelet force could be correlated to disorders with clotting complications, such as chest pain¹⁸⁷, severe coronary artery disease¹⁸⁰ and thromboangiitis obliterans¹⁸⁸. Whole-blood clotting assays were also used to demonstrate that impaired clot contraction occurs in diseases with clotting complications, such as acute ischemic stroke ¹⁸⁹, sickle cell disease ²⁰, systemic lupus erythematosus ¹⁸¹ and polycythemia vera ^{190, 191}. However, it is difficult to directly compare force and contraction assays. Moreover, numerous processes need to be coordinated to control blood clotting, which all affect the biophysical properties of platelets. Nevertheless, these results suggest that platelet function may differ in diseases related to blood clotting as compared to controls.

Microtechnology-based approaches can be applied to study platelet forces at the single cell level^{29, 192, 193}. For example, we developed a hydrogel-based platelet contraction cytometer to measure contractile forces of individual platelets. As expected platelets have lost their ability to exert forces in patients symptomatic bleeding and an actin or myosin disorder. Interestingly, platelets also lost their ability to exert forces in patients that had unexplained bleeding²⁹. Despite having normal values for all standard clinical tests of haemostasis, these patients still had symptomatic bleeding. More research is still required to understand mechanical homeostasis of whole blood clots; however, dysregulation of platelet force is certainly linked to bleeding and clotting.

Assembly and structure of clots

The initiation and assembly of blood clots impact clot properties. Blood clots have the remarkable ability to self-assemble and organize into a variety of architectures. Once coagulation begins, homogeneous, static blood forms a blood clot that contains a core of compressed red blood cells surrounded by a fibrin mesh¹⁹⁴. In flowing conditions, in vivo experiments have demonstrated that blood clots composed predominantly of platelets adopt a core–shell architecture, established by the different activation levels of the platelets within the clots¹⁸³. In this structure, a core of highly activated platelets that support fibrin formation is surrounded by a shell of less activated platelets^195–197 (FIG. 4e). The level of platelet activation is controlled by biochemical signalling events; however, biophysical cues, such as platelet-mediated porosity of the blood clot, also play a key role in controlling gradients of various platelet activators, such as ADP and thrombin¹⁹⁸ (FIG. 4f). Moreover, as coagulation is controlled by a cascade of chemical reactions that occur in the complex hemodynamic environment, conditions affecting mass transfer and chemical kinetics affect blood clot formation¹⁹⁹. In addition, the presence of valves⁷⁸ or certain geometries, such as aneurysms, stenoses and bifurcations ⁴⁷, can influence the adhesion of blood cells and the initiation of clotting. As the clot matures, platelet contraction may redistribute the less adhesive “procoagulant” platelets to the clot periphery, which expose high amounts of phosphatidylserine and enhance fibrin formation²⁰⁰ The different clot architectures can be expected to affect the mechanical equilibrium of the clot, whereas the details of the structure-mechanics relationship remain to be elucidated.

Correlating the mechanical properties of blood clots and its various components with disease requires an understanding of the mechanical properties of the individual components, their interactions and assembly, and the contraction process. The feedback and regulatory mechanisms that maintain optimal stiffness values also need to be investigated; in particular, the individual contribution of each clot component to the overall clot stiffness throughout the clotting process, the interactions that drive clot contraction, the assembled architectures, the translation of microscale changes in cells, platelets and fibrin into macroscale behaviours, and the interactions between platelets, fibrin and other cells. For example, studies of clot component interactions have shown that platelets can fold fibrin into a structure featuring ‘kinks’²⁰¹. Finally, more research is needed to understand whether pathophysiological changes to the mechanical properties of platelets, fibrin and blood clots are associative or causative.

Mechanical homeostasis in blood clots?

Defining mechanical homeostasis as the process by which a biological system maintains a mechanical equilibrium using feedback mechanisms, this concept may also be applied to blood clots (FIG. 4g). Blood clots have an optimal and regulated mechanical ‘set point’ or equilibrium, at which haemostasis is supported instead of thrombosis. Mechanosensitive platelets can decrease the size of clots and stabilize them by exerting forces to regulate their mechanical equilibrium. Therefore, we hypothesize that the blood clot maintains mechanical homeostasis, through mechanical feedback loops. For example, the mechanosensitive behaviour of platelets contributes to a uniform clot contraction process¹⁷⁸. During clot development, activated platelets in the heterogeneous fibrin gel can pull more strongly on denser and stiffer fibrin regions than on softer fibrin, with both regions experiencing the same net changes in strain. Similarly, the mechanosensitivity of platelets³⁹ and the fact that they stiffen with contraction¹⁷⁸ could provide a positive feedback loop mechanism of platelet activation within the growing thrombus, to further activate platelets and trigger the release of granules. Biomechanical and biochemical pathways might then coordinate to maximize platelet forces²⁹. As forces increase, the shorter bond lifetimes of fibrin-fibrin interactions may help to limit the maximum clot stiffness by altering the fibrin architecture. Towards the end of the life of a platelet, when phosphatidylserine is exposed, platelets may even stop applying forces³², suggesting that platelets innately have the ability to tune the mechanical stiffness of the system. Forces applied by platelets may even regulate the growth of a clot since contraction moves less adhesive, soft, procoagulant platelets to the clot periphery, which could play a role in halting the uncontrolled growth of a clot. Blood clots are transient structures that block the flow of blood at the early stages of wound healing. Given the importance of clotting and the involvement of mechanosensitive cells and proteins, mechanical feedback mechanisms, similar to the ones observed in blood vessels, might be involved in clot regulation.

Conclusions and perspective

Blood and blood tissues have mainly been investigated from a biological and biochemical perspective. However, a materials-based framework can offer important insight into such a dynamic system, which is also greatly regulated by physical interactions. In this Review, we discuss the concept of mechanical homeostasis for blood and blood tissues and the idea that competing mechanical interactions are balanced with associated feedback mechanisms. Similar to the concept of cellular tensegrity, which proposes a system in which cells are viewed as prestressed structures capable of sensing and responding to mechanical forces, the concept of mechanical homeostasis in blood suggests dynamic mechanically-driven sensing and response mechanisms of the various blood components. Acto-myosin forces prestress and regulate a structure that is composed of tensile cytoskeletal microfilaments and intermediate filaments that are balanced by elements resisting compression such as microtubule struts and ECM adhesions.. Similarly, blood cells and tissues can maintain a mechanical equilibrium with tensile and compressive elements, which is regulated by feedback mechanisms. Since mechanics is disrupted in disease, this suggests that a mechanical equilibrium may be required to ensure tissue function.

Investigating the concept of mechanical homeostasis in blood might provide new insights not only in basic haematology, but also for diagnostics and therapeutics. Blood vessels, blood and blood clots sense and respond to numerous biochemical and biomechanical stimuli, which are often altered in disease. Technologies based on better recapitulating the in vivo mechanics of blood and blood tissues have enabled new functional studies, such better understanding platelet adhesion, spreading ^{39, 40}, and force ²⁹. Soft materials, such as hydrogels, support the 3D culture of endothelial cells¹⁰⁶ and thus, can be used to investigate how changes in red blood cell properties or haemolytic by-products influence endothelial cell permeability. Furthermore, blood microenvironments often feature multi-cellular structures with collective behaviours that cannot necessarily be extrapolated from single cell studies. For example, platelets and fibrin are relatively well studied at the single cell level, but much less is known about the interaction of these components and how they collectively lead to clot contraction.

Insights into the mechanical and material properties of blood components have led to exciting innovative approaches in diagnostics and therapeutics. For example, creating soft hydrogel microparticles with a similar size and shape than those of red blood cells and with low stiffness (7.8 kPa) allow the optimization of the biodistribution of microparticles in blood by mimicking the behaviour of red blood cells. Interestingly, the low elastic modulus of these hydrogel microparticles increased their circulation time in vivo, from a half-life of <10hrs for stiff particles (>17kPa) to a half-life of 93 hrs for low stiffness (7.8 kPa) particles. This enabled the particles to avoid sequestration in many organs including the lungs, which can lead to pulmonary embolism²⁰². Similarly, by using red blood cells as carriers, clearance and non-optimal delivery of nanoparticles and viral vectors can be avoided²⁰³. Injecting red blood cells loaded with nanoparticles upstream of the target organ, enables delivery of the particles to this organ with substantially increased transfer efficiency. Combining knowledge about margination and fluid mechanics allows the design of a drug delivery system to transport drugs directly to areas of high shear stress. For example, aggregates of nanoparticles coated with therapeutic compounds can be engineered to break apart in high shear regions, allowing targeted drug delivery to the vessel wall²⁰⁴.

The mechanical forces exerted by platelets may provide a new type of biophysical biomarker. Collectives of platelets exerting low forces, typically with an average force below 25nN, are linked to symptomatic bleeding in patients, who show normal haemostasis using existing clinical tests²⁹. Remarkably, this correlation is independent of platelet biomarkers of activation, showing that mechanical forces offer an independent measure of health. Platelet contraction could further be exploited for the design of a drug delivery platform capable of locally modifying the biochemical microenvironment of a clot²⁰⁵. Strategies using soft materials can further be applied to limit biomechanics-induced pathological clotting on implanted materials since platelet adhesion is reduced on soft surfaces^{39, 40}. Similarly, ultrasoft microgels augment blood clotting under physiological flow conditions²⁰⁶. These spherical ultra-low cross-linked microgels with a diameter of 1 μm were capable of spreading into discs on glass that had a diameter of 2 μm and a height of 4 nm. By adding molecular-recognition motifs, these platelet-like particles collapse fibrin networks in vitro and reduced bleeding times in vivo. Therefore, the mechanics and materials properties of blood and blood tissues offer a new tool for diagnosis and treatment of blood-related disorders, showcasing the important contribution of materials science to medicine.