Bio-inspired hemocompatible surface modifications for biomedical applications

Abstract

When blood first encounters the artificial surface of a medical device, a complex series of biochemical reactions is triggered, potentially resulting in clinical complications such as embolism/occlusion, inflammation, or device failure. Preventing thrombus formation on the surface of blood-contacting devices is crucial for maintaining device functionality and patient safety. As the number of patients reliant on blood-contacting devices continues to grow, minimizing the risk associated with these devices is vital towards lowering healthcare-associated morbidity and mortality. The current standard clinical practice primarily requires the systemic administration of anticoagulants such as heparin, which can result in serious complications such as post-operative bleeding and heparin-induced thrombocytopenia (HIT). Due to these complications, the administration of antithrombotic agents remains one of the leading causes of clinical drug-related deaths. To reduce the side effects spurred by systemic anticoagulation, researchers have been inspired by the hemocompatibility exhibited by natural phenomena, and thus have begun developing medical-grade surfaces which aim to exhibit total hemocompatibility via biomimicry. This review paper aims to address different bio-inspired surface modifications that increase hemocompatibility, discuss the limitations of each method, and explore the future direction for hemocompatible surface research.

1. Introduction

1.1. Background of blood-contacting devices

Blood-contacting devices serve as a mainstay for intravenous drug administration, tissue engineering, extracorporeal circulation (ECC), and disease management (Fig. 1). Vascular bypass grafts, catheters, stents, pacemakers, and heart valves are widely used for the treatment of cardiovascular diseases, amounting to 17% of annual health expenditures in the US and the leading cause of death globally [1,2]. By 2030, the American Heart Association projects that 40.5% of the population in the United States will have some form of cardiovascular disease [1], and failure of devices used for treatment will be catastrophic towards morbidity and mortality rates as well as patient cost. Between 2010 and 2030, direct and indirect medical costs attributed to cardiovascular disease are projected to amount to $818 and $276 billion, respectively [1]. One of the most common complications associated with indwelling cardiovascular devices, including peripherally inserted central catheters, is thrombosis (10–20% rate), which has direct implications on patient outcomes [3]. Therefore, improving the likelihood of successful device applications may significantly reduce the cost of treating cardiovascular diseases and improve patient outcomes.

Fig. 1.

Blood-contacting devices are used for various clinical applications including drug administration, extracorporeal circulation, cardiovascular disease treatment, and tissue engineering. Devices used for clinical use range from short-term application (minutes – hours) to permanent exposure (months – years) to the body.

Moreover, extracorporeal blood-contacting devices are critical for patients requiring cardiac and respiratory support or removal of excess water, solutes, or toxins from the blood. Chronic kidney disease affects 14.8% of adults in the US, and end-stage renal disease (ESRD) affects approximately 750,000 people in the US per year [4]. ESRD patients are limited to two treatment options: donor transplantation and dialysis. Dialysis requires on average three different 3–5 h long sessions per week, resulting in frequent ECC applications for circulatory support and subsequent risks of clot formation [4]. Extracorporeal membrane oxygenators (ECMO) are used as cardiopulmonary life-support through mechanical circulation outside the body to oxygenate the blood. In addition to neonates and pediatrics who compose the majority of ECMO patients, respiratory outbreaks such as the COVID-19 pandemic have continued to underline the importance of improving ECMO treatment. Both dialyzers and ECMO devices are extensively utilized for blood filtration but contain large surface areas that, without systemic anticoagulation, would induce widespread thrombosis [5].

Regulating host response upon exposure to the devices remains the largest challenge relating to blood-contacting devices. A major determining factor in the success of blood-contacting devices is their hemocompatibility. Preventing thrombus formation on the surface of devices is crucial for maintaining device functionality and patient safety. Despite the intensive research and development surrounding these devices, blood coagulation and thrombosis remain the largest limitation of many medical devices. When foreign surfaces are exposed to blood, thrombus formation spontaneously and abruptly occurs, ultimately leading to device failure in the absence of anticoagulant or antiplatelet therapies. Device-induced blood clots can impede device functionality, occlude vessels, or break off and move downstream, causing further complications such as pulmonary embolism, kidney failure, deep vein thrombosis, heart attack, or stroke. Thrombosis is the most common cause of vascular access failure in hemodialysis patients [6], and occurs at rates of up to 66% in long-term central venous catheters [7]. While systemic anticoagulation helps reduce the occurrence of device-induced thrombosis, these events still occur. The administration of anticoagulants can also lead to hemorrhaging. Thrombosis and bleeding have been linked to decreased survival rates by 33% and 40% in ECMO patients, respectively [8]. Even with anticoagulant therapy, stenting often leads to late-stage thrombotic occlusion of the stent, manifesting as restenosis and resulting in further systemic thromboembolism [9]. Late-stage restenosis after corrective surgery can lead to further cardiac events, often necessitating patient readmission. Therefore, the use of systemic anticoagulation and antiplatelet therapies requires careful monitoring to prevent device-induced thrombosis while minimizing the risk of bleeding.

In response, researchers have begun to develop surface modifications for medical devices that mimic hemocompatible natural phenomena to reduce the occurrence of device-induced thrombosis and dependence on systemic anticoagulation. This review will address different bio-inspired surface modifications within each category, as well as the limitations and future directions of these devices.

1.2. Mechanisms of medical device-induced thrombosis

When exposed to blood, foreign surfaces are subjected to various complex reactions which can compromise the life span and usability of the device (Fig. 2). Under normal conditions, due to its antithrombotic properties, the endothelium can interact with blood without triggering clot formation [10]. However, when devices are introduced to the blood, the adsorption of physiological proteins initiates the activation of a number of biological processes like the coagulation cascade and inflammation. Overall, medical device-induced thrombosis is the result of fXIIa-mediated thrombin formation and platelet adhesion and aggregation which are both initiated by protein adsorption [11].

Fig. 2.

Schematic representing major biochemical reactions including the intrinsic and extrinsic coagulation pathways, complement activation, and the common pathway of coagulation as a result of the exposure of medical devices. vWF, von Willebrand factor; NO, nitric oxide; TNFα, tumor necrosis factor; IL-1, interleukin 1; HMWK, high molecular weight kininogen; TF, tissue factor; MAC, membrane attack complex; RBC, red blood cell. Black dotted lines represent binding.

Proteins, which constitute a major part of the plasma, are considered to initiate thrombosis by rapidly adsorbing to the foreign surface immediately after it enters the blood [11]. The surface chemistry and physical properties of the device modulate which proteins are attracted to the surface and the strength at which they adsorb. Furthermore, the Vroman effect may occur, a process wherein proteins can be displaced by other proteins over time based on their relative affinity, size, and charge [11]. Generally, smaller proteins adsorb quickly to the surface and are eventually replaced by larger proteins or proteins with greater affinity [12]. Blood is composed of different plasma proteins, several of which play key roles in mediating platelet, leukocyte, and red blood cell attachment.

Platelets are essential for the maintenance of hemostasis, the formation of hemostatic plugs, and the releasing of pro-coagulant signals which ultimately assist in the transformation of prothrombin to thrombin [13]. Platelets interact with fibrinogen attached to the foreign surface by the integrin αII_bβ₃ present on platelets [11]. Due to their high affinity to fibrinogen, platelets can adhere at adsorbed fibrinogen concentrations as small as 7 ng cm⁻² [11,14]. After adhering to fibrinogen, platelets begin to form dendritic pseudopodia and release agonists that further promote aggregation and adhesion of platelets [12]. Von Willebrand factor joins together activated platelets through the αII_bβ₃ complex [15]. Adhered leukocytes can degranulate, releasing platelet-activating factor, interleukins, and TNFα, which further progress platelet activation. Red blood cells adhere independently from the protein monolayer and release adenosine diphosphate (ADP), which further promotes platelet aggregation [11].

Key factors composing the contact-phase system can also bind to the device surface and displace fibrinogen including factor XII (fXII), HMWK, prekallikrein (PK), and factor XI (fXI) [11,16]. Bound fXII can autoactivate into fXIIa, which activates PK into active kallikrein and HMWK into bradykinin, both of which stimulate coagulation and inflammatory responses [11]. Through the intrinsic pathway, fXIIa begins a series of reactions that results in the activation of factor X, ultimately triggering thrombin generation and inflammation [11,16]. Both the extrinsic and intrinsic pathways of coagulation, which share the activation of factor X, begin the conversion of prothrombin to thrombin and the transformation of fibrinogen to fibrin and platelet aggregation, resulting in a dense network of clot formation [11,16].

Exposure of artificial surfaces to blood can also induce the complement system, a key pathway for immune response as the initial line of defense against foreign bodies. The complement system is initiated by three different pathways: classical, alternative, and mannose-binding lectin (MBL). Ultimately, these three pathways result in C3 convertase, which by cleaving C3 into C3a and C3b, promotes inflammation at the site. The generation of C5 convertase also cleaves C5 into C5a and C5b [16]. Together C3a and C5a increase the recruitment, attachment, and activation of leukocytes [11]. Moreover, C5b can subsequently bind to the surface and initiate the production of the membrane attack complex (MAC), triggering a series of inflammatory reactions [16].

In addition to coagulation complications, introducing a foreign material can also induce both acute and chronic inflammatory responses. Similar to medical device-induced thrombosis, inflammation is initially triggered through the adsorption of proteins (ex. complement components, fibrinogen, vWF, vascular cell adhesion molecule (ex. VCAM-1, P-selectin) on the surface, which enable inflammatory cell recruitment and attachment [17,18]. Platelets that attach to the surface generate pro-inflammatory molecules including monomeric C-reactive protein and tissue factors that recruit inflammatory cells (ex. macrophages, polymorphonuclear leukocytes), promoting thrombosis [17,19]. In addition to their roles in device-induced thrombosis, kallikrein, thrombin, and other coagulation enzymes stimulate a local inflammatory response [11]. Over time, monocyte-derived macrophages begin to replace polymorphonuclear leukocytes and degrade the device through the generation of reactive oxygen and nitrogen species as well as hydrolytic enzymes [19]. Fibrous encapsulation is facilitated by later-recruited macrophages, walling off the implant from the body and obstructing device functionality [19]. Excessive inflammatory responses can lead to several major complications including device failure, neointimal thickening, and tissue damage [18,20].

To increase hemocompatibility, researchers are now modifying surfaces to prevent or disrupt the pathways described above. Hemocompatibility assessment guidelines provided by ISO 10993–4 are comprised of five distinct categories: thrombosis, coagulation, platelets, hematology, and immunology [16]. To alleviate undesired side effects from medical device exposure and minimize patient risk, the search for a biocompatible surface that prevents blood component activation in each of these categories has been initiated.

1.3. Current treatment options

Current methods to prevent thrombus formation caused by medical devices fall under two categories: systemic anticoagulation/antiplatelet therapy and the usage of hemocompatible devices. Systemic administration of anticoagulants, mainly systemic heparinization, remains the most widely used technique to control clot formation. However, anticoagulant therapies are among the most common causes of adverse drug-related events and deaths in hospitalized patients. A surveillance study examining adverse drug events with hospitalized Medicare patients found that 13.6% of patients administered heparin and 8.2% of patients administered warfarin experienced an adverse drug event [21]. According to a comprehensive study conducted at Brigham and Women’s Hospital, of the nearly 500 anticoagulant-associated adverse drug events, 48.8% were due to medication errors, 30.5% were due to adverse drug reactions, and 20.7% were due to both medication errors and adverse drug reactions; as a result, death within 30 days of the adverse event occurred in 11% of patients [22]. Patients receiving anticoagulants are likely to have medical conditions (ex. heart failure, ischemic heart disease, chronic kidney disease, stroke) that increase their vulnerability to adverse drug events, increasing the likelihood of longer hospital stay, more complicated treatment requirements, and risk of death [22]. All forms of anticoagulants have been associated with the development of acute hemorrhaging [23]. Incidence rates of major bleeding events during systemic heparinization are reported to occur at 7.3 to 16.7 per 100 person-years [24]. Moreover, systemic heparinization has led to side effects including drug intolerance, thrombocytopenia, and osteopenia [25]. Even after discontinuation of systemic heparinization, 50% of patients that develop heparin-induced thrombocytopenia experience a thrombotic event, and thrombotic complications associated with heparin-induced thrombocytopenia are associated with a mortality rate between 20 and 30% [26,27]. In response, researchers have set out to develop improved hemocompatible devices, averting the need for additional systemic administration of anticoagulants or antiplatelet therapies [28].

Significant research has been conducted in identifying physical, chemical, and biological surface properties that result in antithrombotic behavior. Characteristics including surface charge, the presence of hydrogen bond acceptors, polarity, surface roughness, and hydration forces alter the interactions between the device and blood [16,29,30]. However, even with these surface adaptions, achieving a biocompatible device that exhibits antiplatelet and protein-repulsive activity has proven to be difficult. One of the most widely explored surface modification techniques is the incorporation of polyethylene glycol (PEG). PEG-coated surfaces form tightly bound water layers at the interface that proteins are unable to displace as needed for protein adsorption [31]. Although PEG-incorporated surfaces significantly reduce thrombus formation [32,33], incorporating PEG into polymers generates several drawbacks. PEG is not biodegradable and can trigger the complement pathway, resulting in an immediate or delayed immunological response [34,35]. Although PEG incorporation is generally considered to suppress protein adsorption, exposure to some PEGylated therapeutics has resulted in the generation of anti-PEG antibodies, activation of the complement system, and caused hypersensitivity reactions [36]. However, the exact mechanisms of this are not well-understood.

Other synthetic materials and surface modifications such as the incorporation of poly(2-ethyl-2-oxazoline), titanium oxides/nitrides, polyethylene oxide (PEO), poly(vinyl chloride) (PVC), poly(ethylene), and polysulfone (PSF) have also been utilized to increase hemocompatibility, but despite initial promising results, each method has fallen short of reaching total blood compatibility [10,12,37,38]. For this reason, researchers have turned to emulating hemocompatible and antifouling bodies found in nature (Table 1). Bio-inspired technologies for improving hemocompatibility can be broken down into three categories based on their method of achieving blood compatibility: biopassive methods, bioactive methods, and promotion of endothelial cell growth. Biopassive surfaces do not interact with the environment or trigger an immune response, while bioactive coatings contain or release agents that actively interfere with components that promote thrombus formation and provoke an immune response [12]. More recently, the promotion of endothelial cell growth on the surface of medical devices has been recognized as a final method of preventing thrombus formation, forming a barrier that prohibits the interaction between the device and the surrounding environment, and is particularly necessary for long-term indwelling devices [39]. The remainder of this review will cover technologies developed in each category, the short-comings of each strategy, and future directions for bio-inspired hemocompatible devices.

Table 1.

Bio-inspired hemocompatible surfaces and their origin. (Blue – biopassive strategies, Green – bioactive strategies, Red – promotion of endothelial cell growth strategies).

Nature-inspired hemocompatible surfaces	Origin	Citations
Antifouling micropatterned surfaces	Nelumbo nucifera, shark skin	[40–50, 53–66, 355]
Liquid-infused surfaces	Pitcher plant	[67–86, 356]
Zwitterionic phosphorylcholine-incorporated surfaces	Cellular membrane	[87–109, 142–145]
Passive protein-immobilized surfaces	Dysopsonins	[146–156, 158–163, 168–171]
Nitric oxide-releasing surfaces	Nitric oxide synthase	[83, 86, 105, 170, 173–175, 178–210, 213–219, 221]
Antithrombotic polysaccharide-immobilized surfaces	Glycosaminoglycans	[30, 171, 219, 222–240, 243–256, 258–260]
Thrombin-inhibiting peptide-immobilized surfaces	Direct thrombin inhibitors	[241, 267–283]
Endothelial progenitor cells (EPC)-functionalized surfaces	Endothelial layer	[285–301, 303–305, 307, 309, 311–331]
Endothelium-inspired surface patterning	Endothelial structure	[334–346]

2. Current biopassive methods for improving surface hemocompatibility

The fate of a biomaterial ultimately depends on its ability to prevent recognition by the foreign body response. Biopassive surfaces aim to minimize triggering large adverse reactions, effectively evading surface-induced coagulation by preventing protein adsorption and platelet adhesion (Fig. 3). These modifications invoke a low immune response from the body, but the effectiveness of these devices over long periods of exposure time is still a concern, limiting their use almost exclusively to short-term or single-use applications. This review will discuss bio-inspired passive hemocompatible surface modifications as well as some of the drawbacks and limitations of each technology.

Fig. 3.

Overview of bio-inspired passive/inert surface strategies: (1) antifouling micropatterned surfaces, (2) liquid-infused surfaces, (3) zwitterionic surfaces, and (4) passive protein-immobilized surfaces.

2.1. Biomimetic antifouling micropatterned surfaces

The invention of scanning electron microscopes was a pivotal moment for the development of bioinspired surfaces, granting researchers the ability to study the three-dimensional structures of natural surfaces. The first in his field to utilize this revolutionary technology, German botanist Wilhelm Barthlott explored how the sacred lotus plant, Nelumbo nucifera (N. nucifera), maintains a pristine, dirt-free surface despite growing in muddy waters, thus pioneering the field of superhydrophobic (SH) surfaces and their underlying antifouling characteristics.

Dubbed “the lotus effect”, the self-cleaning ability of N. nucifera is derived from its double-hierarchical surface structure at the micro and nanoscale level (Fig. 4) [40–42]. The resulting structure is capable of trapping air between both the nanoscale wax rodlets and the microscale epidermal cells, which allows for as little as 0.09% of a water droplet’s surface area to remain in contact with the lotus leaf [43]. Therefore, significant effort has been placed into creating synthetic SH surfaces that mimic the lotus effect. Synthetic SH surfaces maintain an anti-fouling air pocket layer through having a SH surface texture coating, by removing base material to achieve SH texture, or a combination of the two methods. Details on the various methods used to create SH surfaces for a variety of applications can be found in several recently published review articles [44–48].

Fig. 4.

SEM (A and B) images showing the double-hierarchal surface structure of N. nucifera composed of a nanoscale spatial distribution of hydrophobic wax rodlets in combination with microscale surface roughness induced by irregular palisade epidermal cell placement. Inspired by the self-cleaning surfaces exhibited by the lotus plant, researchers have begun synthesizing superhydrophobic surfaces (c) exhibiting similar double-hierarchical micro- and nano-features in hopes of passively resisting platelet adhesion, protein adsorption, red blood cell interactions, and immune response (d). (A and B were reprinted/adapted from Applied Physics Letters, 87, 194,112, Microscopic observations of condensation of water on lotus leaves, Cheng et al, Copyright (2005), https://doi.org/10.1063/1.2130392, with the permission of AIP Publishing).[41].

Several reported SH surfaces have shown protein repulsion, decreased platelet attachment, and reduced blood component activation [49–52]. To increase the commercial feasibility of SH surfaces, Nokes et al. developed a roll-to-roll manufacturing technique by shrinking pre-stressed thermoplastic polystyrene with a silver/calcium layer that can mold a SH texture onto any medical-grade plastic surface [53]. Encouraged by the results of SH surfaces, Bark et al. found that SH-coated heart valves decreased in vitro leukocyte and platelet attachment, but despite predictions that the SH coating would give superior boundary layer conditions, the SH coating did not significantly improve hemodynamic flow around the valve [54]. Using computational fluid dynamic simulations, the authors hypothesized that the form drag overcomes energy dissipation over the valves used in the study. However, the authors noted that although the SH surface showed no improvement in the hemodynamics for this study, SH surfaces are still a promising solution for improving hemocompatibility [54].

Until recently, most of the work on SH surfaces for biomedical applications were performed in vitro, but clinical implementations would be best supported by in vivo data to capture the complex, dynamic chemical reactions and systemic effects triggered by the device. Using a combination of O₂ and SF₆ plasma treatment, Brancato et al. modified polydimethylsiloxane (PDMS)-filled polyurethane (PU) sheaths and after an 8-day sheep carotid artery model found that the outer surface of the catheters and the surrounding arterial vessel did not show any sign of clotting or thrombosis [55].

The inconclusive in vivo performance of SH surfaces can be explained by the fact that the stabilities of SH surfaces are often not measured. The air pocket layer in SH surfaces can be disrupted by changes in hydrostatic liquid pressure [56], condensation of surrounding liquid into the micro/nanostructure [57], or external vibrational forces [58]. Additionally, the induced textures used to create SH surfaces are highly susceptible to mechanical degradation from shear stress under flow [59,60]. Once the air pocket layer is disrupted, the underlying surface texture is exposed to the environment, giving blood proteins more surface area to bind onto [61–63]. While SH surfaces may initially repel protein adsorption, the air pocket interface can lead to protein denaturing, which can cause initiation of the coagulation cascade and lead to distal thrombus formation [64]. Therefore, future hemocompatible SH research should focus on the development of robust SH surfaces which can better maintain the stability of the air pocket layer in the dynamic physiological environment.

Drawing inspiration from animals, material scientist Dr. Anthony Brennan first discovered that the rhomboidal arrangement of V-shaped shark denticles protected sharks from biological fouling. The surface topography applies a stress gradient across foulants attempting to attach, ultimately causing the invader to detach [65]. As a result of its superior antifouling efficacy, researchers have begun to mimic the surface topography of shark skin for medical device applications (Fig. 5). Although Sharklet^™ micropatterns have shown efficacy in reducing bacterial colonization and biofilm formation, limited work has tested their antithrombotic properties. May et al. was the first to demonstrate that Sharklet^™ micropatterned thermoplastic polyurethane (TPU) reduced common foulants of central venous catheters including bacteria, platelets, and fibrin sheath formation (Fig. 5B–5I) [66]. This work compared both recessed and protruding Sharklet^™ micropatterned surfaces against unmodified TPU, and found that Sharklet^™ micropatterned surfaces with protruding features (+3SK2x2) best reduced platelet adhesion (~86% reduction) and fibrin sheath formation (~80% reduction) [66]. Future in vivo studies specifically examining the antiplatelet activity of shark-inspired surface patterns should be evaluated.

Fig. 5.

(A) Top-down representation of sharkskin-mimetic Sharklet^™ micropatterned surface, which minimizes biofouling from the surrounding environment. The efficacy of this surface modification was demonstrated in vitro by May et al. Recessed (−3SK2x2) and protruding (+3SK2x2) Sharklet patterned thermoplastic polyurethane (TPU) surfaces significantly reduced platelet adhesion (B-E) and fibrin sheath formation (F-I). (B-I were reprinted/adapted from Clinical and Translational Medicine, 4: e9 An engineered micropattern to reduce bacterial colonization, platelet adhesion and fibrin sheath formation for improved biocompatibility of central venous catheters., May et al, Copyright (2015), https://doi.org/10.1186/s40169-015-0050-9, with the permission of Wiley Online Library) [66].

2.2. Liquid-infused surfaces

As researchers began exploring alternative methods to achieve protein- and platelet-repelling surfaces, another method for attaining anti-fouling properties was again found in plants. Inspired by the waxy microstructure of Nepenthes pitcher plants, liquid-infused (LI) surfaces are a promising option towards the development of a truly hemocompatible surface. The wax of pitcher plants contains hydrophilic, aldehyde-concentrated crystals that are oriented perpendicular to the surface, notably different from the lotus’ hydrophobic waxes [67]. In combination with a smooth surface microstructure, pitcher plants exhibit a highly wettable surface capable of interlocking a thin layer of water in humid conditions (Fig. 6A–B) [68].

Fig. 6.

Peristome surface structure (A, B) of Nepenthes pitcher plant. Arrows indicate the direction towards the inside of the pitcher plant [68]. To mimic the antifouling properties of the Nepenthes pitcher plant (C), polymeric surfaces can be textured and chemically functionalized (1) and infused with a lubricating liquid layer (2) to form a slippery, liquid infused surface. Time lapse experiments of whole blood exposed to bare and omniphobic (FC-70-infused) fluorogels demonstrate the non-wetting, blood-repellent properties associated with liquid-infused surfaces [75]. (A-B were reproduced/adapted from Bohn, H.; Federle, W. Insect aquaplaning: Nepenthes pitcher plants capture prey with the peristome, a fully wettable water-lubricated anisotropic surface. PNAS 2004, 101 (39), 14138–14143, https://doi.org/10.1073/pnas.0405885101 with permission from the Proceedings of the National Academy of Sciences – Copyright (2004) National Academy of Sciences, USA. D was reproduced/adapted from Yao et al. Fluorogel Elastomers with Tunable Transparency, Elasticity, Shape-Memory, and Antifouling Properties. Angewandte Chemie 2014, 126(17), https://doi.org/10.1002/ange.201310385 with permissions from Wiley Online Library).

Inspired by Nepenthes pitcher plants, Wong et al. at the Harvard Wyss Institute reported the first bioinspired LI surface in 2011 [69]. Unlike SH surfaces which rely on hierarchical structure to repel liquids, slippery liquid-infused porous surfaces (SLIPS) utilize nanoscale porosity and capillary forces to tether an intermediate liquid to a polymer surface [70]. To tether a liquid onto a polymer: 1) the intermediate liquid (liquid A) must be nonreactive to the base polymer and the liquid to be repelled (liquid B); 2) the two liquids must be immiscible; 3) liquid A must wet the polymer surface more so than liquid B; and 4) the polymer must be either rough, porous, or able to be swelled with liquid A [71]. Once tethered, the intermediate liquid layer provides an anti-fouling, incompressible homogeneous layer capable of self-healing and self-cleaning [69,72,73]. Demonstrating the physiological efficacy of LI surfaces, Yuan et al. developed a fluorocarbon-infused surface by merging photografting polymerization with osmotically-driven wrinkling to create a rough surface morphology, reducing fibrinogen adsorption by 96% and minimizing thrombus formation [74]. Similarly, Yao et al. fabricated omniphobic fluorogel elastomers using perfluorinated alkyl acrylate monomers and a fluorinated macromolecular cross-linker capable of broad antifouling capabilities including decreased protein adsorption and whole blood repellency (Fig. 6D) [75].

Because LI technology is still developing, few experiments testing the hemocompatibility of these surfaces in vivo have been conducted. In a subcutaneous rat infection study examining the bacterial adhesion and subsequent inflammatory response of LI surfaces, Chen et al. showed that SLIPS-PTFE subcutaneous implants limited inflammatory cell response and attenuated macrophage activation compared to untreated controls [76]. Moreover, polycarbonate connectors and TLP-treated PVC cardiopulmonary perfusion tubing tested in an 8-h ECC porcine model maintained baseline blood flow rates without heparinization, while 4 of the 5 uncoated circuits had completely occluded by the end of the 8-h period [77].

While SLIPS surfaces have shown anti-fouling efficacy under physiological conditions, a variety of limitations remain before clinical application. First and foremost, the loss of lubricant remains a challenge for long-term applications driven by incorrect pairing between the lubricant and polymer, resulting in a cloaking layer [78,79]. To rid the lubricant-polymer pairing issue seen in SLIPS, a slippery surface method was developed that can be universally applied to biomedical polymers dubbed tethered liquid perfluorocarbon (TLP) [77]. By covalently binding perfluorocarbon silane groups to a polymer surface, these perfluorocarbon groups interlock fluorous lubricants, resulting in a low-energy slippery surface that can be universally applied to medical-grade polymers. Badv et al. further improved the efficiency and reproducibility of the TLP process via chemical vapor deposition of the perfluorocarbon silane groups [80].

Like other passive surfaces, the second limitation of LI surfaces is that they do not prevent blood component activation; however, LI surfaces are capable of overcoming this limitation by their ability to load bioactive agents for long-term release [81,82]. With hemocompatibility in mind, future slippery surface research should focus on mimicking the release or exposure of native endothelial agents to better mimic physiological blood vessels. Perhaps the most noteworthy agent released from the endothelium, nitric oxide (NO) produced by the endothelium attenuates platelet activation. Work by Goudie et al. was the first to combine LI surfaces and NO-releasing surfaces (LINOrel) into a silicone oil-infused polymer, resulting in a synergistic reduction in platelet adhesion [83]. The ability of NO-releasing materials to modulate biological responses such as the inflammatory response [84] and formations of thrombi [85,86] warrants further investigation in combination with slippery surfaces. More information on the incorporation of NO for hemocompatible devices can be found in the bioactive methods portion of this review article.

2.3. Zwitterionic surfaces

Selectively permeable and responsible for mediating cellular activity, cell membranes exhibit excellent hemocompatibility, minimizing cell and protein interactions. On the surface of the lipid bilayer of the cell membrane, the predominant polar head group is composed of phosphorylcholine (PC), the polar moiety of phosphatidylcholine and sphingomyelin. Spontaneously assembled, hydrophobic alkyl tail regions are repulsed from the polar physiological surrounding environment, forming a membrane bilayer with PC groups located in the outer leaflet (Fig. 7).

Fig. 7.

Schematic representation of the structure of the cellular membrane lipid bilayer model depicting phosphorylcholine groups present on the outer and inner surface. The structure of 2-methacryloyloxyethyl phosphorylcholine (MPC) is shown above, depicting the PC group mimicking the PC moieties present in the outer leaflet of the cell membrane.

Due to the inert behavior of PC moieties, synthetic amphiphilic copolymers have been the subject of research for several years. The Nakabayashi group in Japan was the first to synthesize a new biocompatible monomer having the phosphorylcholine moiety, 2-methacryloyloxyethyl phosphorylcholine (MPC) [87]. Ishihara et al. later synthesized a phospholipid polar group, MPC-co-n-butyl methacrylate (PMB) and showed that increasing MPC concentration reduced platelet adhesion and aggregation [88]. PMB surfaces have also inhibited fibrin deposition when exposed to human whole blood in the absence of an anticoagulant [88,89]. The MPC moiety creates a biomembrane-like surface via the adsorption of phospholipids in plasma on the surface through self-organization, negligibly interacting with plasma proteins and cells [89].

Various immobilization techniques have been used to introduce MPC to the surface of polymers including reversible addition-fragmentation chain transfer (RAFT), atom transfer radical polymerization (ATRP), and other grafting techniques [90–94]. Kyomoto et al. demonstrated that during photo-induced graft polymerization of MPC on poly(ether-ether-ketone), coating density and thickness of the MPC-containing layer could be modulated by varying photo-irradiation duration and monomer concentration, resulting in a highly tunable hemocompatible surface [95]. Similarly, PC-based surface modifications have been extended to metals such as titanium alloys, exhibiting long-term stability and improved hemocompatibility [96,97]. Natural polymers often plagued by thrombogenicity have also benefitted from PC grafting. PC-modified chitosan via layer-by-layer assembly, grafting, and crosslinking have all resulted in decreased platelet adhesion and protein adsorption after in vitro exposure to plasma [98,99]. Zhu et al. also fabricated O-butyrylchitosan-bonded MPC, a chitosan derivative, imparting anti-thrombogenic properties to chitosan by decreasing in vitro clotting and platelet adhesion [100].

To validate the hemocompatibility of PC in vivo, PC coatings and surface immobilization strategies have been applied to various medical devices. EVAHEART^® LVAS (Sun Medical Technology Research Co), a Ti-based VAD, has been modified with commercially available MPC-co-n-butyl methacrylate (PMB30) and extensively evaluated during in vivo bovine testing [101]. MPC-coated PSF-based fibers used in hemodialyzers prevented blood coagulation while maintaining permeability in an ex vivo rabbit model [102]. Similarly, a PC-based copolymer coated on a polypropylene (PP) hollow fiber membrane oxygenator assessed both in vitro and in an extracorporeal dog model significantly reduced albumin and fibrinogen adsorption as well as platelet adhesion (Fig. 8) [103,104]. Zwitterionic PC-modified catheters with nitric oxide-releasing technology resulted in synergistic antithrombic activity and decreased smooth muscle cell proliferation in a 7-day rabbit model [105].

Fig. 8.

Wang et al. functionalized polypropylene hollow fiber membranes commonly used for artificial lung applications with a crosslinkable MPC-based copolymer coating to reduce protein and platelet interaction (A). Increasing concentrations of poly(MPC-co-BMA-co-TSMA) (PBMT) coated onto the hollow fiber membranes resulted in decreased adsorption of bovine serum albumin and fibrinogen (B). SEM images of bare (C) and PMBT-coated (D) polypropylene hollow fiber membranes after 2 h of whole blood interaction. (Reprinted from Journal of Membrane Science, 452, Wang et al., Hemocompatibility and film stability improvement of crosslinkable MPC copolymer coated polypropylene hollow fiber membrane, 29–36, Copyright (2014), with permission from Elsevier.) [103].

Promising clinical results have also helped PC-based surfaces transition to commercial applications. The BiodivYsio^™ stent (Biocompatibles Ltd, Farnham, UK) is one of the first commercially available stents coated with a PC-containing copolymer. A prospective, multicenter-observational study, SOPHOS (Study of Phosphorylcholine coating On Stents), conducted on 425 patients suffering from stable or unstable angina pectoris found BiodivYsio^™ safe and effective as a stent therapy of native coronary artery lesions [106]. BiodivYsio^™ stents examined in a study on 70 patients treated with primary coronary angioplasty for short term (9 ± 3 days) and long term (7.2 ± 1 month) were found to have no acute or subacute thrombosis [107]. In an assessment of a PC-coated extracorporeal circuit in two groups of 10 patients before and after a cardiopulmonary bypass procedure, De Somer et al. found that PC-coated circuits generated less complement activation and reduced platelet adhesion compared to untreated circuits [108]. The cumulative data showcasing the efficacy and safety of PC-coated medical devices provides a promising prospect for future use.

Over the past few decades, PC-based polymers have proven effective towards minimizing protein adsorption and platelet adhesion and activation. PMB30, a commercially available methacrylate with a PC group in the side chain, has already been approved by the Food and Drug Administration (FDA) in the United States and has been implemented in medical device surfaces, creating excellent hemocompatible and biocompatible properties [89,109].

Apart from MPC, sulfobetaines (SB) and carboxybetaines (CB) are well-known zwitterionic polybetaines that exhibit excellent hemocompatible properties. Although each forms a strong hydration layer that reduces fouling, SB and CB differ by their negatively charged head group ($-{\text{SO}}_3^{-}$ for SB and COO⁻ for CB) [110]. Various strategies have been employed to immobilize SB and CB monomers onto device surfaces.

SB has been incorporated into several different polymers and metals commonly used in blood-contacting applications including PU, PDMS, PP, PSF, PET, titanium and titanium alloys, magnesium alloys, and PTFE [111–121]. Ye et al. found that, through fabricating biodegradable PU containing varied amounts of SB, increased SB concentrations significantly increased water absorption, decreased fibrinogen adsorption, and decreased platelet deposition [119]. Similarly, Tan et al. found that higher poly SB methacrylate content on PDMS led to reduced platelet adhesion, whole blood attachment, and fibrinogen adsorption [117].

Due to the promising blood compatibility properties found in vitro, several groups have begun to modify existing medical devices with SB. SB-coated titanium and magnesium alloys that can be utilized in stent applications have reduced in vitro platelet deposition [112,114]. Carboxyl-functionalized SB-coated PP hollow fiber membranes demonstrated improved thromboresistance without impeding the gas exchange rate required for artificial lung applications [122]. Similarly, Malkin et al. showed that polymethylpentene hollow fiber membranes modified with SB block copolymers reduce platelet deposition from whole ovine blood up to 95% [123]. However, in vivo testing of these materials has been limited. In an in vivo canine model, SB-modified vascular catheters reduced thrombus formation by 99% [124]. Longer in vivo evaluation should be performed to determine the safety and effectiveness of SB modifications for blood-contacting applications.

In addition to SB, CB-based surface modifications gained traction due to their ultra-low fouling capabilities on several different substrates including PU, PS, PP, PVC, PET, silicone rubber, and cellulose [125–132]. CB-based polymer brushes have shown to resist protein and platelet fouling when exposed to plasma or whole blood [133–135]. Extending CB modifications to medical applications, CB-modified PP membranes show superior antifouling activity in blood filtration devices [136]. Moreover, CB-modified PVC delayed protein adsorption, platelet adhesion/activation, and compliment activation, prolonging the lifetime of blood bag materials [127,137]. While evaluating CB’s biocompatibility in vivo, Ukita et al. found that grafting artificial lungs via 3,4-dihydroxyphenylalanine (DOPA) conjugated with poly CB reduced fibrin formation, thrombus formation, and device failure in rabbit and sheep models [138].

Comparing SB and CB surfaces, Cai et al. found that coating PET surfaces with either zwitterion resulted in similar reductions in platelet adhesion/activation and protein adsorption with minimal hemolysis [126]. However, an effectiveness study comparing poly CB methyacrylate, poly SB methyacrylate, and poly PC methacrylate polymer brushes found that CB-based surfaces outperformed the other hydrophilic zwitterionic surfaces, totally preventing plasma fouling [139]. Similarly, Zhang et al. found that both SB and CB polymer brushes coated on gold substrates increased clotting time and lowered platelet adsorption; however, CB polymer brushes resulted in lower plasma protein adsorption [140]. Ladd et al. suggested that CB surfaces may exhibit greater protein resistance due to closer distance between charged groups on the zwitterionic monomers, which results in a more robust hydration layer [141].

Although much progress has been made with zwitterionic polymers, several problems have been identified which few researchers have begun to address. PC moieties can be costly, insoluble, sensitive to moisture, and can be compromised due to the rearrangement and subsequent concealment of PC under the hydrophobic chains (a process of hydrophobic recovery), which decreases the interfacial free energy responsible for antifouling [142–145]. Moreover, the majority of studies conducted on zwitterionic biomaterials still only illustrate the stability of immobilized groups under static conditions or for a short period. Future studies should examine the stability of zwitterionic surfaces in dynamic conditions for long-term or permanent applications.

2.4. Passive protein-coated surfaces

Studies have indicated that immobilizing specific physiological proteins onto the surface of materials can improve their hemocompatibility [146]. Extensively studied in nanoparticle research and development, the adsorption of proteins around a nanoparticle in vivo, commonly known as the protein corona, strongly influences the fate of nanoparticles [147]. In fact, Schottler et al. even go on to argue that ‘pre-loading’ nanocarriers with certain protein classes is necessary to avoid non-specific cellular uptake [147]. A similar phenomenon has occurred throughout the wider application of biomaterials, where evidence has shown that it is not the repulsion of proteins, but the affinity of desired proteins due to surface functionalization, that contributes to biopassive activity (Fig. 9) [147–149]. However, depending on the class of proteins adsorbed to the surface, the biopassive activity of protein-coated surfaces is time-sensitive and may even lead to further activation [150]. Moreover, surface-bound proteins can unfold and reveal antigenic sites, which may trigger further cellular and immune responses [150]. The composition and activity of surface-bound proteins can be carefully tuned by modifying surface chemistry. Parameters including temperature, pH, ionic strength, protein size, net charge, structure, surface roughness, and surface polarity strongly influence protein affinity towards surfaces [151]. By understanding protein-surface interactions, researchers can modify surfaces to attract a particular class of proteins that prevent other coagulation-stimulating proteins from interacting with the surface.

Fig. 9.

First widely explored in the field of nanomedicine (A), pre-coating nanomaterials with select protein classes can modulate the protein corona that forms, assisting in evading immunosurveillance. Similarly, pre-loading biomaterials (B) with particular protein classes can tailor the plasma proteins that are adsorbed to the surface, altering the fate of the device. Adsorption or immobilization of select passive protein layers (C) on the surface of biomaterials reduces non-specific protein adsorption to and platelet interaction with the surface and averts an immune response by decreasing leukocyte adhesion, increasing the hemocompatibility of the biomaterial interface.

Recent research has moved into directly immobilizing proteins onto the surface prior to use [146,152]. Dysopsonins, a class of proteins including albumin and apolipoproteins, can shield nanoparticles from random interaction and internalization by phagocytes [149,153]. Pre-coating materials with dysopsonin proteins has shown to be a promising strategy for directing biological response [154]. Dysopsonin-modified surfaces carry the potential to create a more natural blood-surface interface by reducing non-specific protein interaction and platelet activity [155].

Despite protein coatings being an emerging biomaterial modification, albumin nanoparticles are already on market [149]. Albumin, at a concentration of 35–53 g L⁻¹, is the most abundant protein found in the blood [12]. When used to coat biomedical devices, albumin significantly increases the biopassive behavior by resisting replacement by alternative proteins [156]. Albumin has been immobilized onto polymer surfaces to reduce platelet adhesion and fibrinogen adsorption, extend blood coagulation times, and reduce thrombus formation by ~ 50% [146,156,157]. Krajewski et al. showed that albumin-coated neurovascular stents performed similarly to heparin-coated stents, preserving platelet counts and reducing platelet factor (PF)-4 activation (pro-thrombotic chemokine released from platelets) [158].

More recently, apolipoprotein J is of particular interest due to its contribution to the stealth effect in nanoparticles. More commonly known as clusterin, this protein is a 75–80 kDA heterodimer that controls cell–cell interactions, regulates apoptosis, and polices the transportation of lipids [159]. While some researchers have chosen to pre-coat their materials, proteins can be cost-ineffective and may not be a viable option for commercialization. Tunable surface chemistries may help dictate which proteins will have a greater affinity to the surface - that is, rather than pre-coating a surface with proteins prior to use, the surface of a biomaterial can be modified to attract desired proteins in vivo, thereby generating the preferred protein layer after exposure. Researchers have been able to dictate which proteins adsorb to surfaces by controlling the hydrophobicity, surface roughness, and net charge of surfaces [160]. Researchers have found that modifying materials with surface chemistries that support selective clusterin adherence achieves the same effect as pre-coating the surfaces [147,161]. Schottler et al. and Aoyema et al. demonstrated that PEGylated and poly(phosphoester) (PPE) ylated nanoparticles, which have an affinity for clusterin, showed a sharp decrease in non-specific cellular uptake when pre-exposed to plasma proteins including clusterin [147,162]. Saito et al. hypothesized that clusterin adsorption on poly(2-methoxyethylacrylate)-modified ECC circuits resulted in reduced complement activation [163]. However, a major weakness that remains with pre-adsorbed or preferential natural protein adsorption is that these surfaces would still be subject to Vroman effects over time. Further studies utilizing clusterin for device coatings could elucidate more information on its potential for creating a hemocompatible interface.

Although understanding the effects of different protein adsorption profiles is important for realizing how to improve hemocompatibility, coupling proteins may be necessary for longer-term applications [164,165]. However, such covalent immobilization strategies remain subject to protein degradation over time, which can lead to subsequent platelet adhesion and other thrombotic events despite the initially elevated resistance to platelet-surface interactions [166]. Surface protein coupling for the sake of prevention of displacement by Vroman effects should therefore only be used for short-term blood-contacting applications in order to avoid aging-induced degradation of the coupled protein layer that ultimately compromises its protective properties [167].

Due to its stable formation of amphiphilic and antifouling monolayers, hydrophobins, a class of proteins derived from filamentous fungi, have been used to coat biomedical devices, resulting in decreased secondary protein adsorption [168] and decreased adhesion of biological matter on biliary stents [169]. More recently, Devine et al. fabricated dual hydrophobin-coated/nitric oxide-releasing PDMS, which reduced fibrinogen adsorption after 90 min in a fibrinogen solution and the number of adherent platelets following 90 min of in vitro exposure [170]. Although protein coatings have been shown to decrease coagulation and protein adsorption, this method does not prevent platelet activation. To circumvent this, the hydrophobin-coated/nitric oxide-releasing surface [170] along with albumin-coated/heparin-immobilized biomaterials [155,171] have been some of the first to combine passive protein-immobilized surfaces with active antithrombotic strategies, resulting in synergistic hemocompatible activity. More details on nitric oxide’s and immobilized heparin’s role in improving hemocompatibility are covered in the bioactive section of this review.

Despite the success of protein-coated surfaces in a lab setting, several drawbacks still need to be tackled including long-term storage, sterilization, uncontrollable conformational changes, degradation, and cost [12]. Moreover, due to the Vroman effect, proteins can be displaced over time, altering the composition of proteins attached to the surfaces and minimizing long-term efficacy of these devices [16,172]. Bio-inspired protein-coating technology is a promising answer to increasing the hemocompatibility of biomaterials, but more progress is needed to successfully translate these materials.

3. Current bioactive methods for improving surface hemocompatibility

Another method of mitigating clot formation is the incorporation of bioactive antithrombotic vehicles. Such agents present on or locally released from biomaterials can react with the surrounding environment, directly interfering with protein adsorption or platelet adhesion and aggregation, thereby preventing thrombus formation (Fig. 10). The following section will discuss bio-inspired bioactive surface additives that improve hemocompatibility, as well as some of the drawbacks and limitations of each technology.

Fig. 10.

Bio-inspired active surface strategies to promote hemocompatibility: (1) Nitric oxide-releasing/generating surfaces, (2) antithrombotic polysaccharide surfaces, and (3) thrombin-inhibiting peptide surfaces.

3.1. Nitric oxide (NO)-releasing/generating surfaces

Nitric oxide is a gaseous free radical secondary messenger that attenuates the binding of platelets to adsorbed plasma proteins such as fibrinogen and vWF, thereby preventing their activation [173–175]. NO is also known to mediate immune response, with potent antimicrobial and antiviral properties [176,177]. The endothelial lining of blood vessels consists of a monolayer of cells that continuously synthesize NO from L-arginine at an estimated flux ranging between 0.5 and 4.0 × 10⁻¹⁰ mol cm⁻² min⁻¹, whereby NO and its metabolic byproducts reduce platelet activation by the downregulation of P-selectin expression [178–181]. Biomimetic NO-functionalized surfaces aim to sustain NO fluxes comparable to the native endothelium and have been developed via two main strategies (Fig. 11): (1) release from materials containing an NO donor reservoir and (2) catalytic surface generation of NO from a supplemented or endogenous NO donor.

Fig. 11.

Overview of NO-releasing and NO-generating materials for blood-contacting applications. Similar to the endothelium, controlled NO generation or release from biomaterials into the blood modulates platelet adhesion and aggregation.

NO donors are generally classified as directly or physiologically activated, with some of the most synonymous direct donors being diazeniumdiolates (NONOates) and S-nitrosothiols (RSNOs) such as S-nitrosocysteine (CysNO), S-nitrosoglutathione (GSNO), and S-nitroso-N-acetyl-DL-penicillamine (SNAP). NONOates are known to release NO via enzymatic, chemical, or thermal decomposition with kinetics tunable by the pH microenvironment [182]. RSNOs have variable long-term stability based on steric effects, with NO release mediated by thermal decomposition [183], photolysis [184], and metal ion catalysis [185]. Schoenfisch group provided detailed reviews of various NO macromolecular scaffolds, their applications, and detection of NO release [186–188].

Nitric oxide-generating (NO-gen) materials have been developed from the premise that the NO release from endogenous and supplemented RSNOs is catalyzed by copper ions [189] and other metal ions [190–192]. More recently, work with metal–organic frameworks (MOFs) has improved the stability of metal ions in materials and invigorated interest in developing long-term, self-sustaining materials that can generate NO from endogenous RSNOs [193]. In 2012, Harding et al. demonstrated that the MOF Cu₃(BTC)₂ (BTC: 1,3,5-benzentricarboxylate) shows catalytic activity with CysNO but is limited by its instability under physiological conditions [193]. Follow-up work addressed this structural instability with the alternative MOF Cu (II) 1,3,5-Benzene-tris-triazole (CuBTTri) which generated NO both in a heterogeneous state and when composited within PUs [194]. Recent hemocompatibility studies demonstrate that CuBTTri with supplemented GSNO reduces thrombus formation in human whole blood and inhibits platelet aggregation in PRP [195]. However, despite NO-gen’s promising in vitro capability in these materials to achieve antiplatelet activity, few studies have tested these materials in vivo.

Many polymers and inorganic nanoparticles have been infused with NO donors via blending, solvent swelling, solvent evaporation, and covalent immobilization [196,197]. Early work in developing nitric oxide-releasing (NO-rel) materials for blood-contacting medical devices and their testing in rabbit models are summarized in separate reviews by Frost el al. [198] and Major et al. [199]. Contemporary work has focused on RSNO donors to prolong NO release. Brisbois et al. developed extended antithrombotic surface functionalization via the blending of SNAP into Elast-eon E2As PUs, which showed a three-fold reduction in thrombus area in a seven-day sheep ECC model [200]. Further work developed SNAP-doped silicone tubing which showed > 60% platelet preservation relative to controls in a 4-h rabbit ECC model, while multi-lumens filled with a SNAP-PEG blend showed a 55% reduction in thrombus formation over an extended 11-day rabbit model [201,202]. Wo et al. blended SNAP into a polycarbonate urethane silicone elastomer under the tradename CarboSil 2080 A^™ and achieved a six-fold reduction in thrombus area in a 7 h rabbit ECC model [203,204]. Other hybrid strategies have considered copper [205,206], selenium [207], and zinc oxide [191] nanoparticle topcoats to catalyze the NO release from embedded RSNO-infused CarboSil. Further work with CarboSil composites of GSNO and copper nanoparticles applied to PVC tubing has demonstrated its viability for ECC applications, wherein for a 4-h rabbit model the combination of copper and GSNO maintained almost 90% of the baseline platelet count [206].

Although NO-rel strategies provide highly tunable release, NO donor impregnation has five main problems: (1) donor leaching can be cytotoxic [201,208]; (2) increasing donor concentration does not directly correlate to higher stabilized NO fluxes [202,208]; (3) limited donor storage capacity[203]; (4) thermal lability of NO donors limits shelf-life and viable sterilization techniques [208,209]; and (5) NO release from doped materials cannot prevent biofouling of surfaces [210]. NO-rel materials may illicit a cytotoxic effect through the leaching of RSNOs, their synthetic precursors (RSHs), and their disulfide byproducts (RSSRs). Prior work by Wo et al. has demonstrated that RSNOs may leach nearly 5x as much as either RSH or RSSRs, dependent on the microcrystalline state of the RSNO within the polymer [203]. Several studies have since reported cytocompatibility with decreased platelet activation and donor leaching (<5 wt% of the initially loaded RSNO compound) [83,211].

To stabilize NO release profiles and prevent cytotoxicity, topcoats composed of hydrophobic polymers have been added to NO-rel polymers to modulate diffusion of water into NO-doped layers and subsequent leaching of NO donors [203,208,212]. Strategies that covalently attach NO donors to polymer matrices have also shown success [86,213]. One such strategy based on early work by Frost et al. [184] involves covalently attaching SNAP to silicone rubber, whereby almost 125 days of stabilized NO release was achieved alongside a ~36% improvement in platelet count in a 4-h rabbit model [86]. Most recently, Goudie et al. incorporated alkylamine spacers with a methacrylate brancher onto PDMS to enable denser SNAP attachment with stabilized NO release over 25 days as well as a three-fold reduction in fibrinogen adsorption and a nine-fold reduction in platelet adherence over unmodified PDMS [214]. Ongoing work with covalently attached SNAP and other NO moieties shows promise for the increased efficacy of NO-rel materials.

NO functionalization as an active strategy for preventing platelet activation cannot attenuate blood plasma protein adsorption onto polymeric surfaces. Therefore, multi-defense strategies are required [215]. One such strategy is to mask the surface of hydrophobic NO donor-containing films with a hydration layer, whereby hydrophobic proteins have shown decreased adsorption [216,217]. SNAP-doped CarboSil 2080A base layers top-coated with hydrophilic polymers (e.g., SP60D60 and SG80A) [210] or grafted zwitterions [105,218] have shown resistance to biofouling. Other works have used immobilized heparin [219,220], methods that confer an enhanced anti-platelet and anti-biofouling effect through hydrophobin protein topcoating [170], or liquid infusion via the fabrication of silicone oil-infused, SNAP-doped silicone [83,221]. Ongoing work continues to bridge passive strategies for preventing surface biofouling with active NO release strategies to mitigate platelet activation.

3.2. Antithrombotic polysaccharide-incorporated surfaces

To improve the surface hemocompatibility of biomedical devices, polysaccharides have been grafted onto the surface of blood-contacting materials. Several classes of polysaccharides have exhibited low protein adsorption activity, low platelet aggregation, and no adverse effects on red blood cells [222,223]. Depending on their composition, polysaccharides can be broken down into two classes: homopolysaccharides and heteropolysaccharides [224]. Although some polysaccharides in both classes have been reported to improve the hemocompatibility of materials, most work has been reported with heteropolysaccharides, particularly glycosaminoglycans (GAGs) (Fig. 12).

Fig. 12.

Structures and efficacy of glycosaminoglycans that have been used to render surfaces antithrombotic including heparin (A), heparan sulfate (D), dermatan sulfate(G), and hyaluronic acid (J). Antithrombotic glycosaminoglycan coatings on different medical devices and biomaterials have shown significant reductions in thrombotic complications in vitro and in vivo. Heparin-coated (Hep-Th-P-PPAm) 316L stainless steel stents (C) reduced fibrin sheath formation in a canine iliac artery model after 90 days of implantation compared to bare (B) stents [226]. Biomimetic heparan sulfate (HS)-like coated ePTFE vascular grafts (F) minimized fibrin deposition, thrombosis, and neointimal hyperplasia compared to unmodified grafts (E) in a 20 week ovine carotid model [234]. Increasing concentrations of dermatan sulfate (DS) (H-I) incorporated into a polyurethane backbone significantly decreased protein adsorption and platelet adhesion [229]. Similarly, the hemocompatibility of polyurethane improved after hyaluronic acid (HA) functionalization, decreasing protein adsorption and platelet adhesion (K-L) [247]. (B & C are reprinted/adapted from Biomaterials Yang et al., The covalent immobilization of heparin to pulsed-plasma polymeric allylamine films on 316L stainless steel and the resulting effects on hemocompatibility, 31(8): 2072–2083, Copyright (2010), with permission from Elsevier. E & F are reprinted/adapted from Annals of Vascular Surgery, Wulff et al., Biomimetic Heparan Sulfate-Like Coated ePTFE Grafts Reduce In-graft Neointimal Hyperplasia in Ovine Carotids, 40: 274–284, Copyright (2017), with permission from Elsevier. H & I are reprinted/adapted from Macromolecular Science, Xu et al., Polyurethane/Dermatan Sulfate Copolymers as Hemocompatible, Non-Biofouling Materials, 11: 256–266, Copyright (2011), with permission from Wiley. K & L are reprinted/adapted from Biomaterials, Chuang et al., Regulation of polyurethane hemocompatibility and endothelialization by tethered hyaluronic acid oligosaccharides, 30: 5341–5251, Copyright (2009), with permission from Elsevier.).

Although many have been explored, two strategies have emerged as superior surface modification methods involving hemocompatible polysaccharides: layer-by-layer self-assembly and grafting by covalent immobilization. Layer-by-layer methods involve applying layers of oppositely charged species and are particularly attractive due to their inexpensive and straightforward synthesis [225], but have some drawbacks including lack of robustness and decreased biological activity [226]. To increase stability, researchers have begun covalently bonding polysaccharides to the surface of biomaterials but have limitations such as an increase in device cost, undesired and uncontrollable adverse chemical reactions, and loss of biological activity [225]. This review will examine the use of both homo- and heteropolysaccharides for improving the hemocompatibility of biomedical devices, the limitations that have been met, and improvements that can be addressed in the future.

3.2.1. Antithrombotic heteropolysaccharides

Glycosaminoglycans are a group of long-chain, unbranched heteropolysaccharides that have been widely used for the fabrication of hemocompatible surfaces due to their anticoagulant and antithrombotic properties [227]. Present in all mammalian tissue, GAGs are diverse in both structure and function. Originally believed to only play a role in cellular structure and hydration, GAGs are now understood to regulate many biological activities including anticoagulation [228]. Due to their high hydrophilicity and negative charge, several prevalent GAGs have been shown to improve the hemocompatibility of devices [229].

As the most commonly administered drug in hospitals, heparin has widely been used as a systemic anticoagulant since 1935 [227,230]. Structurally related to the proteoglycan heparan sulfate found on the cell surface and extracellular matrix [230], heparin blocks several serine proteases including thrombin by increasing the effect of serine protease inhibitors, ultimately halting thrombus formation [231]. The anticoagulant activity of heparin-bound surfaces is attributed to its binding to antithrombin, resulting in the inactivation of thrombin and fXa (Fig. 13) [232]. An in-depth analysis of the adsorption profile of plasma proteins also revealed that heparin-coated biomaterials have a decreased affinity for particular blood proteins including C3, fibronectin, and fibrinogen [232]. Therefore, heparin/heparan sulfate immobilization has been widely used to improve the hemocompatibility of polymers [230,233–236].

Fig. 13.

Heparin-bound surfaces exhibit anticoagulant activity through the binding of antithrombin, which inactivates several coagulant enzymes including thrombin and factor Xa.

Because of its promising antithrombotic effects, heparin-coated surfaces have been extended to in vivo medical device applications. Yang et al. demonstrated after implantation into dogs for 30 and 90 days that heparin-coated 316L stainless steel (SS) stents reduced the occurrence of thrombosis while still allowing the adhesion and proliferation of endothelial cells [226]. Although immobilized heparin has shown effectiveness in suppressing thrombin activity, heparin-incorporated surfaces still do not prevent other pro-coagulant activity including platelet activation. To decrease platelet adhesion and activation, Devine et al. immobilized heparin to nitric oxide-releasing surfaces, which synergistically improved the hemocompatibility of silicone rubber in a 4-h ECC rabbit model [219]. Similarly, Qiu et al. showed that the combination of immobilized heparin and selenocystamine (SeCA), a nitric oxide-generating agent, on cardiovascular stent surfaces lowered thrombus weight and maintained blood flow in an ex vivo circulation thrombogenicity rabbit model [237]. Likewise, heparin has been combined with other strategies including albumin [171], collagen [238,239], tanfloc [240], and nitric oxide-releasing [219,220] and generating [237] surfaces. However, several limitations still need to be addressed including high costs, decreased activity once immobilized, and reliance on antithrombin, which varies in levels across critically ill populations [225,241].

Found in the extracellular matrix of mammalian tissues, dermatan sulfate (DS) is both anionic and highly hydrophilic, thereby showing potential for improving antithrombotic activity [229]. Dermatan sulfate, structurally similar to heparin, reduces thrombotic activity by increasing the effectiveness of heparin cofactor II, a serine protease inhibitor, on thrombin [231]. In fact, in the presence of dermatan sulfate, heparin cofactor II-mediated inhibition of thrombin increases 1000-fold [242]. Although research using DS for hemocompatible materials is limited, polymers coated with DS have been shown to reduce non-specific cellular adhesion and protein adsorption [229]. Dermatan sulfate has been used to improve the hemocompatibility of polymers such as PU [229] and PET [243]. Like many polysaccharide-based coatings, methods of incorporating DS without losing activity or longevity of the device are limited. Practical manufacturing and sterilization testing remain to be thoroughly established. Moreover, further in vitro and in vivo testing is needed to assess the hemocompatibility of different DS-incorporated polymers to validate for blood-contacting applications.

Also located in the extracellular matrix, hyaluronic acid (HA) is a non-sulfated GAG that regulates cell adhesion, differentiation, and cell migration [30]. The antifouling capabilities of HA date back to 1999, when Morra and Cassineli demonstrated that HA-coated PS maximized interaction with water, thereby decreasing cellular and bacterial adhesion [244]. This phenomenon has been further extended to reducing protein adsorption and aggregation and inflammation [245]. To improve surface hemocompatibility, HA has been immobilized onto the surfaces of common biomaterials such as metal alloys, PU, titanium, and 316L SS [30,246–248]. PUs combined with HA showed greater effectiveness in reducing protein and platelet adhesion compared to heparin- and PEG-grafted surfaces for vascular graft applications [247]. Similarly, the immobilization of HA on stents implanted in baboons significantly decreased platelet aggregation and thrombus formation [248]. In vivo subcutaneous implantation of HA-based materials in a 4-week mouse model decreased scar tissue formation as a result of decreased non-specific protein and cell attachment [249]. However, although several studies have shown that HA-modified surfaces reduced protein adsorption and platelet adhesion, an exact antithrombotic mechanism has yet to be identified outside of general antifouling activity. Therefore, in addition to robust manufacturing and sterilization improvements, a better understanding of HA-modified material-blood interactions is needed.

3.2.2. Antithrombotic homopolysaccharides

First investigated as a plasma alternative in the 1940s [250], dextran is a hydrophilic, biodegradable branched homopolymer of glucose used to increase the hemocompatibility of surfaces by reducing platelet adhesion and protein adsorption, thereby reducing clot formation [222,251,252]. Produced by certain Leuconostoc and Streptococcus bacteria, dextran is an effective antithrombotic agent, enhancing fibrinolysis, reducing platelet adhesion, and decreasing platelet activation [253]. When bound to a variety of biomaterials such as glass, PET, cobalt-chrome, activated carbon, and nano-scaled biomaterials, dextran has improved antifouling properties by reducing cell, protein, and platelet adhesion [222,223,254,255]. Moreover, dextran-functionalized antimicrobial wound dressings have reduced hemolytic activity and prolonged clotting times [256]. Similarly, Xu et al. demonstrated that dextran- and chitosan-multilayered coatings resulted in reduced albumin adsorption and bacterial adhesion [257]. To improve the hemocompatibility of medical-grade polycarbonate, a common material for hemodialyzers, blood pumps, and oxygenators, Gupta et al. altered the poly-carbonate surface with a dextran-modified poly(vinyl amine) surfactant coating, resulting in a 90% reduction in platelet adhesion under stable flow in whole blood compared to untreated materials [258]. Dextran has also been used in combination with other carbohydrates, molecules, and materials to improve hemocompatibility by reducing protein adsorption, platelet adhesion/activation, and leukocyte attachment as well as prolonging coagulation time [252,259,260].

Despite promising antifouling characteristics, the long-term stability of dextran-coated materials has been rarely addressed [261]. In response, carboxymethyldextrans (CMDs) have been covalently grafted onto material substrates and have shown promising anti-fouling activity against certain cells and proteins [261–266]. As a result of their antifouling characteristics, Michel et al. found CMD-grafted PTFE films demonstrated improved antithrombotic activity in a clotting time assay while improving endothelial cell adhesion [261]. CMD-modified surfaces exhibit steric-entropic repulsive interactions that ultimately repel proteins [262]. However, much like heteropolysaccharide-modified surfaces, dextran-based interfaces face several challenges including manufacturing challenges and lack of robustness. Moreover, there is little research that has evaluated the sterilization and storage stability of these materials.

3.3. Immobilization of thrombin-inhibiting peptides

While many bioactive peptides or peptide-inspired structures have been identified as direct thrombin inhibitors [267], few have been immobilized onto blood-contacting devices. Hirudin, a polypeptide extracted from leech salivary glands, has been recognized as a potent direct thrombin inhibitor but poses a higher bleeding risk than heparin [268]. Like other bivalent inhibitors, hirudin inhibits thrombin by directly binding to the exosite 1 and reactive site of thrombin. Several approaches have tried immobilizing recombinant hirudin but have failed to maintain antithrombotic properties [268]. Lepirudin and desirudin, derivatives of hirudin, have been synthesized to promote similar anticoagulant behavior while prolonging biological activity and stability. Though their affinity for thrombin is 10x weaker than hirudin, these derivatives are still considered potent thrombin inhibitors [269]. Lepirudin has been shown to readily adsorb to PDMS, resulting in more thromboresistant catheters [270]. Analogs based on hirudin have also been engineered to reduce thrombin activity. These analogs contain synthetic peptides composed of D-Phe-Pro-Arg which mimic fibrino-peptide A binding [271]. Synthetic peptides such as D-Phe-Pro-Arg-Pro-Gly [271] and D-Phe-Pro-Arg-chloromethylketone (PPACK) [241] have been successfully immobilized in different biomaterials, resulting in decreased thrombin adsorption and delayed coagulation time. As an advantage compared to heparin-coated surfaces, these inhibitors directly block thrombin and do not rely on the presence of antithrombin, whose concentrations may vary in critically ill patients [241]. However, PPACK is an irreversible inhibitor, binding to and immobilizing thrombin permanently; therefore, the surface will eventually reach a saturation point, limiting the efficacy of these materials for long-term applications [241]. Moreover, PPACK-immobilized surfaces showed an elevated inflammatory response via increased C5a and increased leukocytes on the surface, suggesting that it is a potential activator of complement-driven leukocyte activation [241]. Therefore, more consideration must be made to PPACK-modified surfaces for hemocompatibility.

Perhaps the most extensively used analog of hirudin for the creation of antithrombotic materials is bivalirudin (BV), a short, 20-amino acid synthetic polypeptide highly selective for thrombin and approved by the FDA in 2000 [268,272]. Like hirudin, BV is a bivalent direct thrombin inhibitor composed of two functional regions: an N-terminal region, which binds to thrombin’s catalytic site, and a C-terminal dodecapeptide, which blocks the fibrinogen-binding region of thrombin (Fig. 14) [268,273]. Unlike heparin, BV does not require any cofactor and does not induce platelet activation [272]. The first reported BV-immobilized surface was by Lu et al. in 2012, who demonstrated that BV-modified 316 SS stents resulted in a slight reduction in platelet activation and significantly increased the clotting time compared to control surfaces [268]. Since then, other studies on BV-coated stents have found similar hemocompatible activity. BV-plasma polymerized allylamine (PPAm) stents implanted in dog femoral arteries for five weeks showed less thrombus formation and greater endothelial attachment and growth compared to control stents [272]. BV-eluting stents via a polydopamine (PDAM)-TiO₂ nanotube platform resulted in decreased thrombin and platelet activity for over 60 days [274]. A BV-immobilized phytic acid coating on biodegradable magnesium showed decreased clotting times, improvement of endothelial proliferation, and increased antiproliferative effects towards SMCs [275]. Moving forward, clinical studies are needed to validate BV-functionalized surfaces for improved hemocompatible activity.

Fig. 14.

Direct thrombin inhibitors can be classified as univalent or bivalent, depending on which sites on thrombin they bind to. Univalent inhibitors only bind to the active site, while bivalent inhibitors including hirudin and bivalirudin bind to the active site and exosite 1.

In addition to hirudin-based technologies, other thrombin-inhibiting peptides derived from animal sources have been used for the fabrication of hemocompatible surfaces. Boophilin, a direct thrombin inhibitor from Rhipicephalus microplus [267], was immobilized for the first time by Freitas et al. in 2012 onto self-assembled monolayers of alkanethiols on gold via neutravidin, resulting in delayed plasma coagulation and decreased thrombotic activity [276]. Surfaces grafted with ACH₁₁, a peptide derived from Agkistrodon acutus venom hydrolysates, decreased activated FXa, fibrinogen concentrations, and platelet adhesion [277,278]. Vascular grafts with immobilized ACH₁₁ and cell-adhesive Cys-Ala-Gly peptide decreased neointimal formation and presented no blood coagulation in an in vivo 6-week rabbit carotid artery implant model [278]. Antithrombin present in mammalian circulatory systems [267] has also been used by Klement et al. for the fabrication of an antithrombin-heparin complex surface coating [279]. Antithrombin and heparin activity have been recognized as interdependent – that is, heparin can enhance antithrombin activity by 1000-fold [267]. Similarly, heparin, when bound to antithrombin, accelerates its interaction with serine proteases, resulting in consistent antithrombotic behavior [279]. Antithrombin/heparin-coated PU-based endoluminal grafts best reduced fibrin formation and clot weight in an in vivo rabbit model [279]. Immobilized argatroban, a peptidomimetic synthetic direct thrombin inhibitor [280], on biomaterial surfaces resulted in prolonged clotting times [281] and decreased thrombin activity and thrombus formation in 4-h rabbit ECC models [282,283]. However, similar to heparin-coated surfaces, direct thrombin inhibitors can suffer from reduced activity once immobilized to the surface.[241] Future work of direct thrombin-inhibiting peptide coatings should be evaluated in clinical trials to determine the efficacy of these materials in humans.

4. Current promotion of endothelial cell growth strategies for improving surface hemocompatibility

There is no doubt that the native endothelial surface is the most hemocompatible surface that exists. Measuring at thicknesses as low as 0.2 μm, a healthy endothelium acts as a thin yet highly effective barrier between blood and tissues [284]. Endothelial cells regulate the release of NO, tissue factor, and thrombin inhibitors, effectively preventing thrombus formation. Often used as a final marker for the degree of wound healing, the endothelialization of biomaterials effectively blocks the blood from interacting with the implant. Considered essential for long-term devices and permanent implants, this section discusses bio-inspired methods for improving endothelialization as well as the importance of developing a barrier for medical devices subjected to long-term implantation (Fig. 15).

Fig. 15.

Lining the inner layer of blood vessels, the endothelium exhibits several anticoagulant and antiplatelet mechanisms under normal conditions to prevent blood coagulation.

4.1. Endothelial progenitor cell-functionalized surfaces

Endothelial cells (ECs) express heparan sulfate proteoglycans, tissue plasminogen factor, and NO to prevent platelet plug formation and inflammatory response [285]. However, although pre-coating materials with a layer of endothelial cells appears to be a promising approach to developing thromboresistance, considerable disadvantages include detachment under flow, insufficient coverage, phenotype alterations, and contamination during the long harvesting periods of autologous ECs [286–289]. An alternative strategy to circumvent these problems involves the in situ endothelialization of implanted devices with endothelial progenitor cells (EPCs). Mainly characterized by CD133, CD34, and vascular endothelial growth factor receptor-2 (VEGFR-2) [290–293], EPCs are capable of differentiating into mature ECs. EPCs have often been investigated for treating sites of vascular damage, whereby EPC recruitment has been shown to restore the function of ischemic tissue via re-endothelialization stimulation while also modulating vasculogenesis and angiogenesis in hypoxic tissue [285]. This section focuses on the biomedical application of EPC coatings on polymer surfaces through EPC seeding and EPC capturing to create blood vessel-like surfaces that are hemocompatible and provide a microenvironment suitable for outward cell migration and repopulation of surrounding tissue (Fig. 16).

Fig. 16.

Role of endothelial progenitor cells (EPCs) in the formation of new blood vessels through vasculogenesis and angiogenesis (A). Vasculogenesis is the de novo development of vessels from endothelial progenitor cells (EPCs), whereas angiogenesis is the formation of blood vessels from existing blood vessels. In an effort to create a completely hemocompatible surface, researchers have aimed to endothelialize the surfaces of medical devices such as stents and vascular grafts through two different methods: EPC seeding (B) and EPC capturing (C).

4.1.1. EPC seeding

EPC-seeded scaffolds based on synthetic or natural polymers have seen development since the early 2000s with refinement in EPC isolation, stent design, and biocompatibility testing. The two main approaches to EPC seeding are directly seeding isolated EPCs onto surfaces or using stimulated isolated EPCs to increase EPC homing to the surface. In 2003, Griese et al. demonstrated that EPCs isolated from New Zealand White Rabbits could be cultured to confluency in under three weeks and induce endothelialization of a fibrinogen-coated polytetrafluoroethylene graft, with patency maintained over the four weeks of the study [294]. The combined work of Shirota group around the same time developed EPC-seeding technologies around photocured gelatin-covered stents in a canine carotid artery model [295–298]. Matsumura et al.’s clinical work first demonstrated the versatility of implanted EPC-seeded grafts, showing no postoperative ruptures [299]. Overall, much of the early pioneering research showed favorable in vivo results [291,300,301]. Autologous EC seeded vascular grafts in parallel have consistently shown no significant improvement in patient outcomes compared to autologous vessels and may present issues in long-term graft patency [302]. However, to better affect graft remodeling and endothelialization potential, EPC seeded grafts have taken precedence, based on the need for EPC recruitment and maturation to produce the necessary cell lineages to create stabilized blood flow. The main challenges still present with in vitro seeding of EPCs are as follows: (1) low counts of circulating EPCs in the peripheral blood require large volumes of blood to be harvested for therapeutic use; (2) harvesting EPCs requires one or more invasive surgical procedures; and (3) patients with cardiovascular risk factors have even lower numbers of circulating EPCs [303,304].

4.1.2. EPC capturing

Because of the time constraints and difficulties with culturing isolated EPCs, contemporary interest has shifted towards capturing methods that conjugate scaffolds with homing moieties to increase the affinity of circulating EPCs toward artificial surfaces such as vascular grafts and stents (Fig. 17). Advances in chemical biology over the past twenty years have motivated new bio-inspired synthetic surfaces via the conjugation of different biological small molecules, oligomers, and synthetic derivatives onto scaffolds with application-specific advantages for EPC homing.

Fig. 17.

Artificial surfaces can be modified with molecules (antibodies, aptamers, etc) that selectively bind EPCs, creating a barrier between blood and the foreign body. Blocked vessel passageways (A) stented with an untreated stent frequently result in late in-stent restenosis as a result of delayed healing. However, using EPC capturing techniques (D), modified devices can rapidly attract circulating EPCs (C1), attaching to the surface (C2) and eventually differentiating into a mature EC lining (C3).

Hristov et al. discovered in 2003 that out of three surface markers that identify EPCs – CD34, CD133, and VEGFR-2, only CD34 and VEGFR-2 remain positive after EPCs enter circulation [290]. Therefore, many strategies have immobilized these two surface markers onto scaffold surfaces to attract EPCs. Rotmans et al. demonstrated that prosthetic grafts coated with anti-CD34 monoclonal antibodies exhibit rapid and complete endothelialization in an in vivo porcine model study [305]. Moreover, CD34+ antibody-containing stents have shown considerable success in clinical trials. The Healthy Endothelial Accelerated Lining Inhibits Neointimal Growth-First In Man (HEALING-FIM) study was the first set of clinical investigations with bioengineered EPC capture stents based on a polysaccharide coating with immobilized anti-human CD34 antibodies on a stainless steel stent. These studies demonstrated the feasibility of EPC capture stents, with no stent-related thrombosis observed after six months and rapid buildup of functional endothelial layers with minimal inflammation [306]. The more recent HEALING-II study demonstrated the viability of Genous^™ stents coated with anti-human CD34+ antibodies and found variability in restenosis and cardiac event outcomes based on patients’ EPC numbers, the use of statin therapy, and other conditions [307,308]. Excellent reviews and commentaries by Sethi et al. and Leopold et al. highlight much of the work that various clinical trials have done to improve outcomes for anti-human CD34+ functionalized stents for EPC capturing [309,310].

To date, there are few in vivo applications of immobilized VEGF peptide and VEGFR-2 antibody surfaces for EPC capturing. Melchiorri et al. applied immobilized heparin chemistry to develop separate vascular grafts containing VEGF peptides and anti-CD34 antibodies on biodegradable polyester vascular grafts which were implanted as inferior vena cava channels in mice models and developed superior EPC attachment and endothelialization [311]. Follow-up studies that covalently immobilized VEGF peptides better captured ECs and EPCs due to reduced leaching [312,313]. Li et al. further considered pairings of VEGF peptides and anti-CD34 antibodies in layer-by-layer membrane assemblies with heparin on titanium implants [314]. This study demonstrated further that the combination of anti-CD34 with VEGF had superior EPC attachment under in vitro flow of PRP and exhibited spontaneous endothelialization of cardiovascular implants with decreased thrombosis in in vivo canine testing [314].

Aptamers, oligonucleotides, and peptides that show a high binding affinity to target molecules including CD34+ cells have also been used to capture EPCs. Hoffmann et al. were the first to demonstrate the viability of generating aptamers with high affinity towards porcine CD31+ EPCs via systematic evolution of ligands through exponential enrichment [315]. Grafting several aptamer candidates onto blood-compatible star-PEG-coated polymeric discs allowed for the adhesion of EPCs in an in vitro porcine model and subsequent endothelialization of the surface within 10 days in culture [315]. Chen et al. used CD133+ for EPC capture and endothelialization of tissue-engineered blood vessel scaffolds grafted into the right common carotid artery of diabetic mice, with increased patency observed thirty days post-implantation [316]. Peptide aptamers on PDAM-coated titanium surfaces and oligonucleotide aptamers on heparin surfaces have further demonstrated the ability to pattern scaffolds to influence the distribution of EPCs on stent surfaces [317,318].

To increase endothelial cell and EPC adhesion to blood-contacting devices, many studies have considered peptides derived from endothelial matrix proteins that influence endothelial cell and EPC adhesion and proliferation, such as fibronectin-derived RGD (Arg-Gly-Asp) peptides and laminin-derived YIGSR (Tyr-Ile-Gly-Ser-Arg) peptides [319,320]. In an early study conducted in 2002, Walter et al. demonstrated that the combination of simvastatin and cRGD injection in balloon-injured mice led to increased EPC induction onto damaged carotid arteries and accelerated re-endothelialization [321]. In response, a variety of polymers have been functionalized by RGD motifs such as chitosan [322], ePTFE [323], polycaprolactone(PCL)/collagen [324], and PU [325] to increase EPC homing and help render the material less thrombogenic. Blindt et al. developed cRGD-eluting metallic stents with a BioSpan^® segmented PU topcoat loaded with cRGD [326]. Tested in porcine carotid artery models, the fabricated stents showed increased cRGD tissue levels after four weeks with increased early recruitment of infused EPCs, resulting in reductions in neointimal area and stenosis after 12 weeks [326]. More recently, Zheng et al. extended this principle with electrospun PCL grafts top-coated with RGD peptides, which resulted in revascularization in a rabbit carotid artery model [327]. Across these studies, a common concern was the potential of the RGD motif to also attract platelets. The RGD motif is found in a variety of plasma proteins, including vWF, fibronectin, and fibrinogen. Therefore, one critical limitation of using RGD-binding motifs for EPC homing is that RGD-functionalized surfaces will perpetuate competition between rapid endothelialization and platelet adhesion [328]. To ensure preference for endothelialization, especially in small diameter vascular grafts, other groups have considered the co-immobilization of heparin [329] with RGD and YIGSR moieties to attenuate initial platelet adhesion as well as synthetic peptides with RGD motifs [330] or peptides selected for low platelet affinity [331].

Despite the many promising design strategies for EPC capturing hemocompatible surfaces, especially for coronary stent applications, there remain some critical issues with their performance when compared to untreated stents or other biopassive and drug-eluting alternatives. Early EPC capture stent trials demonstrated no significant reduction in neointimal hyperplasia because of EPC capture compared to traditional bare-metal stents [306]. These results highlighted one critical limitation of EPC capture for stents – the coating covers only stent struts, and therefore, interstrut space is not expected to exhibit early functional endothelial lining [306]. Further concerns have been raised about the sterilization methods for EPC capture stents, as gamma irradiation may compromise the immunoaffinity of the conjugate antibodies. In addition, the use of murine monoclonal antibodies may trigger an immune reaction in patients with suitable human anti-mouse antibodies. Further clinical trial work with the Genous^™ EPC capture stents based on early designs in the HEALING trials highlighted that EPC capture stents may lead to greater late lumen loss and target vessel failure rate compared to drug eluting stents with paclitaxel [332]. Most recently, a clinical trial of 1,300 patients was prematurely suspended after evaluation of ongoing trial data showed significantly worsened target lesion failure rates after 12-months implantation compared to drug-eluting stents [333]. These less than exciting results underline a central theme of hemocompatible materials research, emphasizing the need to investigate combination strategies, especially with other biopassive surface strategies or drug-elution, to potentially achieve greater clinical outcomes [5,28].

4.2. Endothelium-inspired surface patterning

As the required implantation time of medical devices increases, complete endothelialization of the device surface is required to promote total hemocompatibility. To date, most research has focused on the incorporation of biological molecules to recruit endothelial cells, which require continuous delivery and may invoke an immune response and rejection. Therefore, researchers are now examining the interaction between blood and micro/nanostructures present in the endothelium. The endothelial lining plays a crucial role in the regulation of blood activity, effectively forming a barrier between blood and the rest of the native tissue. SEM imaging shows the aortic intima surface is not smooth but rather rough on a micro-scale, suggesting that native endothelial cells are organized on a micron level and that this surface patterning can greatly affect both endothelial and SMC behavior (Fig. 18) [334–336]. Therefore, there has been significant interest in the development of blood-contacting materials modified with endothelium-like micro/nanostructures that prevent blood from interacting with a foreign surface.

Fig. 18.

SEM image of cross-section (A) and interior view (B) of the endothelium of a carotid artery [335]. To promote rapid endothelial growth on cardiovascular stents, Liang et al. used a femtosecond laser (FSL) to fabricate a vascular smooth muscle cell-biomimetic surface texture on 316L stents (C) [339]. As a result, endothelial cell growth, alignment, proliferation, and migration were markedly improved on FSL surfaces over a 72 h period (G-I) compared to non-machined surfaces (NMS) (D-F). (A & B were reprinted/adapted from Tissue and Cell, 27(2), Pham et al, Quantitative characterization of endothelial cell morphologies depending on shear stress in different blood vessels of domestic pigs using a focused ion beam and high-resolution scanning electron microscopy (FIB-SEM), 205–212, Copyright (2015), with permission from Elsevier. C-I were reprinted/adapted from Biomaterials, Liang et al, Biomimetic cardiovascular stents for in vivo re-endothelialization, 103: 170–182, Copyright (2016), with permission from Elsevier.).

Self-endothelialization is critical for vascular stent applications, where total healing is necessary to prevent late in-stent restenosis and thrombosis. Previous studies have indicated that modifying the micro/nanostructure of materials can dictate cellular behavior, promoting adhesion, proliferation, and cell orientation of target endothelial cells [337,338]. To promote endothelialization on 316L SS stents, Liang et al. fabricated vascular smooth muscle cell (VSMC)-inspired patterned surface interfaces, resulting in selective promotion of endothelial adhesion, proliferation, and migration after 30 days of implantation in rabbits [339]. Similarly, Ti-based stent surfaces [340,341] and PLGA [342] have been modified to mimic the natural structure of the vessel wall, resulting in not only improved tissue regeneration, but also endothelium-like cell alignment and cellular function. These results suggest that micro/nanostructures influence both endothelial attachment and functionality. Micro-patterning allows for fine-tuning of the morphology and functionality of endothelial cells to produce similar behavior and alignment to that under blood flow shear stress [343]. Li et al. simulated human vascular endothelial cell morphology under blood flow shear stress through micro-patterning and showed regulation of NO, vWF, prostacyclin, thrombomodulin, tissue factor pathway inhibitor, and fibronectin secretion [344]. Therefore, surface patterning carries the potential to not only increase endothelial adhesion, but also modulate endothelial functionality, morphology, and alignment similar to what is found under in vivo conditions.

Interestingly, mimicking the surface structure of blood vessels has also been shown to result in decreased platelet adhesion. During in vivo circulation, endothelium-like PDMS consisting of submicron-ridges and nano-protuberances significantly reduced platelet adhesion and activation as a result of a flow boundary layer that caused fewer platelet-platelet collisions [345]. Bio-inspired surfaces have spurred a re-examination of the previously accepted paradigm that increasing surface roughness increases platelet adhesion due to increased surface area exposure. Recent work has found that defining a surface as rough is too general of a parameter to predict platelet behavior, arguing that surface roughness can be divided into three categories based on the size of the structures: macroscale (>2 μm), microscale (<2 μm and > 50 nm), and nanoscale (<50 nm) [334]. Increasing roughness on the macroscale range induces greater contact area for platelets, which are size-wise on that scale, while changing the surface roughness on the microscale may decrease the contact points for platelets, thereby decreasing adhesion. Surfaces containing nanoscale topography behave as smooth surfaces. When replicating the surface topography of a rabbit’s heart valve, which is composed of a cobblestone-like structure covered in tenuous villus, on PDMS, the surface became superhydrophobic and reduced platelet adhesion [346]. More information on superhydrophobic surfaces and their influence on protein adsorption and platelet adhesion can be found in the biopassive section of this review.

Although the fabrication of endothelium-like structures for medical devices has resulted in increased endothelialization and decreased platelet adhesion, several drawbacks remain. Because these surfaces are subject to flow conditions, the surface topography may degrade over time, altering the initial modifications made to the surface. Moreover, although these devices passively promote endothelialization of medical devices in the long run, the maturation of an endothelial layer coating the device takes time. Before complete endothelialization, proteins can still bind to the exposed surface and subsequently trigger the coagulation cascade and an inflammatory response. Therefore, these devices still require systemic anticoagulation during the time of implantation. Future improvements to endothelium-inspired surface patterning should consider combinations with local bioactive methods of preventing clot formation during the endothelialization process.

5. Conclusions and future directions

Medical implants routinely fail due to protein adsorption and platelet adhesion and activation, resulting in thrombus formation. Occlusive clot formation can prevent device function and result in surgical complications and local tissue necrosis. Additionally, large thrombi can embolize, leading to severe complications such as heart attack and stroke. To prevent thrombus formation, the current gold standard in clinical practice is the systemic administration of anticoagulants including heparin [347]. By preventing the activation of the coagulation cascade, systemic administration of heparin has significantly lowered the risk of thromboembolism in numerous clinical studies [348–352]. However, systemic heparinization has resulted in serious complications such as post-operative bleeding and heparin-induced thrombocytopenia. Due to these complications, the systemic administration of antithrombotic agents remains one of the leading causes of clinical drug-related deaths [353,354].

To circumvent problems that have arisen due to anticoagulant and antiplatelet therapies, bio-inspired hemocompatible surface modifications have been engineered to provide localized hemocompatibility. Nature provides a template for the most hemocompatible materials and thus has inspired researchers to design materials that mimic natural phenomena ranging from gaseous free radicals emitted from the endothelial lining to the microstructure of different plant surfaces. Each surface modification strategy (biopassive methods, bioactive methods, and promotion of endothelial cell growth) is associated with several common advantages and disadvantages (Table 2). Because of the interdependent and complex nature of biochemical reactions in the blood leading to thrombus formation, researchers have begun integrating the strategies discussed in this review to create more effective, multifunctional platforms [28]. By doing so, this next generation of bio-inspired hemocompatible surfaces can overcome some of the shortcomings exhibited by individual strategies as well as target multiple components of medical device-induced thrombosis to better optimize hemocompatibility.

Table 2.

Common advantages and disadvantages of different bio-inspired hemocompatible surface strategies.

Surface Modification Strategy	Common Advantages	Common Disadvantages
Biopassive Methods	Reduced protein adsorption Low immune response Broadly antifouling (antimicrobial)	Delayed tissue integration/healing Susceptible to degradation Time-sensitive
Bioactive Methods	Potent effects Localized activity No adverse systemic effects	Loss in activity Temporary release profile Targets single component of thrombosis
Promotion of Endothelial Cell Growth	Multiple antithrombotic characteristics Promotes wound healing Better efficacy for permanent applications	Adverse immune response Prolonged development Complex manufacturing process

Even though significant strides have been made in the fabrication of these devices, several challenges currently exist. While many surfaces have shown promising short-term hemocompatible effects, devices designed for long-term exposure still require the administration of anticoagulants or antiplatelet therapy over time. While some materials have slowed the rate of protein adsorption and platelet adhesion, devices can become compromised after prolonged exposure to blood. Furthermore, surface modifications can cause hemolytic activity toward red blood cells and invoke an undesired immune response, especially when incorporated with biomolecules. While endothelialization on device surfaces has shown promise in preventing thrombosis and restenosis for long-term or permanent exposure, the growth and development of an endothelial layer take time, and therefore these devices require anticoagulant or antiplatelet therapy during the time of implantation. Moreover, although endothelialization is the understood ultimate method of hemocompatible surfaces, practical manufacturing methods remain the primary concern. Finally, medical devices should exhibit long-term stability under storage, show mechanical durability, and should be capable of being sterilized. To fully implement these surface modifications into the market, each of these challenges should be fully considered.

Biographies

Megan Douglass is a senior engineer at a leading medical device company. Megan received her Ph.D. from the University of Georgia in 2021 under the guidance of Dr. Hitesh Handa, where she focused on the development and evaluation of hemocompatible and antimicrobial surface modifications for medical devices.

Mark Garren received his B.S. in Chemical and Biomolecular Engineering from the Georgia Institute of Technology in 2017. He is currently a PhD candidate under the guidance of Dr. Hitesh Handa at the University of Georgia. His research is focused on gasotransmitter and reactive species strategies for modulating biological responses to polymeric materials to improve infection resistance and hemocompatibility of medical devices.

Ryan Devine is currently a R&D Engineer working on tissue engineered products in the private industry. Ryan received both his B.S. and Ph. D. degrees in Biomedical Engineering from the School of Chemical, Materials, and Biomedical Engineering at the University of Georgia in 2017 and 2022, respectively. Under the guidance of Dr. Hitesh Handa, Ryan’s doctoral dissertation focused on the development and study of slippery, nitric oxide-releasing materials for the improvement of medical device hemocompatibility.

Arnab Mondal received his Ph.D. in Engineering from the University of Georgia in 2021 under the direction of Dr. Hitesh Handa. He currently works with Dr. Elizabeth Brisbois as a postdoctoral researcher at the School of Chemical, Materials and Biomedical Engineering, College of Engineering, University of Georgia. His research interest includes leveraging nitric oxide-based polymeric materials for developing medical devices for antimicrobial and hemocompatibility applications.

Hitesh Handa is an associate professor in the School of Chemical, Materials and Biomedical Engineering at the University of Georgia. Dr. Handa’s area of focus is in translational research for development of medical device coatings, wound healing materials, therapeutic nanoparticles, and microfluidic artificial lungs. With his experience in biomolecular interactions, materials/surface science, polymeric coatings, blood-surface interactions and animal models, his goal is to bridge the gap between the engineers and clinical researchers in the field of biocompatible materials.