Toward an Artificial Endothelium: Development of Blood-Compatible Surfaces for Extracorporeal Life Support

Abstract

A new generation of extracorporeal artificial organ support technologies collectively known as extracorporeal life support (ECLS) devices are being developed for diverse applications to include acute support for trauma-induced organ failure, transitional support for bridge to organ transplant and terminal support for chronic diseases. Across applications, one significant complication limits use of these life-saving devices: thrombosis, bleeding and inflammation caused by foreign-surface induced blood interactions. To address this challenge, transdisciplinary scientists and clinicians look to the vascular endothelium as inspiration for development of new biocompatible materials for ECLS. Here we describe clinically approved and new investigational biomaterial solutions for thrombosis such as immobilized heparin, nitric oxide-functionalized polymers, “slippery” non-adhesive coatings, and surface endothelialization. We describe how hemocompatible materials could abrogate the use of anticoagulant drugs during ECLS and by doing so radically change treatments in critical care. Additionally, we examine several special considerations for the design of biomaterials for ECLS, including: 1) preserving function of the artificial organ, 2) longevity of use and 3) multifaceted approaches for the diversity of device functions and applications.

1. Introduction.

Extracorporeal life support (ECLS) has evolved significantly since initial development of circulation components in the 1870–1880s and assembly of the first prototype in the 1920s by Sergei Brukchonenko [1]. In 1953 Dr. John H. Gibbon, Jr. used a new ECLS prototype to provide 26 minutes of heart and lung support during surgical repair of a cardiac defect in a human [2]. This initial device weighed over 1,000 kg and required multiple specialists to operate. Since then, ECLS technology has become significantly safer and less invasive, and has diversified to include capabilities for multi-organ failure including cardiac, pulmonary, renal and hepatic support. ECLS is now often performed outside of surgical suites in the intensive care unit and can be utilized for weeks and months of support, including bridge-to-transplant applications. Further, ECLS has been used out of hospital for cardiopulmonary resuscitation and during inter-hospital transport of unstable patients [3, 4]. The military now utilizes ECLS in combat hospitals and for transcontinental aeromedical transport of combat casualties [5, 6]. Additionally, ECLS has been introduced as a multifaceted intervention after combat trauma [7]. Future applications include destination therapies for chronic diseases using miniaturized, wearable ECLS devices that could provide support out of hospital in the patient’s home [8].

Despite the range of critical care applications and patient populations for which ECLS has been utilized, the major challenge limiting use of ECLS is disturbance of coagulation. The large surface area of interaction between blood and internal device components subjects circulating blood to detrimental effects [9]. The result is both thrombotic and hemorrhagic complications, which can coexist simultaneously.

When blood contacts foreign surfaces in the ECLS circuits, adsorption of plasma proteins rapidly occurs. Platelets exposed to the proteins become activated, adhere to the circuit surface and release procoagulant factors to recruit and activate other platelets. Platelet and contact pathway activation initiates the coagulation cascade, leading to thrombin generation and conversion of fibrinogen to insoluble fibrin - stabilizing the clot [10]. Circulating red blood cells, platelets and leukocytes continuously become activated with exposure to the developing thrombus, causing release of soluble pro-inflammatory mediators and activation of complement factors [11]. As activated blood is returned to systemic circulation, tissue factor release and pro-inflammatory factor expression occur within the endothelium [10]. Turbulent blood flow, pressure drops and shear stress are generated by the blood pump and result in hemolysis, introduction of procoagulant microbubbles, endothelial collagen exposure and von Willebrand factor release from damaged cells [11]. Cumulatively, this causes severe risk of thrombus formation in the circuit and the patient, requiring administration of systemic anticoagulants to all patients receiving therapy.

Simultaneous with thrombus formation, activation of blood and immune cells causes consumption of platelets and clotting factors, resulting in a latent hypocoaguable state and significant bleeding. Additionally, administration and maintenance of therapeutic anticoagulation in ECLS patients is exceptionally challenging due to continuous blood trauma, high transfusion requirements and underlying coagulopathies resulting from initial injury [12]. Because of this, iatrogenic bleeding (secondary to anticoagulant administration and disturbance from circulation) is one of the most common complications associated with ECLS therapy [13].

Anticoagulation for ECLS is typically performed with unfractionated heparin, direct thrombin inhibitors or combinations thereof; however, there are no universal guidelines for administration resulting in highly variable, center-specific hemostasis management protocols [12]. Lack of consensus on the type of anticoagulant to use, form of monitoring test to be employed and the therapeutic target range for each test results in suboptimal outcomes [14].

To address these challenges, research scientists and industry partners are developing biocompatible surface coatings to be applied to the ECLS circuit, providing localized anticoagulation at the blood-biomaterial interface. Inspiration for these coatings is derived from the vascular endothelium, the endogenous blood-contacting interface that covers over 3,000–6,000 m² in adult humans [15]. The vascular endothelium serves as more than a passive barrier that separates blood from foreign sources of activation. The endothelium actively regulates blood fluidity and prevents hemostasis by continuously modifying surface chemistry, excreted factors, flow distribution and permeability. Further, surface chemistry and functional characteristics of the endothelium are extremely heterogenous, varied to support tissue-specific needs and fluctuating metabolic demands [16]. The intricacy and adaptability of the endothelium emphasize the complexity of developing an artificial hemocompatible surface that would resemble its properties.

Blood-biomaterial interactions during ECLS are dependent on several factors, including: 1) composition and surface characteristics of the material; 2) rate and path of blood flow; 3) duration of blood exposure; and 4) status and composition of the blood. These factors vary depending on the type of organ support the device is designed to supply, and the specific configuration of the circuit - which varies by manufacturer. For this reason, a universally hemocompatible material currently does not exist; and further, a material that is biocompatible for one ECLS application may not be compatible for another. In this review, we discuss both clinically available and investigational hemocompatible materials for ECLS, including the biological inspiration that led to their development and the pre-clinical and clinical studies demonstrating their efficacy.

2. Biomaterials for ECLS – coatings in the clinic.

ECLS circuits consist of vascular access catheters, a blood pump, a membrane or functional unit (for the specific type of organ support) and connective tubing (Figure 1). Generally, circuits are composed of hydrophobic materials including polyvinyl chloride (PVC), silicone, polycarbonate, polymethylpentene (PMP), polyurethane and/or polypropylene [9]. Application of coatings over the surface of these polymers or incorporation of bioactive substances into the polymer matrix could alter the physical and mechanical properties of the polymer, potentially altering device performance. Additionally, the intended blood flow rate the devices are designed to support varies significantly depending on the application – for example, a rate of 3 mL/kg/min is recommended for adult respiratory support; whereas, a rate of 60mL/kg/min is recommended for adult cardiopulmonary support [17]. This span of flow rates will have very different implications for anticoagulation solutions. The size of the devices is also an important consideration as devices are modified to accommodate neonatal, pediatric and adult populations. Finally, the duration of use of the device must be considered to ensure that the coating is functional for the time of intended use. These considerations are integral in development of biomaterials for ECLS and will be discussed in context of the clinically available coatings described below (see summary Table 1).

Figure 1.

Diagram of standard extracorporeal life support system of partial lung support. Panel A schematic of extracorporeal life support system for cardiopulmonary or partial lung support highlighting the following 5 key components of the circuit: 1) membrane oxygenator/artificial lung, 2) blood pump, 3) vascular access catheter, 4) connective tubing and 5) heat exchange fibers integrated within the membrane oxygenator. Panel B schematic of a cross-section of the gas exchange fibers in the membrane oxygenator. The schematic demonstrates how by diffusion oxygen (O₂) travels from the sweep gas (100% O₂) in the hollow inner lumen of the membrane fibers into the blood that surrounds the fibers; and carbon dioxide (CO₂) moves from the blood into the membrane fiber lumen. Panel C schematic of a cross-section of the heat exchange fibers that are integrated into the membrane oxygenator. The lumen of the fibers is filled with heated water (H₂O). As blood flows through the fibers, heat is transferred to the blood to maintain the blood at body temperature (37 ˚C).

Table 1. Surface Coatings for Extracorporeal Life Support.

Summary of surface coatings for extracorporeal life support with brief description of mechanism of action (MOA) and primary element of foreign-surface induced thrombosis addressed. Antithrombin III (ATIII), poly(ethylene oxide) (PEO), Poly(2-methoxyethylacrylate) (PMEA), surface modifying additives (SMAs).

		Primary Target
Coating	MOA	Thrombin	Platelets	Protein Adsorption
Heparin	Inhibits thrombin formation via ATIII and other cofactors	X
Albumin	Passivating agent, competitive inhibition of fibrinogen adsorption			X
Phosphorylcholine	Modifies surface charge; hydrophilic characteristics form a surface hydration buffer to limit protein adsorption			X
Synthetic Polymers (PEO, PMEA, SMAs)	Alter surface hydrophilicity to inhibit protein adsorption and platelet interactions			X
Liquid Lubricant Layers	Lubricant layer prevents protein adsorption and adhesion of blood components; tether layer prevents washout of lubricant into blood			X
Nitric Oxide Releasing Materials	Release nitric oxide directly or via catalysis to inhibit platelet activation		X
Direct Thrombin Inhibitors	Inhibit thrombin without a cofactor	X
Glycosaminoglycan Polymer Brush	Physical barrier to protein adsorption			X
Factor XIIa Inhibitor	Inhibit contact pathway activation	X	X
Endothelial Cell Seeding	Mask foreign polymer surface with a layer of endothelium	X	X	X

2.1. Heparin.

Immobilized heparin, first described in 1963 by Dr. Vincent Gott, is one of the first coatings applied to ECLS and is currently the most widely utilized [18]. Heparin binds to and accelerates the activity of antithrombin III (ATIII), an enzyme that inhibits thrombin, Factor Xa and other clotting enzymes, and also performs numerous anti-inflammatory functions (Figure 2) [19].

Figure 2.

Schematic of clinically available surface coatings utilized for extracorporeal life support. Heparin Coating: Antithrombin III (ATIII) binds to heparin which induces an activating conformational change in ATIII. Activated ATIII then inhibits thrombin and other proteases. Albumin Coating: Immobilized albumin occupies surface binding sites to competitively inhibit adsorption of other pro-coagulant, platelet-activating proteins such as fibrinogen. Phosphorylcholine Coating: Phosphorylcholine produces an interface that mimics the outer leaflet of cellular membranes, masking the foreign surface and altering the surface hydrophilicity.

Within the vascular endothelium, heparin sulfate residues are anchored to the endothelium by a proteoglycan core. These residues bind and localize ATIII on the endothelial surface, and are key components of the glycocalyx – a protective barrier between the blood and endothelial cells that performs mechanotransduction and micro-regulatory functions [20].

Heparin coatings have been utilized clinically for over 4 decades, initially for cardiopulmonary bypass (CPB) and eventually for ECLS [21]. Clinical trials and meta-analyses comparing heparin-coated to uncoated circuits for CPB and ECLS showed that heparin reduced transfusion requirements [22–24], inflammatory cell activation [23, 25], complement activation [26, 27] and ICU length of stay [22–24]; however, platelet count and activation did not differ [28, 29]. Studies that distinguished between covalently bound versus ionically bound heparin found that covalent binding improved outcomes [26], likely due to better retention of the heparin molecule which can be displaced via ion exchange [9].

While the above approach has made a significant impact on medical care, most reported benefits have been observed with short-term (~6h) use [9]. The long-term efficacy for multi-day ECLS support has not been established. Currently no data supports utilization of immobilized heparin without supplemental anticoagulation for ECLS [30]. In fact, even when used in combination with systemic anticoagulants, this approach may be insufficient. For example, Lehle and colleagues examined heparin-coated oxygenators from 28 patients receiving ECLS for respiratory failure for mean duration of 11 days with continuous heparin infusion [31]. They observed significant thrombus formation and cellular deposits or “pseudomembranous” layers of 30–45 μm (primary wall thickness of gas exchange fiber is 75 μm). These layers can impede gas exchange and cause occlusion, requiring circuit exchange during treatment. Pre-clinical comparison of heparin-coated ECLS circuits used with and without continuous heparin administration in an ovine lung injury model demonstrated no difference in platelet aggregation and activation, thromboelastography or coagulation factors throughout 10 hours circulation [32]; however, after 10 hours significant thrombus deposition was observed on membranes from the group that did not receive continuous heparin infusion [33]. Out of necessity, heparin-coated circuits have been clinically utilized without systemic heparinization (or with low-dose heparin protocols) for hours to days in trauma patients with bleeding complications; however, these patients are generally transitioned to heparin infusion protocols once bleeding is resolved [34, 35].

In addition to insufficient longevity and efficacy for ECLS, heparin will not dissociate thrombin already bound to fibrin on material surfaces and will not inactivate platelets [10]. Additionally, extended exposure to heparin can invoke heparin induced thrombocytopenia (HIT). This condition is normally treated by discontinuation of heparin therapy; however, when a heparin-coated circuit is in use, this may require complete circuit withdrawal [36].

2.2. Albumin.

Introduced in the 1980s, immobilized albumin has been applied to ECLS materials and is most often used clinically when HIT is suspected. In the vasculature, albumin is incorporated in the endothelial glycocalyx via interactions with proteoglycans and glycoproteins where it serves an important role in controlling the charge and permeability of the glycocalyx [20, 37]. Albumin is the most abundant plasma protein and is thought to be inert to platelet activation and pro-thrombotic reactions when adsorbed on a biomaterial surface [38]. This is because albumin lacks specific binding sequences for platelets, leukocytes and coagulation enzymes. In this way, immobilized albumin (Figure 2) acts as a “passivating agent” or competitive inhibitor to limit adsorption of pro-coagulant proteins – primarily fibrinogen, but also vWF, fibronectin, vitronectin and immunoglobins [38]. The presence of albumin on the surfaces, however, may not be long-lasting due to displacement by pro-coagulant proteins as a result of the Vroman Effect [39]. In a 2-hour ex vivo extracorporeal circulation study comparing albumin coating to uncoated controls, albumin coating reduced plasma concentrations of activation markers of platelets, complement, coagulation and immune cells. Furthermore, oxygenators had reduced fibrinogen adsorption at end of study [40]. While albumin coated ECLS circuits have been successfully utilized in the clinic, specific data demonstrating efficacy and comparing this approach to other coatings is lacking.

2.3. Phosphorylcholine (PC).

PC is a neutral, zwitterionic phospholipid that composes the surface of the lipid bilayer of cell membranes. PC coatings have been developed to modify the surface charge and hydrophilicity of foreign polymers to resemble that of endothelial cells (Figure 2) [9]. In addition to serving as a physical barrier, phospholipids participate in cell-to-cell communication through alteration in charge and composition [41]. A 3h in vitro assessment of a PC coating applied to oxygenator segments using ovine blood did not show an appreciable reduction in platelet deposition versus heparin coating [42]. When PC coating was compared to uncoated materials during CPB, platelet count was preserved and transfusion requirements were reduced [43, 44]; however, when compared to heparin- and albumin-coated materials, PC was similar and did not show additional benefit during use or in post-operative outcomes [45, 46]. Another CPB study suggested that heparin and PC coatings may reduce inflammation compared to albumin [47]. A miniaturized extracorporeal lung support system with PC-coated surfaces was utilized for inter-hospital transport with minimal systemic anticoagulation and low incidence of coagulopathic complications [48]. Similar to albumin, while successful use of PC coatings during ECLS has been reported [49–51], there is no evidence to support use without continuous systemic anticoagulation.

2.4. Other synthetic polymer surfaces.

Synthetic polymer surfaces that are not directly modelled after biological molecules have been developed to increase hydrophilicity and biocompatibility of ECLS surfaces. This includes poly(ethylene) oxide (PEO), poly 2-methoxyethylacrylate (PMEA) and surface modifying additives (SMAs) [9]. PEO, which alone increases hydrophilicity and decreases protein adsorption [52], is incorporated in the Trillium® coating (Medtronic; Dublin, Ireland) which has a functional layer of PEO, sulfate/sulfonate groups and heparin. Applied to CPB, Trillium® was comparable to other heparin-coated materials [53–55] and preserved platelet count [53, 56], reduced platelet and granulocyte activation [55], and reduced complement activation versus uncoated controls [56]. PMEA is generated on PVC and polypropylene via plasma activation and is composed of a hydrophobic backbone along the polymer surface and a weakly hydrophilic surface on the blood-contacting face [57]. The hydrophilic face attracts a thin boundary layer of water that prevents adsorption and deformation of proteins [57]. SMA, which can be blended into the base polymer or applied to the surface, is a triblock copolymer with alternating hydrophobic and hydrophilic regions thought to compete for platelet binding sites on adsorbed proteins [57]. PMEA and SMA materials have demonstrated questionable efficacy in the CPB literature, with some authors reporting comparable results to heparin-coated circuits [58], others reporting modest preservation of platelet count [59] and/or decreased transfusion requirements [60], and still others reporting no appreciable benefit [57, 61]. Additionally, in a pediatric CPB study, investigators reported transient leukopenia, elevated respiratory quotient and C-reactive protein levels when using PMEA versus heparin membranes, which may have caused post-operative systemic inflammatory response syndrome [62]. Current use of these synthetic surfaces without supplemental anticoagulation is not advised.

3. Biomaterials for ECLS: investigational coatings.

As detailed above, clinically available coatings and polymers show benefit versus uncoated surfaces, but still require systemic anticoagulation to prevent thrombosis during extended use. Here we describe investigational surface coatings with promising features for application to ECLS (see summary Table 1).

3.1. Slippery, non-adhesive surfaces.

One novel approach to inhibit protein adsorption and thrombosis during ECLS is utilization of liquid lubricant layers with slippery, “omniphobic” characteristics. One such coating is tethered liquid perfluorocarbon (TLP), which is a bilayer coating that utilizes 1) a covalently-bound perfluorocarbon tether layer applied to the substrate/native polymer and 2) a thin, mobile liquid surface layer of perfluorocarbon lubricant [63]. The tethered layer attracts and anchors the liquid lubricant, preventing washout into overlaying fluid under flow. TLP is a modification of a coating approach called “SLIPS” or “slippery, liquid-infused, porous surfaces.” SLIPS were inspired by the Nepenthes pitcher plant which has a roughened, porous rim that attracts a liquid water layer via capillary forces to create a slippery surface preventing insect attachment [64]. Leslie, et al showed that TLP applied to acrylic and polysulfone reduces fibrinogen adsorption and platelet adhesion following incubation in whole blood [63]. Additionally, TLP was also shown to reduce thrombus formation in human whole blood during real-time clot formation using thromboelastography [65].

A key consideration in application of lubricant layers like TLP to ECLS is whether the lubricant is retained on the device surfaces under ECLS-relevant flow conditions for the duration of intended use. Howell, et al showed that similar lubricant layers were stable following 16 hours circulation at dialysis-like flow rates of 10–90 mL/hr [66]. Additionally, TLP-coated PVC tubing remained patent in a porcine arterio-venous shunt model under maximum flow rate of 300 mL/min without administration of heparin; whereas, most uncoated controls became occluded [63]. Further, the feasibility and safety of applying TLP coating to complete ECLS circuits (tubing, catheter, blood pump and membrane oxygenator) was assessed in healthy swine for 6 hrs circulation (1 L/min blood flow rate) without systemic anticoagulation [67]. TLP did not alter gas exchange efficiency of the membrane lung or cause untoward effects compared to heparin coated controls; and reduced thrombus deposition on the membrane fibers as assessed by SEM. This suggests the lubricant layer is retained at a functional level for the specific flow rates and study durations tested; however, further evaluation in a multi-day testing format is needed. In addition to TLP, liquid lubricant layers have been applied to ECLS circuit materials using other methods, such as chemical vapor deposition of a perfluorinated organosilane on coronary catheters [68] and swelling of silicone oil into the polymer matrix of silicone tubing [69]; however these materials have yet to be evaluated under ECLS-relevant flow rates.

3.2. Nitric oxide (NO) releasing materials.

Another bioinspired, active approach to prevent foreign surface-mediated thrombosis during ECLS is development of NO-releasing materials (Figure 3) that deliver NO to the blood-biomaterial interface at fluxes similar to those from healthy endothelial cells in the vasculature, estimated to be 0.5–4 ×10⁻¹⁰ mol cm⁻² min⁻¹ [70]. In the endothelium NO is a potent vasodilator, prevents platelet aggregation and adhesion, inhibits monocyte activation, and also has antimicrobial properties [71]. Various NO donor species such as diazeniumdiolated dibutylhexanediamine (DBHD/N₂O₂) [72] and S-nitroso-N-acetylpenacillamine (SNAP) [73] have exhibited antithrombogenic properties when incorporated into medical polymers and evaluated in a 4-hour extracorporeal circulation model in rabbits. In this model, coatings are applied to PVC tubing that form an arterio-venous shunt when connected to the carotid artery and jugular vein via two angiocatheters. Thrombus deposition area, platelet count and activity (platelet P-selectin expression and aggregometry) and plasma clotting time are assessed. To investigate NO-release capacity of these materials in vivo beyond the 4-hour circulation time frame, Brisbois and colleagues evaluated cannulas fabricated from SNAP-doped Elast-eon™ E2As polymer following implantation in the jugular veins of sheep for 7 days [74]. Thrombus area was reduced in the SNAP-doped group, and bacterial adhesion was reduced by 90% compared to controls. Additionally, post-circulation NO-flux from the SNAP-doped catheters was 0.6 ± 0.3 × 10⁻¹⁰ mol cm⁻² min⁻¹, still within the range of endothelial NO-flux. Further information on development, materials testing, and applications of NO releasing materials can be found in a recent series of reviews by Schoenfisch et al [75].

Figure 3.

Nitric oxide functionalized materials create surface localized pockets of nitric oxide capable of eradicating planktonic bacteria, dispersing onset biofilms, and attenuating platelet activation in ECLS applications.

A limitation to the use of NO-releasing materials is that only a finite amount of NO donors can be incorporated into the polymer matrix. Recent studies suggest that extended NO-release for over 30 days may be achievable. For example, when SNAP was covalently attached to poly(dimethylsiloxane) and coated onto silicone rubber tubing, NO flux was in a physiologically relevant range for 40 days [76]; however, during this evaluation the coated materials were simply incubated in PBS – with no exposure to blood or ECLS-relevant flow conditions. Another study extended the 4-hour rabbit model to 11 days where a SNAP-poly(ethylene glycol) (PEG) composite was applied to a silicon rubber catheter and implanted in the jugular vein [77]. At the end of 11 days, approximately 35% of the initial SNAP content was retained; again, demonstrating extended NO-release albeit under significantly lower flow rates and shear stress than during ECLS. It is also important to note that current studies investigate these materials applied to tubing and catheters only, without consideration of the ECLS membrane or blood pump.

Another solution to extend performance of NO-releasing materials is utilization of a NO catalyst to generate NO from bio-available NO-donors, such as S-nitroshothiols (RSNOs) [78]. These catalysts could continuously produce NO during ECLS (provided the NO-donors were present in circulating blood) rather than pre-loading the circuit materials with finite levels of synthetic donors (Figure 4). This can be accomplished using immobilized selenium or copper-based species to stimulate breakdown of RSNOs to NO and corresponding free thiols. For example, organoselenium species such as ebselen and other diseleno-species were shown to generate physiologically-relevant NO-flux following 5 days incubation in porcine plasma [79]. More recently, polyurethane films prepared via polymerization with diseleno chain extenders have shown reduced platelet adhesion and activation following incubation in rabbit platelet rich plasma for 1 hour; and had sustained catalytic activity over 30 days with physiologically relevant NO generation in 50 μM GSNO solutions under reducing conditions [80]. This class of materials requires further testing in vivo under flow conditions.

Figure 4.

Nitric oxide surface functionalization strategies include nitric oxide generation from endogenous donors by embedded catalysts, the spontaneous degradation of embedded donors, and combination strategies of the two.

Another subset of NO-generating materials are metal-organic frameworks (MOFs) such as the copper-based H₃[(Cu₄Cl)₃-(BTTri)₈] (H₃BTTri = 1,3,5-tris(1H-1,2,3-triazol-5-yl)benzene), or CuBTTri [78]. CuBTTri accelerated NO release from S-nitrosoglutathione (GSNO) nearly 65-fold relative to thermal decomposition [81] and CuBTTri impregnated polymers have shown significant NO release from a variety of RSNOs [78, 81]. CuBTTri polymers have yet to be evaluated in vivo under flow, but CuBTTri with exogenous GSNO has been shown to reduce platelet aggregation in whole blood in a recent ex vivo study [82]. In addition to copper-based MOFs, NO-generating copper nanoparticles have been incorporated into polyurethane and applied to PVC tubing and angiocatheters for evaluation in the 4-hour rabbit model [83]. The coating preserved platelet count and function, reduced thrombus area and fibrinogen Aү dimer formation and prevented monocyte activation, so long as a continuous infusion of an exogenous RSNO was used in combination with the coating. These effects were not observed when supplemental RSNO was not administered, suggesting endogenous RSNO concentration may be insufficient for this approach. This is a fundamental obstacle for NO-generating materials as they rely on endogenous concentrations of RSNOs or a sustained supplement, as well as variable reducing conditions to sustain physiologically relevant flux.

Whereas most of these studies tested NO-catalyzing materials at high RSNO concentrations of 50 μM, there remains poor consensus in literature on circulating concentrations of these low molecular weight RSNOs in human blood. Most sources suggest a range from ~2.8 nM to 10 μM, therefore further testing with lower RSNO concentrations may be warranted [84]. To address the potential limitation of endogenous RSNO availability, combination coatings were recently developed that incorporated NO-generating copper and selenium nanoparticles with synthetic NO donors into polymer composites producing NO-fluxes as high as 11.7 ± 3.6 × 10⁻¹⁰ mol cm⁻² min⁻¹, causing up to a 99.8% reduction in bacterial adhesion and 92% reduction in platelet adhesion in vitro compared to controls [85, 86]. Another combination coating of copper nanoparticles and GSNO applied to PVC was tested in a 4 hour rabbit ECC model, wherein 89.3% of the baseline platelet count was maintained [87].

Recent synergistic strategies (see Figure 5) have incorporated the potential of NO to prevent platelet activation with other agents that target thrombin generation or plasma protein adsorption onto the biomaterial surface. One such method is the coordination of a thrombin inhibitor with a NO donor. Major et al. developed an Elast-eon™ infused with a DBHD/N₂O₂ donor and the thrombin inhibitor argatroban immobilized to the surface to the effect of generating physiologically-relevant NO flux, maintaining platelet function, and attenuating thrombus formation in a 4 hour rabbit arterio-venous shunt model [88]. Other multi-targeted methods include modifying the hydrophilicity, hydrophobicity, or surface charge of medical polymers in combination with NO-releasing substances. For example, CarboSil-SNAP doped films top-coated with super-hydrophilic polymers such as SG80A and SP60D60 have shown up to a 4-fold reduction in fibrinogen protein adsorption with unexplored application in ECLS [89]. Further study of these and other NO-releasing hydrophilic materials for ECLS is warranted.

Figure 5.

Nitric oxide strategies work in synergy with other methods that immobilize thrombin inhibitors or modulate surface hydrophilicity/hydrophobicity for enhanced resistance to surface fouling and attenuation of platelet activation.

3.3. Immobilized direct thrombin inhibitors (DTIs).

Similar to immobilization of heparin, immobilized DTIs like bivalirudin and argatroban could prevent thrombus formation at the biomaterial surface, while reducing systemic complications. Unlike heparin, these agents do not require a cofactor and can inhibit thrombin already bound to fibrin clots [90]. Argatroban has been applied to polyurethane angiocatheters and PVC tubing and evaluated in vivo in the 4-hour rabbit arterio-venous shunt model [91], with blood flow ranging from 59–80 mL/min. The coating reduced total area of thrombus formation on the tubing but did not preserve platelet count. Thrombin clotting time in circulating blood was elevated in the argatroban group, suggesting some leaching may occur [91]. The efficacy and longevity of these materials at ECLS-relevant flow rates have yet to be assessed, and application of the coating to the membrane will require further investigation.

3.4. Glycosaminoglycan polymer brushes.

The endothelial glycocalyx consists of an extensive layer of glycosaminoglycan containing proteoglycans organized into “brush-like” structures of 100–200 nm that provide a physical barrier to protein adsorption. Attempts to recreate organized glycosaminoglycan-rich structures is challenged by the fact that these residues are strongly polyanionic, making assembly electrostatically unfavorable [92]. To overcome this, a technique was developed to form polyelectrolyte multilayers that couple polycations like chitosan with polyanions like hyaluronan [93]. This technique was used to form polymer brushes rich with sulfated glycosaminoglycans found in the glycocalyx, organized into structural domains as in the endothelium. The materials successfully inhibited protein adsorption and fibrin formation but require further evaluation in vitro under flow conditions [92]. Additionally, this approach has been coupled with nitric oxide releasing donor species to combine multiple thromboregulatory mechanisms as occurs in the endothelium [94].

3.5. Factor XIIa inhibitor.

Coagulation fXIIa is instrumental in thrombosis during blood exposure to foreign surfaces, initiating contact pathway activation [38]. To target this, application of corn trypsin inhibitor (CTI), a selective fXIIa inhibitor, was applied to polyurethane catheters using PEG conjugation to immobilize CTI on the catheter surface [95]. These catheters were evaluated in a rabbit model where time to occlusion was extended 2.5-fold compared to uncoated controls. There was no difference in prothrombin time or activated partial thromboplastin time from circulating blood in either group, suggesting that CTI does not leach from the surface. Others reported reduced fibrinogen adsorption using PEG conjugated CTI on polyurethane following incubation in plasma [96]. Further studies are needed to evaluate this approach applied to other ECLS circuit components under flow.

3.6. Endothelial cell seeding.

Rather than replicate specific thrombo-regulatory aspects of the endothelium, the concept of applying a layer of endothelial cells over the surface of medical polymers to achieve a completely hemocompatible surface has been investigated. This is accomplished by seeding cells onto the surface directly prior to blood exposure or to formulate a surface that will promote attachment and differentiation of circulating endothelial progenitor cells on the biomaterial surface to self-assemble an endothelial layer in vivo [97]. The pre-endothelialization approach, which has been utilized for vascular grafts and stents, is now being adapted for ECLS applications. For example, endothelial cells isolated from human umbilical cord blood were applied to segments of polymethyl pentene (PMP) from a hollow-fiber gas exchange membrane [98]. Because endothelial cells will not adhere to a strongly hydrophobic surface like PMP, a layer of titanium dioxide was applied to allow for cell seeding; however, the titanium dioxide coating alone reduced the oxygen transfer rate of the fibers by 22% compared to controls, prior to addition of the cell layer. Endothelial cells adhered to the titanium dioxide coated fibers forming a monolayer, and withstood shear stress at physiological levels (30 dyn/cm²). Additionally, the seeded endothelial cells remained in a non-activated state, determined by expression of pro-inflammatory/prothrombotic markers. While this is a provocative approach, feasibility of manufacturing such a product – including shelf-life, sterilization, cell source, cost and the ability to apply the cells to complete ECLS circuits must be established.

4. Conclusions.

The vast number of strategies aimed to develop a hemocompatible surface for ECLS emphasize the importance and complexity of this problem. After reviewing limitations of clinically available coatings and understanding how investigational coatings are assessed, it is apparent that there is a disconnect between how these coatings are utilized in clinic and how coatings are developed and optimized in the lab. For example, while it is important and feasible to assess novel coatings using stagnant laboratory assays in whole blood or plasma during the initial optimization phase, it is necessary to establish if a coating will perform with similar efficacy and durability under flow conditions in vivo using animal models that closely resemble the physiologic complexity and intricacies of patient care. Use of arterio-venous shunt models does not generate flow rates that would be sufficient for most forms of ECLS, and also does not account for hemolysis and shear induced by the blood pump. Additionally, the majority of studies evaluate coatings applied to tubing only, while the membrane is the most common site of clotting complications [99]. This is likely the most challenging component of the circuit to address due to the large surface area, complex structure and possibility of altering performance of the native polymer – such as gas transfer and permeability of membrane fibers designed for extracorporeal pulmonary support. The condition of the blood in investigational studies is also important as injury status, species, hydration, anesthesia medications and anticoagulants will alter the coagulation response of the blood. Finally, and potentially most importantly, many of the studies mentioned in this review were conducted for only a few hours, as many coatings for ECLS were initially developed and evaluated for cardiopulmonary bypass utilizing a 6-hour time frame. Studies of extended duration specific to certain ECLS applications are necessary to ensure that the biomaterials can perform for the duration of intended use. For example, the average duration of ECLS for adult respiratory support is 271 hours, more than 45 times the current 6-hour testing protocol [100]. Based on the testing limitations discussed in this review, we have identified 5 key areas of consideration for assessment of biomaterials for ECLS (see Figure 6).

Figure 6.

Five key areas of consideration for evaluating biocompatible surfaces and materials for extracorporeal life support.

In summary, it is important for biomedical scientists, chemists, engineers, clinicians and industry to work together to understand the challenges of clinical care and patient management, the fundamentals and properties of materials science, the intricacies of coagulation and thrombogenesis, and the challenges of developing and manufacturing a medical product. The future of ECLS will depend on these types of interactions to develop a robust solution for ECLS without anticoagulation.