Zwitterionic poly-sulfobetaine coating and antiplatelet liposomes reduce fouling in artificial lung circuits

Abstract

The artificial lung has provided life-saving support for pulmonary disease patients, and recently afforded patients with severe cases of COVID-19 better prognostic outcomes. While it addresses a critical medical need, reducing the risk of clotting inside the device remains challenging. Here, a two-step surface coating process of the lung circuit using Zwitterionic polysulfobetaine methacrylate was evaluated for its non-specific protein antifouling activity. It was hypothesized that similarly applied coatings on materials integrated (IT) or nonintegrated (NIT) into the circuit will yield similar antifouling activity. The effects of human plasma pre-conditioned with nitric oxide-loaded liposome on platelet (plt) fouling was also evaluated. Fibrinogen antifouling activities in coated fibers were similar in the IT and NIT groups. It however decreased in coated polycarbonate (PC) in the IT group. Also, plt antifouling activity in coated fibers was similar in the IT and NIT groups and was lower in coated PC and Tygon in the IT group compared to the NIT group. Coating process optimization in the IT lung circuit may help address difference in the coating appearance of outer and inner fiber bundle fibers, and the NO-liposome significantly reduced (86%) plt fouling on fibers indicating its potential use for blood anticoagulation.

Graphical Abstract:

Illustration of protein antifouling coating of ECMO circuit. Zwitterionic polysulfobetaine methacrylate (pSBMA) was grafted onto the surfaces of artificial lungs and circuits using a one pot polymerization coating approach in water (TRIS buffer) after the entire circuits had been treated with UVO-plasma. Polydopamine linkers facilitate pSBMA attachment to surfaces. The effects of coating and human blood plasma treated with nitric oxide-loaded liposome on the non-specific protein antifouling activity in devices were evaluated.

Introduction

Chronic lower respiratory diseases were the sixth leading cause of death in the US in 2019–2020 [1], and for patients with severely impaired lung function, respiratory support is required to restore life sustaining blood gas exchange. The artificial lung in extracorporeal membrane oxygenation (ECMO) has historically provided life-saving respiratory support for many lungs disease patients and has recently afforded the sickest COVID-19 patients a chance to survive [2]. The device transfers gases based on a hollow fiber membrane technology that allows mass transfer between blood and gas flows, a principle central to the function of an artificial lung. The membrane technology has also been used in various separation applications, including liquid purification, gas separation, energy applications, and biomedical applications, such as pharmaceutical filtration, and drug delivery [3,4]. Notable impacts of mass-transferring membranes are in the areas of ECMO and hemodialysis where in both applications, a concentration gradient drives mass diffusion across structurally defined membranes which remain in contact with blood. Since they are in direct contact with blood and make up the largest surface area in each of the technologies, the membrane materials’ blood compatibility is essential. Their interactions with blood clotting factors and immune cells leading to activation have required critical attention and mediation [5,6]. Blood clot fouling on the fibers and the subsequent degradation of their membrane mass transfer efficiency can be prevented using approaches that are designed to eliminate nonspecific protein fouling [7]. Material properties, including the ability to prevent the activation of the contact system of blood coagulation cascade and to avoid adverse reactions with blood have been viewed as critical hemocompatibility metrics for long-term ECMO and for lifetime use of other blood-contacting devices [6].

The clinical standard for avoiding clot formation during ECMO is systemic infusion of anticoagulant heparin. In order to limit the bleeding risk that is inherent with systemic heparinization, heparin surface coatings [8] that inactivate thrombin and hydrophilic polymer-based coatings including poly(ethylene glycol) [9,10], poly (2-methoxyethyl acrylate) [11], and phosphorylcholine [12,13] have been applied to medical devices to minimize protein and platelet adhesion onto the membrane surface [6,14].

Since surface-immobilization of heparin can reduce its anti-thrombin activity, other lung membrane modifications which include nitric oxide generating and nitric oxide releasing approaches [15–18] and zwitterionic/hydrophilic coatings [19–22] are under investigation. While the anti-thrombogenic and anti-fouling activities of these methods are effective, they are limited by flow-induced erosion, loss of activity over weeks during use, and the blood incompatibility of unreacted pre-coating materials [6]. In addition, the modification of the porous gas exchange membrane surfaces of the artificial lungs with hydrophilic coatings can lead wetting of the pores and poor gas exchange.

Besides hydrophilic coatings, superhydrophobic/omniphobic coatings that are intended to reduce wetting and therefore fouling are under investigation with reported studies showing promising protein and blood anti-fouling activities [22,24]. These coatings have been developed using physicochemical modification techniques such as chemical vapor deposition [25], nanoparticle coating [26–28], and spray coating of the lung membrane surfaces by electrospinning. [29] However, the surface inertness of membrane materials that are being coated are challenging to modify and limit utility of these methods.

Another class of antifouling surface modification methods intended for limiting protein and whole blood deposition on blood-contacting medical devices are based on utilizing fluorinated polymers [30,31]. Notable among this type of surface coating technology is tethered liquid perfluorocarbon (TLP) which impacts a slippery property to surfaces on which it is applied. TLPs are also omniphobic/amphibobic and can prevent non-specific protein fouling and wetting of the gas exchange hollow fiber membranes of artificial lungs [32–34]. While TLP have been well received as a different paradigm to biofouling control, keeping the liquid tethered onto the modified surface under flow conditions in long term applications and their optimization for superior anti-thrombogenicity over heparin coating remain unsolved and are the aims of several research groups [31–34].

The hollow fiber membrane material of the artificial lungs circuit studied in this work is a PP fiber coated with thin, solid wall of gas permeable polydimethylsiloxane (PDMS) to reduce the potential of plasma from the blood wetting and leaking through the membrane wall and the potential for gas embolus. They were also treated with UVO plasma to address the barrier of surface inertness to modification. Zwitterionic polysulfobetaine methacrylate was grafted onto the surfaces of artificial lungs circuits using a one pot polymerization coating approach in water after the entire circuits had been treated with UVO-plasma. To evaluate this two-step coating process, the multi-component and multipolymeric artificial lung circuits were coated using polydopamine to function as linkers to the surfaces of circuits and as attachment sites for subsequent polysulfobetaine methacrylate grafting. Coated devices were subsequently tested for non-specific protein antifouling activity. It was hypothesized that test outcomes from coatings present on separate materials and when present on fully manufactured device will result in an equivalent anti-fouling activity if the same coating process and test methods were followed. To test the hypothesis, polycarbonate, polydimethylsiloxane, and PDMS-coated polypropylene hollow fibers lung materials in these two groups were evaluated for protein antifouling (fibrinogen and platelets). In addition, the effects of human plasma pre-conditioned with nitric oxide loaded liposome solution at varied dosing on platelet fouling of hollow fibers were evaluated to determine the effects of the interaction of NO-loaded liposomes with human blood plasma.

Experimental Section

Materials and Reagents

The artificial lung materials including Tygon ND-100-65 tubing, polydimethylsiloxane (PDMS), PDMS coated polypropylene fibers, polycarbonate, and the artificial lungs (US 10,251,989 B2 and EP 3 129 080 B1) were provided by Advanced Respiratory Technologies LLC (ART). Dopamine-hydrochloride, sulfobetaine methacrylate (SBMA), sodium periodate (NaIO₄) and TRIS-buffered saline (pH 8.5) prepared in the lab from TRIS sachets for coating were purchased from Sigma-Aldrich. The Jelight UVO-Cleaner model 144AX series and UV light source was generated from Biochemguard BSL2 safety hood were used for plasma treatment of the surfaces of IT and NIT medical device materials and for grafting coating.

Fibrinogen assay materials include human fibrinogen powder, bovine serum albumin (BSA), Horseradish peroxidase, conjugated anti-fibrinogen antibody, O-phenylenediamine (OPD), citrate-phosphate buffer, hydrochloric acid (HCl) and phosphate buffered saline (PBS) were purchased from Sigma-Aldrich. For the LDH assay, an LDH assay kit was purchased from ABCAM. Citrated pooled male blood plasma was purchased from ZenBio, and centrifugations were done using a Beckman Coulter Allegra X-30R centrifuge.

Characterization of Plasma Treated Lung Hollow Fibers

To determine the ozone exposure duration and fiber structure relationship, PDMS coated PP hollow lung fiber mats (6 cm × 6 cm) were exposed to UV light at different time durations (20, 30, 40, 50, 60, and 90 minutes), at least N=3/condition, using the UVO-Cleaner model 144X series and compared baseline mats (no UVO exposure). Differences among treated and untreated mats were qualitatively analyzed by visual and tactile inspection. In doing so, an initial assessment of fiber mat compliance and their morphology chances in relation to UVO exposure could be understood to guide the initial UVO exposure stage of the two-step modification of the polymer surface.

Coating of Lung Circuit Materials

Integrated and non-integrated lung circuit materials were coated following the same two-step zwitterionic DOPA-SBMA grafting approach. Polymer substrates integrated and nonintegrated into lung circuits were first exposed to UV light for ozone-treatment followed by exposure to DOPA-SBMA in water (TRIS buffer pH 8.5) for the grafting of polySBMA brushes to surfaces. For UV treatment of the artificial lung device circuit or NIT materials, samples were placed inside a Jelight UVO-Cleaner model 144AX series and ran for 20 minutes. The coating solution was formulated in a large 1000 mL beaker, with the initial step of adding 600 mL of Tris-buffered saline and then dissolving 1.2g of Dopamine-HCl in 600 mL of the TRIS solution. Next, SBMA was dissolved into the solution at a DOPA:SMBA ratio of 1:15. After this, 5 mM of sodium periodate was added to the coating solution. Lung circuits were then primed using a 60 cc syringe and subsequently exposed to UV light in the biosafety hood for 2 hours while flipping the circuit intermittently. After the UV light treatment, the circuit was drained and rinsed with DI water until the effluent was clear and stored in a −20 degrees Celsius fridge for autopsy and surface analyses.

XPS Surface Characterization of Grafted Coatings on Lung Circuit Materials

The XPS spectra were collected using a monochromatic 1486.7 eV Al Kα X-ray source on PHI VersaProbe II X-ray Photoelectron Spectrometer with a 0.47 eV system resolution. The energy scale was calibrated using Cu 2p3/2 (932.67 eV) and Au 4f7/2 (84.00 eV) peaks on a clean copper plate and a clean gold foil. The lung fiber samples (coated and uncoated) were mounted onto a 60 mm diameter round stainless steel (SS) sample holder and the sample surfaces were grounded using SS metal mask covers with 5 mm diameter holes for X-ray exposure.

Evaluation of Protein Fouling on Coated Lung Circuit Materials

The fibrinogen and lactate dehydrogenase (LDH) assays were used to evaluate the protein antifouling activity of coated IT and NIT circuit materials. For the fibrinogen assay, samples were incubated in 1 mL of 3 mg/mL fibrinogen in well plates for over 90 minutes then washed three times with PBS buffer. Afterwards 1 mL of 1 mg/mL of BSA was added and incubated for 90 minutes, then washed again with PBS buffer (3x) and transferred into new wells. A 1 mL of 1:1000 dilution of HRP conjugated fibrinogen antibody (IgG 10 – 20 mg/mL) in PBS was added and incubated for 30 minutes and then washed (3x) with PBS buffer and then transferred to new wells. Additionally, 1 mL of OPD in 1 M citrate-phosphate buffer with 0.03% hydrogen peroxide at a pH of 5.0 was added and then incubated away from light for 30 minutes. For the reaction to be halted, 500 uL of HCl was added to the mixture and the fibrinogen concentration was evaluated using a UV-vis (492 nm) spectrophotometer.

For LDH assay preparation, the assay kit was thawed for 20 minutes, and while the thawing was underway, pooled adult human plasma was prepared to obtain PRP. To prepare PRP, plasma sample tubes were centrifuged at a hard spin at 400G and 3000 rpm to separate the plasma into two regions, in which the lower one-third was platelet rich plasma (PRP) and the upper two third of the tube contained the platelet poor plasma (PPP). Platelet pellets were formed the bottom of the tube and the upper two-thirds at the centrifuged blood plasma was removed for the pellet to be dispersed by gently shaking the tubes. Finally, calcium chloride (0.2 M) was added to the PRP (1:1 v/v) to reverse the effects of citrates before testing.

Coated samples (~1 cm × 1 cm) were incubated in calcium chloride spiked citrated PRP for 90 minutes at 37°C and then rinsed three times with PBS buffer and transferred into new wells. 300 μL of PBS and 10 μL of 10x lysis buffer were added and incubated for 45 minutes before adding 50 μL of the reaction mix. Samples were then incubated for 30 minutes away from light followed by adding 50 μl of HCl to stop well reactions. To detect the LDH activity of lysates from adsorbed platelets, the light absorbance by the developed well solutions were measured at 490 nm and 680 nm wavelength and the 680 nm readings were subtracted from the 490 nm measurements.

Fouling Testing of Lung’s Hollow Fiber with NO/liposome Solution Conditioned Plasma

To analyze the ability of NO encapsulated in liposomes to affect platelet fouling on the surface of lung fibers, varied Lioposome NO (LipoNO) solution doses and lung fibers were incubated, and platelet adhesion was assessed. The LipoNO solution was prepared using commercially purchased lipids and synthesized DMHD-N₂O₂ donor as previously described [35]. The liposome construct consisted of phospholipid scaffolding materials which comprised 1,2-Distearoyl-sn-glycero-3-phosphocholine (DSPC) lipids. Loading of NO into liposome was conducted via standard thin film hydration method. Briefly, 2.6 mL NO donor solution [1 mg/mL NO donor in PBS: Isopropyl alcohol (9:1 v/v)] was added to thin lipid film for NO donor encapsulation during liposome formation. 2 mL of the resulting solution was then suspended in 300 mL of DI water and recirculated in a custom tangential flow filtration system to remove free NO donor. Confirmation of NO donor encapsulation in liposome was supported by the analysis of encapsulation efficiency (EE) calculated as the total released NO from filtered LipoNO solution over that released from unfiltered NO donor solution × 100. The calculated EE was 49.87%. NO release measurements were conducted by chemiluminescence using a GE 280i NO Analyzer. 100 μL solutions were injected into 10 mL PBS (pH 5.5, 37°C) for release measurements. After passive loading of NO into liposome via thin standard film hydration method, the LipoNO solution was stored at −20°C until they were used in the fouling experiment. Fibers (~1 cm × 1 cm mat, fiber-to-fiber voids inclusive) were incubated in PRP conditioned with LipoNO solution at varied amounts described in Table 1. A Floid imager was used to inspect the surfaces of fiber for qualitative assessment while a standard LDH assay was used to quantify the adsorbed platelets.

Table 1.

Test conditions for platelet adsorption on fibers (N=5 samples/condition) incubated in PRP conditioned with NO loaded liposomes at different doses.

Groups: PP hollow fiber (uncoated or coated) incubation (1hr) w/recalcified PRP	Static Test	Expected Outcome
1. PP fiber and PRP only	LDH, PP Imaging	Highest absorption
2. PP fiber in PRP/liposome (1:1 by volume)	LDH, PP Imaging	Similar outcome as group 1
3. PP fiber in PRP/NO-liposome (1:0.5 by volume)	LDH, PP Imaging	Lower absorption
4. PP fiber in PRP/NO-liposome (1:1 by volume)	LDH, PP Imaging	Lowest absorption

Results and Discussion

Hollow Fibers’ Physical Response to UV light Exposure at Varied Durations

PDMS coated PP fibers, making up the majority surface area of the artificial lung, were qualitatively analyzed for their ozone exposure-fiber structure relationship to provide insights to the ozone sensitization limits of increasing surface radical generation. Below this limit, fiber structures should not be damaged to increase the likelihood of blood plasma leakage during the use of the lungs for respiratory support. Fibers exposed to higher durations of ozone greater than 30 minutes exhibited noticeably different gross morphological appearances such as curling and stiffening and an overall brittleness. While the longitudinal outlines of the fiber wall appeared largely intact by visual observation within these exposure times, the potential effects of the structural changes including the low-occurring pitting features on the surfaces of fibers on blood and gas flows during the gas exchange are unknown. We summarize, in Figure 1 (top row), the morphological state of fibers exposed to ozone for different durations. Also, images from different exposure timeframes showing structural changes can be found in the Supporting Figure 1 and Supporting Table 1. The observable changes in the fiber mats included stiffening and curling indicating the existence of an UV exposure duration-Fiber structural property relationship. While these observations were absent in fiber mats that were exposed to radical oxygen for 20 minutes or less, the state of fiber walls post ozone treatment and their ability to support gas exchange without increased failure rates due to wall damage or blood plasma effusion into the luminal space of the fiber wall requires investigation in the optimization of ozone treatment conditions. It can be observed in Figure 1 (bottom row) that the integrated lung circuit which undergoes a two-step surface modification for the application of anti-fouling coatings is first placed in a UV treatment chamber for ozone exposure, then primed with the coating solution, coated, drained, and then dried. Along these modification stages, any fiber structure changes are not easily observed and would require disassembly of the device for detailed characterization. Another important coating process factor was grafting coverage on the fiber bundle of the lung. Among other factors, the coverage would be influenced by the substrate’s hydrophilicity, grafting chemistry, and the substrate’s geometry. The integrated lung circuit has a uniquely complex fiber bundle geometry and may require the exploration of ways to enhance the graft of antifouling coating material candidates. In so doing, an optimized set of coating process parameters could be determined to achieve improved coating uniformity in the fiber bundle. For example, instead of repeated side-to-side flipping of the integrated lung circuit which was done in this study for even coating of the fiber bundle, a CO₂ flush before priming the lung with and establishing circuit flow of the coating solution, and optimal circuit orientation could enhance coating uniformity. See Figure 2.

Figure 1.

Lung circuit materials modifications: A) fiber mats exposed to UVO plasma for durations and morphology changes of fibers (N=5 samples/duration). Fiber images from exposure durations between 20 and 90 minutes are presented in supporting figures. The interaction of radical oxygen singlet with fibers for over 20 minutes led to stiffening and noticeable curling of fibers (A1: untreated fiber mat, A2: 20 minutes treated fiber mat, and A3: 90 minutes treated fiber mat). B) steps of zwitterionic pSBMA coating of the artificial lungs. B1 shows lungs before 20 minutes UV light exposure, B2 shows integrated lung circuit primed with coating solution after two hours of grafting under UV light source, and B3 shows coated lungs after being rinsed with DI water.

Figure 2.

An investigation of lung fiber bundle coating uniformity. Lungs primed with DOPA/SBMA coating solution 2 minutes A) and 20 minutes B) into coating of lungs circuit, and coated fiber bundle autopsied from devices after 2 hrs of coating followed by rinsing with DI water. Bundle showed qualitatively different degrees of coating stains on outer and middle mats of fiber bundles C).

pDOPA/pSBMA Polymerization and Composition of Graft Coating

To assess the polymerization DOPA/SBMA in TRIS, the coating solution consisting of SBMA, dopamine, and sodium periodate in TRIS buffer was evaluated for its absorbance dynamics using UV-Vis spectroscopy for up to 3 hours after mixing to observe any absorbance changes during polymerization of SBMA. Absorbance scans, over 200–800 nm wavelength of the coating solution, ran before UV light exposure and repeated for 3 hours when the solution was under UV light-assisted polymerization showed both increased absorbance and spectral broadening. The absorbance spectral shape of coating solutions pre and post UV light were almost identical (Figure 3).

Figure 3.

The UV-vis absorbance spectral shape (λ= 200nm to 800nm, only up to 400nm shown) of DOPA/SBMA coating solutions under 3hrs of exposure to UV light. The spectrum broadened and increased in intensity of absorbance as pDOPA/pSBMA forms.

Their absorbance intensities, however, increased with time likely due to macromolecule and polymer formation in solution as SBMA monomers polymerize and get conjugated onto dopamine through carbonyl-amine reactions. These assemblies are thought to occur during the coating of integrated lung circuit although it is not clear whether the surface grafting with the pSBMA is by a “graft-to”, “graft-from” or a combination of the two methods. To explore the coating methodology further, the surfaces of the coated lung circuit, particularly the gas exchange fibers representing the largest surface area of the device, were reviewed for chemical composition using X-ray photoelectron spectroscopy. As expected, XPS results showed differences in surface chemistries among coated and uncoated lung hollow fibers. The detected elements as a function of energy required to excite orbital electrons that were present on uncoated fibers were Carbon (C1s), Silicone (Si2s, Si2p), Oxygen (O1s, OKLL) while those on coated fibers were Carbon (C1s), Silicone (Si2s, Si2p), Oxygen (O1s, OKLL) and Sodium (NaKLL, Na1s), Nitrogen (N1s), Calcium (Ca2p3), Chlorine (Cl1s, Cl2p), and Sulfur(S2p). Surface silicone was detected in both uncoated and pSBMA-coated fibers, but expectedly in diminished amount on coated fibers due to coverage by the pSBMA coating on PDMS-coated PP fibers. See Figure 4. The presence of Sodium in the XPS spectrum is likely due to the use of sodium periodate (oxidizing agent) to assist the coating of polysulfobetaine methacrylate as it facilitates the polymerization of dopamine which anchors polysulfobetaine methacrylate to the surfaces of the artificial lungs’ materials. It is believed that the presence of calcium is likely due to the use of TRIS buffer for the coating of artificial lungs.

Figure 4.

X-ray photoelectron spectroscopy wide scans of lung fibers (coated and uncoated (N=3 samples/group). Two separate scans (scan 1 and scan 2) are presented showing uncoated controls in the left column and coated pDOPA/pSBMA coated fibers to the right. Sulfur and Nitrogen elements were present on coated fibers while detected Silicone amounts decreased in the coated group likely due coating coverage on the PDMS-coated PP lung fibers.

Biocompatibility of Artificial Lung Circuit Materials

While the scope and inferences of blood biocompatibility assessment using protein incubation assays under static test conditions is limiting, they nonetheless provide an initial biocompatibility screening tool to evaluate the potential of surface modification methods and success criteria metrics for guiding their optimization. Fibrinogen, a key blood clotting protein, and platelets which are also essential for clot formation as well as their adsorption levels on circuit-integrated and nonintegrated lung materials were used as initial screening measures for assessing the potential thromboresistance of pSBMA-coated lung circuit materials for clinical ECMO.

Adsorbed Protein

Key observations made from the adsorption of fibrinogen onto lung materials were that 1) although all the non-integrated polymers were modified using the same two-step coating process-a UV light treatment and pDOPA/pSBMA graft coating-there were variations in their protein absorption levels, and 2) fibrinogen adsorption level on each coated substrate was significantly lower than fouling on uncoated material control (Figure 5A1). The differences in fibrinogen fouling on NIT coated materials were not explored with additional tests as they mostly didn’t rise to significance. It is speculated, however, that the elemental compositional differences in the surfaces of substrates may be influencing the fouling variations, as these variations show up from fibrinogen fouling on uncoated control materials. It can be observed in Figure 5A2, that relative fibrinogen fouling on pSBMA-coated materials, as weighted by fouling on their uncoated material controls, show significant differences. The material, percent fouling, and p-values were CPC 36.47±2.05, p=0.001; Tygon 52.79±8.06, p=0.01; PDMS-coated PP fiber 45.24±7.07, p=0.004; PDMS 32.50±1.27, p=0.002; housing PC (HPC) 76.39±1.84, p=0.045, and in the IT group (Figure 5 panels B1 and B2) the values were: CPC 70.15±15.99, p=0.23; Tygon 31.15±4.48, p=0.001; PDMS-coated PP fiber 38.24±4.55, p=0.003. Note that only the fibers, Tygon, and CPC materials were able to be removed nondestructively by lung circuit autopsy for evaluation.

Figure 5.

Fibrinogen adsorption on artificial lung materials (N=5 samples/material group). Absorbance from fouling reaction media (A1) and relative fouling (A2) on materials not integrated into the artificial lungs circuit, and similar data presentation (B1, B2) for materials modified post integration into the artificial lungs circuit.

Platelet Adhesion on Graft Coatings

Platelet adsorption, as measured by lysed lactate dehydrogenase from surface bound platelets was significantly decreased due to pSBMA coating on lung circuit materials. It can be observed in Figure 6, top row panel that relative platelet fouling on coated surfaces was varied. Adhered platelets data expressed by material, percent adhesion, and p values were CPC 24.59±0.14, p=0.004; Tygon 18.87±0.71, p=0.001; PDMS-coated PP fiber 15.68±0.64, p=0.009; PDMS 38.89±0.14, p=0.037; housing PC 22.22±0.57, p=0.001 in NIT materials group. And in the IT materials group (Figure 6, bottom row panel) they were CPC 25.78±0.95, p=0.048; Tygon 45.24±1.86, p=0.002.

Figure 6.

Platelet adsorption on artificial lung materials (N=5 samples/material group). Absorbance from fouling reaction media (A1) and relative fouling (A2) on materials not integrated into the artificial lungs circuit, and similar data presentation (B1, B2) for materials modified post integration into the artificial lungs circuit.

Again, it is important to note that only the fiber, CPC and Tygon materials were nondestructively accessed from the artificial lung circuit, and thus they were used as the study group to compare the antifouling activities of materials integrated into the lung circuit or not integrated. In so doing, we gained insights into antifouling activities of materials, especially the lung fiber bundle, modified post integration into the lung circuit. For the number of lung circuits tested (N=2), it made no significant difference whether the fibers were integrated or nonintegrated in terms of fibrinogen fouling. It should be noted that the variation of pSBMA coatings’ antifouling activity across fiber bundle thickness was not analyzed. For Tygon and CPC, fibrinogen fouling in the NIT materials group was noticeably different (p=0.09 Tygon, p =0.12 CPC) than in the IT group as only their luminal surfaces of materials in the IT group were modified with pDOPA/pSBMA coating. See Figure 7A. A similar observation was made from the analysis of platelet fouling where integrated samples showed significantly elevated fouling on Tygon (p = 0.021) but not on fibers or CPC materials in the IT group compared to the NIT materials group. See figure 7B. It should be noted that as the fouling assays test for luminal and outside surfaces, the IT and NIT comparison of CPC and Tygon is not as important as fouling data from coatings on NIT materials alone. The more important comparison is about fibers which make up the largest material surface area of the ECMO device.

Figure 7.

Differences in fibrinogen (A and platelet (B fouling between artificial lungs circuit materials (N=5 samples/material group) modified with coatings before or after their integration in the lungs circuit.

Platelet Adhesion on Lung Hollow Fibers after Exposure to Blood Plasma Preconditioned with NO/Liposome Doses.

It was expected that the addition of NO-loaded liposome (LipoNO) solution to recalcified blood platelet rich plasma could lead to interactions with platelets, potentially through the unloading of NO cargo from liposomes to affect the platelet’s membrane integrin GP IIb/IIIa binding to fiber surface bound fibrinogen. Furthermore, increasing the number of NO loaded liposomes could lead to further decreases in platelet adsorption.

Adsorption of platelet, as measured by the LDH assay (Figure 8), on fibers showed marked reductions when the fibers were incubated in PRP containing NO loaded liposome doses. In panel A (uncoated fibers), the bar graph labeled 1 represents fouling outcome from the incubation of fibers in PRP only (control). Compared to bar graph 1, bars 2 (57.50 ± 11.12 %, NS), 3 (22.91 ± 6.92 %, p < 0.01), and 4 (15.53 ± 3.10 %, p < 0.001) respectively represent fiber fouling outcomes after they were incubated in PRP : empty liposome solution (1:1 v/v), PRP : NO-loaded liposome solution (1:0.5 v/v), and PRP : NO-loaded liposome solution (1:1 v/v). In panel B, middle bundle fiber material of coated artificial lungs was incubated in plasma pre-conditioned with different doses of NO loaded liposomes (LipONO) and the quantity of adhered platelets was evaluated using the LDH assay as was done for uncoated fibers. Similarly, compared to bar graph 1 (control), fouling outcomes for bars 2, 3, and 4 were (58.10 ± 17.42 %, NS), (13.20 ± 10.48 %, p < 0.01), and (1.20 ± 0.43 %, p < 0.001) respectively. The combination of coating and LiPONO led to increasingly higher reductions in measured platelet adhesion with higher LiPONO dosing than LiPONO alone. PRP : NO-loaded liposome solution (1:0.5 v/v) and PRP : NO-loaded liposome solution (1:1 v/v) compositions respectively led to additional 9.7% and 14.3% reductions in platelet fouling.

Figure 8.

Platelet fouling, as measured by LDH assay, of lung fibers after incubation in human blood plasma containing NO loaded liposome. Bars 1, 2, 3, and 4 of the bar graph respectively represent fouling outcomes from when fibers (uncoated in panel A and coated in panel B) were incubated in PRP alone, PRP : unloaded liposome (1:1 v/v), PRP : NO-loaded liposome (1:0.5 v/v), and PRP : NO-loaded liposome (1:1 v/v) solutions. Diminution of fouling on fibers independent of dilution effects and due to NO-loaded liposome can be observed by comparing bars 2 and 4 in each panel (A and B). The corresponding optical images of the surfaces of fouled fibers (uncoated) are shown in the image cluster. N=3–5 samples/group.

Similarly labeled optical images also provided qualitative insights on platelet fouling on fibers (uncoated) which showed reduced fouling at higher LipoNO dosing. Much work is needed here to understand the types of interactions that this LipoNO construct, and its derivatives can have with platelets to develop NO delivery vehicles that effectively shields the highly reactive NO from erratic and off-target reactions and deliver its NO cargo with high specificity to platelets for acquiescence. In addition, SEM images of platelets adhesion on coated and uncoated fibers that were taken are shown in Figure 9. These images were collected to assess platelet antifouling activity of coated fibers.

Figure 9.

Scanning electron microscopy of platelet fouling on artificial lungs’ fiber surface. Top row images represent fouling on uncoated gas exchange membrane surfaces while bottom row shows fouling on poly(SBMA)/polyDOPA coated versions.

Conclusion

For the first time, a two-step surface modification of the artificial lung circuit using zwitterionic pDOPA/pSBMA graft coatings for protein antifouling within the lung device was investigated to assess its potential for enhanced hemocompatibility in the lung and its potential extension to related blood-contacting applications such as dialysis membranes, vascular grafts, and ventricular assist devices. Conclusions directly supported by the study outcomes data include 1) the physical changes to the PDMS-coated PP lung fibers initiating beyond 20 minutes of UV-ozone exposure, 2) significant enhancement of the antifouling activity of coated circuit-IT materials compared to uncoated controls, 3) variation in fouling (platelet and fibrinogen) across material types, and 5) the observation that antifouling activities of NIT compared to IT fibers were not significantly different while fibrinogen or platelet fouling were higher in the IT Tygon or CPC groups. Overall, it was determined that an optimization of the coating process, especially for the IT lung circuit, is needed as the coating appearance was noticeably different between outer and inner fiber bundle fibers. It is thought that factors including flowing of the coating solution, reducing air bubble pockets within the fiber bundle and dynamic orientation of the ECMO circuit during coating, if well addressed, can enhance the coating uniformity on the fiber bundle.

The conditioning of PRP with LipoNO doses led to significant reductions in platelet adsorption on the PDMS-coated PP lung fibers. While the mechanisms of interaction between NO-loaded liposome and platelets were not studied, the antifouling effects could be explained based on known reaction of the NO molecule with platelets.

As platelet adhesion and aggregation are mediated by the binding of membrane glycoproteins to plasma proteins including von Willebrand factor (VWF) and fibrinogen [36–38], it is speculated that the LipoNO solution might be affecting those adhesion reactions. It is unlikely that plasma VWF-mediated platelet adhesion has a significant role in the antifouling outcome observed, as VWF-dependent platelet adhesion occurs optimally under high fluid shear stress conditions (> 5,000 s⁻¹) and VWF proteins, which exist as multimers in plasma, are not freely present in normal blood due to their storage in granules [38]. On the other hand, adsorbed plasma fibrinogen, confirmed in this study, can bind to the integrin αIIbβ3 (GPIIb-IIIa complex) which are localized on platelet membrane for adhesion to occur. This reaction can be affected by NO’s interaction with platelets resulting in the reduction of fibrinogen’s binding affinity for the GIIb-IIIa complex. The coating process, although requires significant optimization, shows promising antifouling results from this initial protein fouling screening, and the NO loaded liposomes’ interaction with PRP led to significant decreases in platelet adhesion on lung fibers.

Three different in vitro tests were conducted, namely fibrinogen adsorption onto the various materials that make the lung circuit, platelet adsorption on the lung circuit material using normal human plasma, and platelet adsorption using human plasma pre-conditioned with NO cargo loaded liposomes at different doses. This allowed us to perform an initial evaluation of the artificial lung materials’ ability to resist fibrinogen and platelets adhesion as these coagulation factors play integral roles during the contact, extrinsic, and common phases of the blood coagulation pathway. Optimization studies can thus be conducted to enhance these effects separately from any improvement of the artificial lung design to limit flow-induced thrombosis.