Bioinspired polymer nanoparticles omit biophysical interactions with natural lung surfactant.

Herein, we report the attenuated impact of bioinspired nanoparticles on the essential function of lung surfactant. Colloidal particles made from poly(lactide) caused a significant loss of surfactant protein B (and C) from a natural lung surfactant accompanied by a decline in surface activity under static conditions and surface area cycling. No such perturbation of lung surfactant composition and function was observed for polymer nanoparticles coated with bioinspired poly(2-methacryloyloxyethyl phosphorylcholine) (PMPC). More specifically, increasing the PMPC-coating layer thickness (≥3 nm) and density (dense conformation, distance of individual polymer chains of ≤3 nm) on the polymer nanoparticle surface diminished bioadverse events. PMPC-coated poly(lactide) nanoparticles provoked a less severe perturbation of the utilized lung surfactant when compared to colloidal counterparts coated with poly(ethylene glycol). Overall, a steric shielding of colloidal drug delivery vehicles with bioinspired PMPC can be considered as a valuable approach for the rationale development of biocompatible nanomedicines intended for lung delivery.


Introduction
Polymer nanoparticles allow for an optimized drug delivery to the lungs Ruge et al. 2016;Garbuzenko et al. 2017). Together with the inhaled approach, the controlled release properties of colloidal drug delivery systems offer significant potential for a localized therapy (lung selectivity) accompanied by less frequent administrations (lower burden of therapy) (Beck-Broichsitter, Merkel, and Kissel 2012). Meanwhile, diverse types of nanoparticles (e.g. inorganic (Hu et al. 2015;Valle, Wu, and Zuo 2015) and polymer-based (Beck-Broichsitter et al. 2011;Beck-Broichsitter et al. 2014a;Daear et al. 2015)) are suspected to adversely affect the essential lining layer of the terminal airspace (i.e. lung surfactant), which supports the gaseous exchange (Borden 2014). Nanoscale particulate matter is shown to influence the structure and composition of lung surfactant ultimately impairing its biophysical activity (Arick et al. 2015;Hidalgo, Cruz, and Perez-Gil 2017). Importantly, such an inhibition of lung surfactant functionality needs to be considered during the design of nanoscale drug delivery vehicles (Hidalgo, Cruz, and Perez-Gil 2017;Guagliardo et al. 2018), especially when treating subjects suffering from airway diseases. One promising strategy to obtain lung surfactant-"compatible" polymer nanoparticles was recently introduced by covering their surface with hydrophilic poly(ethylene glycol) (PEG) (Beck-Broichsitter et al. 2014b;Beck-Broichsitter, Bohr, and Ruge 2017;Beck-Broichsitter 2018). Meanwhile, surface coatings of polymer nanoparticles mimicking a cell-like nature (e.g. phospholipids (Sengupta et al. 2005) or cell membranes of erythrocytes (Hu et al. 2011) found applicability to minimize unwanted interactions with the biological environment and, thus, to optimize the therapeutic outcome of vascular drug delivery (Beck-Broichsitter, Nicolas, and Couvreur 2015a).
Here, we expanded on this approach and coated poly(lactide) (PLA) nanoparticles with a bioinspired phosphorylcholine-based poly(methacrylate), namely poly(2-methacryloyloxyethyl phosphorylcholine) (PMPC), mimicking the zwitterionic group of phospholipids (Monge et al. 2011;Gong and Winnik 2012), the main component of lung surfactant (Lopez-Rodriguez and Perez-Gil 2014; Parra and Perez-Gil 2015). After a thorough physicochemical characterization, the compatibility of plain, PEGylated and PMPC-coated PLA nanoparticles with a natural lung surfactant (i.e. large surfactant aggregates (LSA) from rabbit source) was investigated by monitoring the remaining surfactant protein B (and C) content and surface activity under static conditions and surface area cycling. The current observations help to increase our understanding of the interplay between nanoscale drug delivery vehicles and the essential lung surfactant system and, thus, guide a way for the development of biocompatible nanomedicines for lung administration.

Materials
All materials were obtained from Sigma-Aldrich (Steinheim, Germany) unless otherwise noted. Chemicals and solvents were of analytical grade and used without further purification.

Synthesis and characterization of bioinspired polymers
Poly(lactide)-block-poly(2-methacryloyloxyethyl phosphorylcholine) (PLA-b-PMPC) diblock copolymers were synthesized from 2-hydroxyethyl 2-bromoisobutyrate (HEBI). Ring-opening polymerization of lactide was first initiated from the free hydroxyl function of HEBI. The resulting PLA-Br macroinitiator was then used for the consecutive atom transfer radical polymerization reaction of MPC to give the respective PLA-b-PMPC (Supplemental material, Materials and methods section and Figure S1). The synthesized diblock copolymers are hereafter abbreviated as PLA-b-PMPC x , with x displaying the M n (in kDa) of the PMPC block. The polymers were analyzed by proton nuclear magnetic resonance spectroscopy and gel permeation chromatography and for their phosphorus content (Supplemental material, Materials and methods section, Table S1 and Figures S2 and S3).

Isolation of LSA
Male rabbits (2.5-3.5 kg; Charles River, Sulzbach, Germany) were utilized as LSA donors. All experiments described herein were performed in accordance with the German Law on the Use and Protection of Laboratory Animals (TierSchG) (Bundesministerium der Justiz und f€ ur Verbraucherschutz -Tierschutzgesetz \(TierSchG\)). The Federal Authorities for Animal Research of the "Regierungspr€ asidium Giessen" (Giessen, Germany) approved the study protocol (GI 20/10 A1/2012). Animals were allowed free access to water and food and were housed under controlled environmental conditions (constant temperature and humidity with a 12 h dark-light cycle). The rabbits were killed by an intravenous application of a lethal dose of pentobarbital/ketamine. A catheter was immediately placed into the trachea and lungs were lavaged three times with 50 ml of ice-cold isotonic sodium chloride solution. After filtration through sterile gauze and sedimentation of cells (300g, 15 min, 4 C), the supernatants (from 20 animals) were pooled and stored at À80 C until further processing. LSA were then isolated from the bronchoalveolar lavage fluid by high speed centrifugation (48,000g, 60 min, 4 C) (Beck-Broichsitter et al. 2014a;G€ unther et al. 1999a;Ruppert et al. 2003aRuppert et al. , 2003b.

Quantification of phospholipids
Lipids were extracted from the surfactant preparations according to Bligh and Dyer (Bligh and Dyer 1959). Phospholipids were then quantified by means of a colorimetric phosphorus assay (Rouser, Fleischer, and Yamamoto 1970). The amount of LSA and small surfactant aggregates (SSA) in the samples was determined as outlined by G€ unther et al. (1999a).

Quantification of hydrophobic surfactant proteins B and C
Samples were assayed for surfactant proteins B and C using rabbit-specific enzyme-linked immunosorbent assays (Kr€ amer et al. 1995;Schmidt et al. 2002) according to the manufacturer's protocol (MyBioSource, San Diego, CA). Briefly, samples were transferred to microtiter plates pre-coated with monoclonal surfactant protein B or C antibodies and then a polyclonal, biotin-labeled antibody was added. Next, an avidin/horseradish peroxidase conjugate was utilized for complex formation and signal amplification. 3,3 0 ,5,5 0 -tetramethylbenzidine served as the substrate, and the reading was performed at a wavelength of 450 nm.

Incubation of polymer nanoparticles with lung surfactant
LSA stock suspension and polymer nanoparticle stock suspensions were combined to meet the desired final phospholipid (PL) (i.e. 2 mg/ml) and polymer nanoparticle concentration in isotonic saline solution containing 2 mM Ca 2þ . First, mixed aliquots were incubated for 60 min (without shaking, static incubation conditions) at 37 C. Second, samples were rotated end over end (surface area cycling, dynamic incubation conditions) for predetermined time periods at 0.5 Hz and 37 C (approximately nine-fold surface area change (from $1.1 cm 2 to $9 cm 2 )) per cycle (G€ unther et al. 1999a;Ruppert et al. 2003aRuppert et al. , 2003b.

Biophysical studies
The surface activity of the samples was assessed using the oscillating bubble technique (pulsating bubble surfactometer, Electronetics Corp., Amherst, NY) (Beck-Broichsitter et al. 2011;G€ unther et al. 1999a). Briefly, measurements were performed at a constant PL concentration of 2 mg/ml in isotonic sodium chloride solution containing 2 mM Ca 2þ at 37 C. After incubation (conditions as outlined above), samples of 35 ml were transferred to the disposable sample chamber, and the adsorption rate was measured. Therefore, a bubble of minimal radius (0.4 mm) was created and while maintaining the bubble at that minimal size without pulsation, the pressure difference across the air/liquid interface was monitored. Next, pulsation was started by sinusoidally oscillating the bubble radius between 0.4 and 0.55 mm (equal to an approximately twofold surface area change per cycle). The cycling rate was set to 20 cycles/min. The pressure difference across the air/liquid interface was recorded continuously. The surface tension was then calculated using the Young-Laplace relationship. c ads (surface tension after film adsorption) and c min (surface tension at the minimum bubble radius during film oscillation) values were read after 12 and 300 s, respectively.

Statistics
All measurements were carried out at least in triplicate and values are presented as the mean ± standard deviation (SD) unless otherwise noted. To identify statistically significant differences, one-way analysis of variance with Bonferroni's post t-test analysis was performed (SigmaStat 3.5, STATCON, Witzenhausen, Germany). Probability values of p < 0.05 were considered statistically significant.
During the further physicochemical characterization, we observed a shift to more neutral f-potential values for the PLA(1)/PLA-b-PMPC 1 (1) and PLA(6)/PLA-b-PMPC 5 (1) formulations, which is an indicative for the formation of polymer nanoparticles with a core-corona structure (core: hydrophobic PLA and corona: hydrophilic, zwitterionic PMPC). X-ray photoelectron spectroscopy measurements of polymer nanoparticles, which revealed the presence of phosphorus associated with the MPC units of the polymer, supported this hypothesis ( Figure S5). The theoretical PMPC-coating layer thickness amounted to $1 and $3 nm for PLA(1)/PLA-b-PMPC 1 (1) and PLA(6)/PLA-b-PMPC 5 (1) nanoparticles, respectively (Matsuda et al. 2008;Kikuchi et al. 2012). It was found that an individual PMPC chain occupied $5-9 nm 2 on the nanoparticle surface, which is an indicative of a brush conformation in the case of a PEGylated surface (Beck-Broichsitter 2018; Rabanel, Hildgen, and Banquy 2014). Finally, the utilized colloidal formulations alone were not surface active.

Interactions of polymer nanoparticles with LSA under static incubation conditions
Lung surfactant covers the terminal airways, where it maintains a low surface tension and, thus, enables the gaseous exchange (Lopez-Rodriguez and Perez-Gil 2014; Parra and Perez-Gil 2015). LSA, the freshly secreted fraction of lung surfactant, are mainly composed of phospholipids ($80%) and further contain a minor but relevant amount of surfactant proteins (e.g. hydrophobic surfactant proteins B and C: together 1.1 ± 0.2% of phospholipids; for further details on the composition of LSA from rabbit source please consult G€ unther et al. (1999a) and Schmidt et al. (2004). A "complex" interplay between phospholipids and the hydrophobic surfactant proteins maintains a low surface tension at the air/liquid interface during breathing (Lopez-Rodriguez and Perez-Gil 2014; Parra and Perez-Gil 2015). During expiration/inspiration (change of the alveolar surface area), hydrophobic surfactant proteins significantly support the formation of complex interfacial surfactant films (Hobi et al. 2016;Zhang et al. 2011). A multilayered "reservoir" underneath increased the mechanical stability and fluidity of the lung surfactant film during the reciprocating cycles of compression and expansion (Borden 2014;Parra and Perez-Gil 2015;Bachofen et al. 2005). When assessed by the pulsating bubble surfactometer, LSA revealed c ads and c min values of 22.9 ± 2.0 and 2.5 ± 1.1 mN/m, respectively, a surface activity that is typically observed in vivo (Zuo et al. 2008).
Next, we investigated the effect of polymer nanoparticles on LSA composition and biophysical function under static conditions (no surface area cycling of the samples). Admixtures of plain PLA nanoparticles (!0.1-1.0 mg/ml) translated into a relevant loss of surfactant proteins B and C from ( Figure 1) and a dramatic decrease in surface activity (Figure 2) of the surfactant preparation in a dose-dependent manner.
A direct interaction of the plain PLA nanoparticles with essential components of lung surfactant (Hu et al. 2013(Hu et al. , 2017, such as hydrophobic surfactant proteins (Raesch et al. 2015;Kumar et al. 2016;Whitwell et al. 2016) (no relevant depletion of the PL content and/or conversion of LSA to less surface active SSA was detected) provoked the observed  Table S2. LSA were used at a PL concentration of 2 mg/ml (surfactant proteins B and C content associated with the (non-cycled) LSA: 0.5 ± 0.1 and 0.6 ± 0.1% of phospholipids, respectively) throughout the experiments. Values are presented as the mean ± SD (n ! 4). No SD bars are shown, if the SD fell into the symbol. The ruled rectangles depict the relative amount of hydrophobic surfactant proteins associated with the LSA without added polymer nanoparticles. Statistically significant differences (p < 0.05, color according to symbol code): ( Ã ) LSA incubated with magnetic polymer nanoparticles vs. LSA; ( †) magnetic PMPC-coated polymer nanoparticles vs. magnetic plain PLA nanoparticles; ( §) mPLA (6) dysfunction of the lung surfactant preparation. A decline of the "free" surfactant proteins B and C content caused a distinct architecture of the lung surfactant "machinery" at the air/liquid interface having less favorable biomechanical properties (Beck-Broichsitter et al. 2014b), with surfactant protein C showing slightly higher adsorption tendency toward the plain PLA nanoparticles (Figure 1(B)). Consequently, c ads and c min values of up to $30 and >10 mN/m, respectively, were registered for LSA in the presence of plain PLA nanoparticles. Such drastic shifts in composition and biophysical function of lung surfactants are usually only seen for severe airway diseases (Zuo et al. 2008;Griese 1999).
As of late, unwanted interactions with the biological environment were minimized when nanoscale drug delivery vehicles were modified with biomimetic (cell-like) coatings (Beck-Broichsitter, Nicolas, and Couvreur 2015a). As an example, surfaces equipped with zwitterionic polymers are believed to offer efficient stealth properties and are, therefore, considered as alternatives to commonly utilized PEGylation (Monge et al. 2011;Gong and Winnik 2012;Schlenoff 2014;Gulati, Stewart, and Steinmetz 2018). Inspired by this knowledge, we designed biodegradable nanoparticles coated by bioinspired PMPC, mimicking the zwitterionic group of phospholipids, the main component of lung surfactant (Lopez-Rodriguez and Perez-Gil 2014; Parra and Perez-Gil 2015). A thicker PMPC-coating layer was found to be more effective to circumvent interactions with the utilized lung surfactant. PLA(6)/PLA-b-PMPC 5 (1) nanoparticles (PMPC shell thickness of $3 nm) prevented a relevant depletion of the hydrophobic surfactant proteins ( Figure 1) and an impairment of surface activity (Figure 2). This observation is in general agreement with what has currently been discovered for PEGylated colloidal drug delivery vehicles. However, PEGylated nanoparticles could not completely overcome interactions with natural lung surfactants such as LSA ( Figure S8) (Beck-Broichsitter et al. 2014b;Beck-Broichsitter, Bohr, and Ruge 2017;Beck-Broichsitter 2018). In this respect, nanoscale drug delivery vehicles coated by bioinspired PMPC could be an alternative for lung administration, especially when the perturbation of the lung lining layer is triggered by a depletion of relevant components of lung surfactant (e.g. hydrophobic surfactant proteins).
Next, the impact of the distance between the individual PMPC grafts was investigated (Tables S3 and S4, Figure S9). When comparing the performance of the PLA(6)/PLA-b-PMPC 5 (1) (chain-to-chain distance of $3 nm (Table 1)) and PLA(12)/PLA-b-PMPC 5 (1) (chain-to-chain distance of $6 nm (Table  S3)) nanoparticles, it was evident that a PMPC graft distance close to or below the individual chain length featured hydrophobic surfactant proteinrepellent properties and an attenuated perturbation of lung surfactant function.
Overall, plain polymer nanoparticles revealed a significant impact on LSA composition and function under static conditions (no surface area cycling of the samples). This adverse effect was significantly attenuated when testing their bioinspired counterparts. With respect to the relationship between protein depletion and molecular weight/grafting density of the utilized PMPC, our findings indicated a general agreement with what was reported for PEGylated surfaces (Rabanel, Hildgen, and Banquy 2014;Gref et al. 2000). However, we also observed relevant differences in performance between the two polymers in the current study, which could be attributed to their distinct chemistry (PMPC is charged, PEG is neutral). The shielding properties of PMPC vs. PEG may further depend on protein characteristics (e.g. size and hydrophobicity).

Interactions of polymer nanoparticles with LSA under surface area cycling
Numerous airway diseases are associated with a disturbance of lung surfactant homoeostasis (Griese 1999;G€ unther et al. 1996;Griese, Birrer, and Demirsoy 1997;G€ unther et al. 1999b;Hohlfeld et al. 1999;Lopez-Rodriguez et al. 2017). As an example, the lavage material from patients suffering from the acute respiratory distress syndrome has an increased amount of SSA, the main metabolic product of LSA (Magoon et al. 1983;Lewis, Ikegami, and Jobe 1992). In parallel with the transition of LSA to SSA, a relevant loss of surfactant protein B in and a remarkable decline of surface activity of the remaining LSA content was observed (G€ unther et al. 1996). Said in vivo conversion of lung surfactant composition and performance can be simulated in vitro by a surface area cycling approach (G€ unther et al. 1999a;Ruppert et al. 2003aRuppert et al. , 2003bInchley et al. 1999;Veldhuizen, Yao, and Lewis 1999). When we utilized said procedure, the LSA fraction from rabbit source underwent a continuous change in composition and surface activity ( Figure S10). Specifically, the decrease in the LSA and surfactant protein B content was accompanied by a loss in adsorption rate and dynamic surface tension-lowering behavior, which accords with the main characteristics of lavage material from diseased subjects (G€ unther et al. 1999a;Griese 1999). In this respect, a further perturbation of "diseased" lung surfactants by colloidal drug delivery vehicles should be strictly ruled out.
Despite the significance, only Schleh et al. (2009) andBeck-Broichsitter (2016a) have studied the effect of nanoscale particles on the surface activity of impaired lung surfactants. These authors further emphasized, that the interactions of colloids with lung surfactant should be studied under dynamic conditions (i.e. surface area cycling). Therefore, we investigated the impact of nanoscale drug delivery vehicles after and during the conversion process of LSA to SSA by (1) adding polymer nanoparticles to cycled LSA ( Figure 3) and (2) cycling freshly harvested LSA in the presence of polymer nanoparticles ( Figure 4). An admixture of plain PLA nanoparticles to cycled LSA translated into a further perturbation of the impaired LSA performance (no relevant depletion of the PL content and/or further conversion of LSA to SSA was detected), which was again triggered by a decline of the remaining surfactant protein B (Figure 3). The addition of polymer nanoparticles coated with bioinspired PMPC (i.e. PLA(6)/PLA-b-PMPC 5 (1)) led to no detectable interference with the cycled LSA. Accordingly, for the treatment of airway disease, which encounters lung surfactant dysfunction, additive adverse effects of non-coated colloidal drug delivery vehicles (i.e. chance for "worsening" the biophysical malfunction) need to be considered.
Depending on their size (Kreyling et al. 2013) and degradability (Beck-Broichsitter, Merkel, and Kissel 2012), inhaled nanoscale particulate matter may show a distinct clearance/retention pattern from/in the lungs. Especially "persistent" nanoobjects could come into contact with newly secreted LSA and, thus, potentially influence the LSA to SSA conversion (rate) during the breathing cycle. Plain polymer nanoparticles seemed to "catalyze" the transition of freshly harvested LSA to inferior SSA during surface area cycling (no relevant decrease of the PL content was detected), accompanied by a drastic depletion of surfactant protein B and a decrease in surface activity. By contrast, PLA(6)/PLA-b-PMPC 5 (1) nanoparticles did not affect the kinetics of lung surfactant conversion (Figure 4).
The exact mechanisms leading to lung surfactant conversion are mostly undetermined. Surface area cycling and the presence of a serine-active carboxylesterase (also called "convertase") were described as responsible variables for the transformation of LSA to SSA (Ruppert et al. 2003a(Ruppert et al. , 2003bGross and Schultz 1992;Krishnasamy et al. 1997). Accordingly, a low surface tension (i.e. a highly compressed lung surfactant film with a multilayered surfactant "reservoir" underneath) is required for the "convertase" to approach its substrate (i.e. surfactant protein B, which is linked to the lung  Table S2). The remaining LSA fraction of the cycled surfactant preparation was used at a PL concentration of 2 mg/ml (surfactant protein B content associated with the cycled LSA: 0.18 ± 0.04% of phospholipids) throughout the experiments. Values are presented as the mean ± SD (n ! 4). No SD bars are shown, if the SD fell into the symbol. The ruled rectangles depict the relative amount of surfactant protein B associated with (A), the adsorption (B) and the dynamic surface tension-lowering behavior (C) of the remaining LSA fraction after surface area cycling without added polymer nanoparticles ( Figure S10). Statistically significant differences (p < 0.05, color according to symbol code): ( Ã ) cycled LSA fraction incubated with (magnetic) polymer nanoparticles vs. cycled LSA fraction ( Figure S10); ( †) (magnetic) PLA(6)/PLA-b-PMPC 5 (1) nanoparticles vs. plain (magnetic) PLA nanoparticles. For an interpretation of the color code, readers are referred to the online version of the article.
surfactant "machinery"). A less dense organization at the air/liquid interface, as occurring when adding "classical" inhibitors of lung surfactant function (e.g. plasma proteins or oleic acid), retarded the conversion rate of LSA to SSA. Particularly in the case of in vitro cycling experiments, adsorption of the hydrophobic surfactant proteins to the surface of the cycling tubes also needs to be considered. However, a previous report recovered less than 1% of the initially provided surfactant protein B content from the cycling tubes upon extraction with organic solvents (G€ unther et al. 1999a). Although plain colloidal drug delivery vehicles were identified as inhibitors of lung surfactant function (Beck-Broichsitter et al. 2014a, 2014b, no such decrease, but an elevation of the lung surfactant subtype conversion rate was observed. This must be due to their unique inhibition mechanism, namely the remarkable adsorption capacity for surfactant protein B during the surface area cycling procedure (Figure 1) (Beck-Broichsitter et al. 2014b;Raesch et al. 2015;Kumar et al. 2016;Whitwell et al. 2016). The remaining "free" surfactant protein B found in close proximity to the interfacial lung surfactant film could then be degraded by the "convertase" and/or adsorb to the surface of the cycling tubes. Polymer nanoparticles featured with a protein-repellent coating layer composed of PMPC did not  Table S2). The surface activity experiments (in (C) and (D)) were performed with the remaining LSA fraction of the cycled surfactant preparation at a PL concentration of 2 mg/ml throughout. Values are presented as the mean ± SD (n ! 3). No SD bars are shown, if the SD fell into the symbol. Statistically significant differences (p < 0.05, color according to symbol code): ( Ã ) cycled LSA incubated with (magnetic) polymer nanoparticles vs. cycled LSA ( Figure S10); ( †) (magnetic) PLA(6)/PLA-b-PMPC 5 (1) nanoparticles vs. plain (magnetic) PLA nanoparticles. For an interpretation of the color code, readers are referred to the online version of the article.
interfere with the lung surfactant machinery (i.e. no relevant adsorption of surfactant protein B) and, thus, the conversion rate of LSA to SSA remained unaffected.

Conclusions
To summarize, we investigated the in vitro effect of polymer nanoparticles on the essential biophysical performance of a natural lung surfactant. Plain polymer nanoparticles revealed a relevant perturbation of lung surfactant function, due to a depletion of essential surfactant protein B (and C), in a static setting and under dynamic conditions of surface area cycling. Nanoparticle concentrations, which were shown to be effective in vitro, are on the upper end of concentrations, which will be reached in vivo after inhalation delivery. However, accumulation of polymer nanoparticles might be observed when frequent dosing is necessary (e.g. chronic diseases), which would worsen the surfactant/nanoparticle balance. A coating of the polymer nanoparticles with a bioinspired, protein-repellent phosphorylcholine-based poly(methacrylate) layer (i.e. PMPC), rather than PEGylation avoided a perturbation of lung surfactant composition and function. PMPCcoated PLA nanoparticles can be considered compatible with the lung surfactant "machinery" present at the air/liquid interface and, thus, represent a valuable platform for the therapy of lung diseases.