Pulmonary surfactant and drug delivery: Vehiculization, release and targeting of surfactant/Tacrolimus formulations

Abstract

This work explores the potential for strategizing pulmonary surfactant (PS) for drug delivery over the respiratory air-liquid interface: the interfacial delivery. The efficacy of PS- and interface-assisted drug vehiculization was determined both in vitro and in vivo using a native purified porcine PS combined with the hydrophobic anti-inflammatory drug Tacrolimus (TAC), a calcineurin inhibitor. In vitro assays were conducted in a novel double surface balance setup designed to emulate compression-expansion dynamics applied to interfacially connected drug donor and recipient compartments. In this setup, PS transported TAC efficiently over air-liquid interfaces, with compression/expansion breathing-like dynamics enhancing rapid interface-assisted diffusion and drug release. The efficacy of PS-assisted TAC vehiculization was also evaluated in vivo in a mouse model of lipopolysaccharide (LPS)-induced acute lung injury (ALI). In anesthetized mice, TAC combined with PS was intra-nasally (i.n) instilled prior administering i.n. LPS. PS/TAC pre-treatment caused greater TAC internalization into a higher number of lung cells obtained from bronchoalveolar lavages (BAL) than TAC pre-treatment alone. Additionally, the PS/TAC combination but not TAC or PS alone attenuated the LPS-induced pro-inflammatory effects reducing cells and proteins in BAL fluid. These findings indicated that PS-mediated increase in TAC uptake blunted the pro-injurious effects of LPS, suggesting a synergistic anti-inflammatory effect of PS/drug formulations. These in vitro and in vivo results establish the potential utility of PS to open novel effective delivery strategies for inhaled drugs.

Graphical abstract

1. INTRODUCTION

Current drug delivery strategies are mainly designed based on depositing the therapeutic principle directly on the place of action (topic) or upon absorption and distribution (systemic) into the body and subsequent targeting. Pulmonary surfactant-driven drug vehiculization goes further and opens a new perspective in the current respiratory drug delivery paradigm.

Pulmonary surfactant (PS) is a membrane-based lipid-protein material synthesized and secreted to alveolar spaces by type-II pneumocytes. It reduces the surface tension of the layer of water coating the respiratory surface, preventing pulmonary collapse during breathing, while at the same time it establishes a first barrier against potentially harmful entities at the distal airways[1]. To fulfill these vital functions, PS has the capability to efficiently adsorb into the alveolar air-liquid interface and spread long distances over it. Considering its composition and interfacial properties[2], PS may simultaneously (1) solve the solubility problem of hydrophobic drugs, which can be intercalated between surfactant phospholipids, (2) enhance drug distribution over the respiratory surface once deposited at the upper or conductive airways, and (3) assist in reaching places in the lungs that are barely accessible with the current delivery strategies, such as in the distal alveolar spaces[3].

Antibiotics[4–7], antimicrobial peptides[8], recombinant adenoviral vectors[9], antioxidant enzymes[10], corticosteroids[11–14], vaccines[15] or nanoparticles[16] have already shown both in vitro and in vivo synergistic effects when administered in combination with PS. Nevertheless, to the best of our knowledge, few studies have addressed the molecular mechanisms behind the capability of PS to transport drugs through air-liquid interfaces and none to know the insights involved in the release process. Here we have developed a novel dynamic setup based on previous static models developed in our laboratory[13] and elsewhere[17], incorporating breathing-like interfacial dynamics into interfacially connected compartments. Following the squeeze-out model, which explains the interfacial behavior of PS subjected to compression-expansion cycling[18], we hypothesized that surface dynamics may play a key role during drug release and therefore contribute to enhancing the vehiculization process. Specifically, we wanted to test whether drug exclusion may happen through interface-associated 3D structures formed during breathing dynamics[19], as occurring with unsaturated phospholipids[18] or meconium components[20]. To model the entire barriers developed in the lungs in situ (i.e. branched structure, mucus, immune system or the air and liquid flows), we also analyzed as a proof of concept the distribution, targets and potential synergistic therapeutic effects of PS/TAC formulations in a lipopolysaccharide (LPS)-induced mouse model of acute lung injury (ALI). LPS is widely accepted as model to simulate ALI since it activates innate immune responses via Toll-like receptors (TLR), especially TLR-4 [21, 22] and stimulates the release of inflammatory mediators, including TNF-α, IL-1β and IL-6 [23]. This triggers activation of inflammation and the resulting damage to lung tissues and PS [24, 25].

As a model of hydrophobic drug, we used Tacrolimus, also known as FK506 or TAC. This drug is an immunomodulatory compound that has been widely used in transplantation of human lungs [26, 27] and other solid organs [28, 29] to prevent and treat allograft rejection. Apart from the effects on T-cells and adaptive immune responses, TAC also influences innate immune response processes [30–33]. Conjugated with FK binding proteins, TAC inhibits calcineurin, a phosphatase involved in intracellular signal transduction and blocks mitogen-activated protein kinases (MAPKs) cascade and NF-κB. This inhibits pro-inflammatory cytokine production (e.g. IL-1β and IL-6) and NOS, and triggers the expression of anti-inflammatory cytokines (e.g. IL-10, TGF-β and TNF-β) [34–36]. TAC also induces apoptosis in inflammatory cells via caspases 3 and 9, shifting to an anti-inflammatory environment [36]. Furthermore, some studies suggest that TAC firstly induces the activation of steady-state macrophages blocking the constitutive inhibition of TLRs by calcineurin [30, 37]. Therefore, TAC may constitute an effective treatment to modulate the activity of macrophages and neutrophils and to equilibrate the balance of pro-/anti-inflammatory cytokines reducing the inflammation process.

With this in mind, the present work addresses 1) the ability of PS to transport TAC via air-liquid interfaces, 2) whether breathing-like surface dynamics enhances drug diffusion and release, 3) the feasibility of surfactant-assisted interfacial vehiculization in vivo, and 4) the potential synergistic effects of PS/drug combinations using TAC as an example. Altogether, we have determined some of the main factors implicated in inhaled surfactant/drug vehiculization along with the synergistic effects and the potential targets that may be associated with surfactant/drug formulations.

2. MATERIALS AND METHODS

Unless otherwise stated, all chemicals were purchased from commercial suppliers (i.e. Sigma-Aldrich®, Merck KGaA or Macron Fine Chemicals™) and used without further purification. Water was from a Merck-Millipore Direct-Q3 purification system and further distilled for its usage in the surface balance experiments.

2.1. The vehicle: pulmonary surfactant

Whole native PS (NS) was purified from bronchoalveolar lavage (BAL) fluid obtained from porcine lungs, following a protocol optimized at our laboratories (see Suppl. Information) [38–40]. Then, organic extract (OE) was obtained by subsequent lavages with Chloroform/Methanol (2:1 v/v) following the Bligh and Dyer method [41]. To prepare aqueous suspensions from the OE, an appropriate volume of the extracted material (solubilized in chloroform/methanol [2:1 v/v]) was dried under a nitrogen stream to form a dry film and then under vacuum for two additional hours to remove organic solvent traces. The dried films were subsequently hydrated with a buffered solution (i.e. 5 mM Tris, 150 mM NaCI, pH 7.4 for in vitro assays and PBS for in vivo assays) at 45 °C, a temperature above the melting temperature of PS phospholipids. The hydration was performed during one hour with vigorous shaking every 10 min.

2.2. The passenger: Tacrolimus

Tacrolimus (TAC or FK506, +99%) was purchased from Sinoway Industrial (Xiamen, China) in powder form. To synthesize the fluorescent derivative of TAC bearing a Nile Blue analogue (abbreviated TAC·NBA, Figure S1, Suppl. Information; λ_abs^max = 569 nm; λ_em^max = 655 nm in 100 mM pH 7.4 PBS buffer), the C22 carboxymethoxyloxime of TAC [42] was reacted with the aminopropyl-NBA following a similar procedure to that described by Glahn-Martinez et al. [43] for mycophenolic acid labeling (see supplementary information for details).

2.3. The road: surface troughs

The classical Langmuir-Blodgett and Wilhelmy troughs have been combined in this work creating a novel set-up to study surfactant-promoted drug vehiculization under static and dynamic conditions. The Langmuir-Blodgett trough was purchased from NIMA technologies. The Wilhelmy trough was designed in collaboration with the UCM Research Support Centre (CAI). Surface pressure sensors were purchased from NIMA technologies (Coventry, UK). Wilhelmy plates (21 mm perimeter; 24 mm x 10 mm x 0.5 mm) were made with a grade 41 Ashless Filter Paper (Whatman®, GE Healthcare Life Sciences). The interfacial bridge was made of a hydrated filter paper (Whatman No. 1). Under limiting conditions of surfactant concentration (<10 mg/mL), differences in spreading rate were observed depending on the size of the interfacial bridge. However, at the higher concentration used in the in vitro experiments (50 mg/mL), those differences are negligible. It is still possible that the cellulose matrix of the paper bridge could interact and perhaps retard, or impair, the interface-associated lateral diffusion of some PS structures, but this may mimic, at least conceptually, potentially analogous effects of the mucus layer at the upper and conductive airways. Figure 1 illustrates the static and dynamic devices in detail to better understand the experiments and the information that can be obtained from each configuration used in this work.

Figure 1: Dynamic vehiculization trough and experimental design.

Image (A) and schematic representation (B) of the double through setup. (C) Experimental sequence of experiments and main questions and outcomes addressed to understand surfactant and interface-assisted drug vehiculization and release.

2.3.1. Static vehiculization trough

To evaluate the potential of surfactant to transport TAC through an air-liquid interface, a static vehiculization trough setup was used. As described previously, it consists of two Wilhelmy surface troughs, one acting as a donor (315 mm²) and the other as a recipient (106 cm²), connected by an interfacial bridge (made of hydrated filter paper; Whatman No. 1; 1 cm x 5 cm) [13, 17, 44].

2.3.2. Dynamic vehiculization trough

In order to study the effect of interfacial compression/expansion dynamics on surfactant-driven drug vehiculization, a new dynamic setup was designed based on the static vehiculization trough. A Wilhelmy through (10 cm²) and a Langmuir-Blodgett trough (area variable from 185 cm² to 58 cm²) were connected with an interfacial bridge (made of hydrated filter paper; Whatman No.1; 1 cm x 7 cm) to develop the dynamic vehiculization setup (see Figure 1). This novel device is highly versatile and brings the possibility to further study the interfacial behavior of PS and its potential as a drug delivery system.

A small volume of any surfactant sample, combined or not with a drug, can be deposited upon injection into the aqueous subphase, or by spreading or aerosolization directly onto the interface, of the donor trough. This donor compartment would be somehow mimicking the first area of deposition at the upper airways where a therapeutic formulation would be delivered. Once adsorbed into the donor interface, surfactant spreads through the interfacial bridge (analog of conductive airways) and reaches the Langmuir trough (representing the large and dynamic surface of distal airways). Only the fraction absorbed into the interface of the donor trough, and then spread through the interfacial bridge, reaches the Langmuir-Blodgett trough and can be further studied. This method overcomes potentially erroneous interpretations derived from the formation of subjacent structures in the subphase during the traditional way of application (by spreading of an organic solution or an aqueous suspension directly onto the interface), which may act as reservoirs of new material.

As shown in Figure 1C, different scenarios can be modeled by changing the configuration of the device. In a first configuration, the bridge is removed just before performing compression-expansion cycles at the Langmuir trough leaving both interfaces disconnected (2a). In this case, the surfactant film that has already spread to the recipient compartment is isolated and the structure of the film formed, associated to the eventual release of TAC over the cycles, can be analyzed avoiding the adsorption of new material coming from surfactant reservoirs in either the subphase and/or the surface of the donor trough. Alternatively, the compression-expansion cycles can be performed in the presence of the bridge with both interfaces connected (2b). This allows analyzing how interfacial dynamics enhances the diffusion of surfactant and drug along the interface. Additionally, surface pressure can be further monitored after the dynamic cycles either in the absence or in the presence of the bridge (3a and 3b, respectively). Removing the bridge after completing the cycles and monitoring the change in surface pressure for a certain period of time brings the possibility to study the potential readsorption of interfacially-associated membranes and surfactant structures that were previously excluded from the interface during cycling. Leaving the bridge allows evaluating whether, after successive interfacial compression-expansion cycles, the compression-driven excluded material can be replaced by new material coming from the reservoirs in the donor compartment.

2.4. Captive Bubble Surfactometer

In order to determine the potential effects of TAC on surfactant performance, a Captive Bubble Surfactometer (CBS) was used. CBS allows for modelling the alveolar dynamics under physiological-like conditions of temperature, humidity and pH, and compression-expansion cycling rates. An aqueous suspension of 200 nL of the appropriate surfactant/TAC combination, at a phospholipid concentration of 25 mg/mL, was injected close to the interface of a small air bubble (5 mm diameter, 0.05 cm³). The bubble was formed at the bottom of an agarose cap confined inside a cylindrical sealed thermostatted chamber (T = 37 °C). The bubble is then subjected to subsequent compression and expansion cycles by moving up and down a piston driven by a computer-controlled engine. The experiment is continuously recorded by a video camera for further analysis. Based on the height and diameter of the bubble, the volume, surface area and surface tension can be calculated by a specific software [45, 46].

2.5. Intercalation and release of TAC into and from interfacial phospholipid films

To study whether TAC intercalates and interacts within surfactant phospholipids at the air-liquid interface, interfacial monolayers of DPPC (the most abundant phospholipid in PS) have been assessed in the Langmuir-Blodgett trough in combination with three different proportions of TAC (1%, 5% and 10% w/w with respect to phospholipids). DPPC films exhibit a very conspicuous structure upon compression, which is very sensitive to perturbation as a consequence of the presence of minor amounts of any compound. Thus, they are particularly useful to show the potential interactions and effects of TAC within interfacial lipid films. Each DPPC/TAC mixture, dissolved in Chloroform/Methanol (2:1 v/v), was spread at the air-liquid interface to form an interfacial film. After 10 minutes waiting for organic solvents to evaporate, the compression isotherms (barrier speed: 65 cm²/min) were obtained. In order to study the effect of the drug on the interfacial lateral structure of the films by epifluorescence microscopy or AFM, the interfacial films were transferred onto glass or mica plates, respectively, before and after completion of five compression/expansion cycles (barrier speed: 65 cm²/min). For epifluorescence preparations, the COVASP method was followed (barrier speed: 25 cm²/min; dipper speed: 5 mm/min) [47, 48] to transfer the film under the complete compression isotherm on the same plate. For AFM studies, a fixed pressure was selected instead and was maintained constant during transfer to the mica support (dipper speed: 5 mm/min). The samples for epifluorescence observations were conveniently labelled as described in the Fluorescence microscopy section. The experiments were carried out in a thermostatic chamber at 25 °C and dark conditions.

2.6. Static surfactant-driven interfacial vehiculization

An aqueous aliquot (15 μL at 50 mg/mL) of reconstituted OE combined with 10% TAC (9% TAC + 1% TAC·NBA by mass) was deposited at the surface of the donor trough. After 30 minutes for the PS/TAC combinations to spread over the interfacial bridge, reaching and covering the recipient interface, the material at the interface and tightly associates areas were carefully collected with a custom-built trap connected to a pump, to measure TAC·NBA fluorescence. This process was performed twice, one right after 30 minutes of donor-to-recipient spreading and a second one after 30 extra minutes to accumulate more fluorescent molecules and enhance the detection by fluorescent spectroscopy. Changes in surface pressure were continuously monitored during the experiments in both donor and recipient troughs (T = 25 °C).

2.7. Dynamic interfacial vehiculization and release

An aqueous aliquot (15 μL at 50 mg/mL) of the reconstituted OE of PS in combination with 10% TAC (8% TAC + 2% TAC·NBA by mass) was deposited into the donor trough. After 30 min for the PS/TAC combinations to spread through the interfacial bridge and reach and cover the recipient interface, different experiments were performed (see Figure 1). The films at the recipient Langmuir-Blodgett trough were first subjected to 10 compression/expansion cycles (barrier speed: 65 cm²/min) in the presence or in the absence of the interfacial bridge to evaluate the effect of cycling dynamics on the interfacial trip and release process. Additionally, the surface pressure was monitored during further 30 minutes once the cycles were completed, in the presence or in the absence of the bridge, to evaluate whether the material excluded during cycling is able to re-adsorb again (no bridge) or is replaced with new coming material diffusing from the donor interface (with bridge).

The interfacial spreading and vehiculization were also evaluated in interfaces previously covered by PS emulating what would be the endogenous surfactant coating the airways. The interface was firstly coated with limited amounts of NS (1 μL labelled with 1% mol BODIPY-PC at 50 mg/mL) to minimize the formation of surface-associated reservoirs in both donor and recipient troughs. After 35 min to allow for NS coating of both donor and recipient interfaces, a second addition of surfactant was applied at the donor compartment (14 μL at 50 mg/mL) either under static or dynamic conditions. In this case, the sample was a combination of OE and 10% TAC (8% TAC + 2% TAC·NBA by mass), emulating how an additional dosage of vehiculizing surfactant could work.

After each experiment, the interfacial films were transferred onto glass or mica plates for structural observations following the COVASP method [47, 48]. The interface was also carefully collected for measuring the amount of TAC·NBA fluorescence appearing in the recipient trough. Changes in surface pressure were continuously monitored during the experiments in both donor and recipient troughs (T = 25 °C).

2.8. Fluorescence spectroscopy

In order to investigate the vehiculizing potential of PS, fluorescence of the TAC·NBA probe was measured in both donor and recipient troughs using an Aminco Bowman Series 2 spectrofluorometer (λ_excitation = 590 nm; λ_emission = 650 nm). The emission spectra were recorded at 25 °C and sensitivity was set differently for the samples from the donor and recipient troughs as considerably more fluorescent molecules were collected from the donor compartment. Every sample was stored at 4 °C under dark conditions until being measured to avoid photobleaching of the sample.

2.9. Epifluorescence microscopy

To analyze the structure of DPPC monolayers in the presence of TAC, images were taken from interfacial films doped with 1% (molar ratio) NBD-PC (phosphatidylcoline labelled with the fluorescent probe nitrobenzoxadiazole; Molecular Probes, Thermo Fisher Scientific), transferred as Langmuir-Blodget films onto glass coverslips under continuous compression conditions (barrier speed: 25 cm²/min; dipper speed: 5 mm/min). Images were acquired with an ORCA R2 10600 (Hamamatsu Photonics K.K.) camera coupled to an epifluorescence microscope (Leica DM 4000B, Leica Microsystems). In the pertinent experiments, the red fluorescent derivative of TAC was added conveniently as a trace (1% by mass with respect to lipids). When the amount of drug was higher than 1% by mass, the remaining was completed with non-modified TAC.

For the experiments performed in vivo, 1.5% by mass with respect to lipids of TAC·NBA was used to track the drug over the lungs and to determine which type of cells isolated from bronchoalveolar lavages take up TAC. The images were taken under an Olympus IX-81 inverted fluorescence microscope.

2.10. Atomic force microscopy

To analyze the lateral organization of phospholipid films and the three-dimensional structures of PS under compression, samples were prepared following the same method used for epifluorescence microscopy but transferring the films onto mica plates. The measurements of DPPC monolayers were taken by using an AFM MultiMode SPM (Digital Instruments, Inc./Veeco Metrology Group). In the case of the experiments performed with surfactant films, a NanoWizard II AFM (JPK Instruments, Berlin) was used. Samples were imaged in contact mode, using MLCT and MSNL-10 silicon nitride tips B, C and D, with a spring constant of 0.01, 0.03 and 0.1 N/m respectively. The setpoint was kept constant to 0.1 N/m and the scanning rate adjusted between 0.5 and 2 Hz. The data were processed using the JPK Data Processing software.

2.11. In vivo model for Acute Lung Injury

To test the efficacy of surfactant-assisted drug delivery in vivo, we used LPS (5 mg/Kg, Escherichia coli O111:B4, Sigma-Aldrich)-induced model of ALI. We determined drug uptake in alveolar cells and associated therapeutic benefits. The fluorescent version TAC·NBA was used in a proportion of 1.5% wt with respect to surfactant phospholipids. Experiments were carried out in the following 6 groups of anesthetized, Swiss Webster mice (8-10 wk, males, 30-40 g) in which we i.n. instilled the indicated solutions, sequentially at 0 and 1 h:

1. PBS (30 μL) at both time points.

2. PBS (30 μL), then LPS (5 mg/Kg).

3. OE in PBS (10 mg/mL, 50 μL), then LPS (5 mg/Kg) to test surfactant effects.

4. TAC·NBA in PBS (50 μL; 7.5 μg dissolved in 1 μL dimethyl sulfoxide [DMSO] mixed with 49 μL PBS), then LPS (5 mg/Kg).

5. TAC·NBA/OE in PBS (10 mg/mL, 50 μL), then LPS (5 mg/Kg).

6. TAC·NBA/OE in PBS (10 mg/mL, 50 μL), then PBS (5 mg/Kg).

Five hours after the last instillation, the mice were anesthetized and exsanguinated through the heart for euthanasia prior to BAL collection. To anesthetize mice, 100% oxygen and 3% isoflurane was supplied for induction and intra-peritoneal injection of ketamine (100 mg/kg) and xylazine (5 mg/kg) for anesthesia maintenance. The Institutional Animal Care and Use Committee of Columbia University Medical Center approved all animal procedures.

2.12. BAL collection and analyses

In mice euthanized by exsanguination under anesthesia, 1 mL ice-cold PBS was instilled into the trachea and, after a gentle massage over the chest, the BAL fluid was withdrawn. Then, the BAL was analyzed for A) cell counting, B) determining protein content, C) observing the presence of red blood cells (RBC), D) evaluating TAC uptake by BAL cells and E) identifying cells morphologically. For counting cells, 9 μL of a BAL aliquot was added to Turk’s solution (1:10 v/v) containing gentian violet, a nuclear stain and 1-2% acetic acid, which destroys red blood cells (RBCs). The Turk’s solution enabled accurate cell counting in a haemocytometer. Cells were counted in four different squares (0.04 mm² each; four aliquots from each BAL). Subsequently, the remaining BAL was centrifuged (15 min; 4 °C; 800 G) to pellet cells for microscopy and the supernatant BAL saved for protein analysis.

BAL supernatant was analyzed by the bicinchoninic acid (BCA) assay for measuring the protein content. The pellet was directly observed by naked eye for the presence of RBC and then resuspended in PBS (1 mL). Subsequently, the cell suspensions were centrifuged in a Cytospin™ 4 Cytocentrifuge (5 min; 4 °C; 2000 rpm) after diluting conveniently to have proper cell density for further analysis. Finally, the cytospun cells on slides were fixed with methanol (99%) and part of them additionally stained with Giemsa (30 min) for fluorescence and bright-field microscopy, respectively. Epifluorescence microscopy (Olympus 1X81) was done to test for TAC labeling on BAL cells as an indication of drug uptake. Bright-field microscopy of Giemsa solution-stained cells revealed cell morphology, which enabled identification of cells based on cell size and nuclear morphology.

The samples stained with Giemsa could not be observed to evaluate the TAC uptake, as this colorant is also visible with the same filter used for TAC·NBA, complicating the correct interpretation of the images by inducing false positives (see Figure S2). Therefore, the samples were prepared separately to distinguish the cell populations (Giemsa staining) and observe the TAC·NBA fluorescence (no stain, just the fluorescently labelled TAC).

2.13. Statistics

Data were expressed as mean (standard deviation). Multiple results comparisons were conducted using One or Two Way ANOVA test following by post-hoc analysis with Tukey test when appropriate. Two-pair comparisons were carried out by paired or unpaired t-test when appropriate. Analyses were performed with Sigma Plot 11 (Systat Software, San Jose, USA) and p<0.05 was considered to be significant.

3. RESULTS

3.1. Surfactant-assisted drug vehiculization in vitro

3.1.1. TAC intercalates into surfactant lipids preserving the interfacial performance

As shown in Figure 2A, regardless of the presence of up to 10% TAC into the complexes, the aqueous suspensions of OE maintained very rapid interfacial adsorption properties reducing surface tension of water to an equilibrium tension of 23-25 mN/m in 5 seconds (p=0.807). The slow quasi-static compression/expansion cycles revealed a tendency that the higher the amount of TAC into the surfactant, the more area reduction is needed to reach a minimal surface tension below 5 mN/m in the first cycles. This is only achieved after large exclusion plateaus that are larger the higher the amount of drug. This suggests that TAC takes some room at the interface upon adsorption, intercalated between surfactant lipids, and that compression of the drug/surfactant film is progressively excluding the drug out from the interface, likely along the plateau. Remarkably, the compression needed to reach the minimal surface tensions is progressively decreasing over the cycles. Subsequently, when rapid breathing-like dynamic cycles were performed (bottom panels in Figure 2A), a similar interfacial behavior was observed in the presence and absence of TAC, with practically no hysteresis and a minimal surface tension of <2 mN/m reached with less than 20% area reduction. This suggests that probably the drug was excluded and released during the compression/expansion cycles. This occurs at a level that the ultimate interfacial performance able to produce minimal tensions with limited area reduction is not compromised.

Figure 2: Effect of TAC on the interfacial performance of surfactant films.

(A) Assessment in the CBS. Interfacial adsorption kinetics of aqueous suspensions of OE in the absence and in the presence of TAC (10% by mass) is represented as surface tension-time (γ-time) isotherms (upper panels). Mean of 3 experiments is represented. Error bars represent the standard deviation. Representative γ-area isotherms of three independent experiments are compared in the absence and in the presence of increasing amounts of TAC (1%, 5% and 10% by mass) during slow quasi-static (middle panels) and breathing-like rapid dynamic cycles (lower panels). The decreasing scale of grey represents each cycle; 1^st to 4^th for quasi-static and 1^st, 10^th and 20^th for dynamic. (B, C) Incorporation and effects of increasing amounts of TAC into DPPC monolayers. The graphs represent the compression isotherms of DPPC at 25 °C in the presence of increasing amounts of TAC (1%, 5% and 10% by mass with respect to DPPC) before (A) and after (B) five compression/expansion cycles. Surface pressure (Π) = γ_H2O – γ_sample. (D) Epifluorescence images showing how compression-expansion cycles promote the interfacial exclusion of TAC (bar scale = 50 μm). Representative images taken before and after five compression-expansion cycles at the Π indicated on each image. Traces (1% by mass) of NBD·PC and TAC·NBA were added to detect interfacial distribution of DPPC and TAC. Representative colours; green: NBD-PC; red: TAC·NBA. The histogram of each image was adjusted to better differentiate the structures in the image (fluorescence of each dye after cycles was reduced because of interfacial exclusion. Not shown).

To confirm the behavior of TAC upon compression of drug/phospholipid interfacial films, increasing amounts of TAC (1%, 5% or 10% w/w) were incorporated into DPPC monolayers formed upon spreading of organic solutions onto a Langmuir-Blodgett surface balance. Figure 2B represents the surface pressure-area (π-area) isotherms of films formed by DPPC (the main phospholipid in lung surfactant) in the presence of increasing amounts of TAC. The area per molecule at the lowest Π rises as the amount of TAC increases, indicating that TAC certainly intercalates and occupies some space between DPPC molecules, at the air-liquid interface. At Π≈10-12 mN/m, DPPC begins to undergo a lateral reorganization from liquid expanded (L_e) to liquid condensed (L_c) phases, as observed in the so-called phase transition plateau in the isotherm and in the formation of highly packed domains (black areas in Figure 2D, excluding the fluorescent bulky probe in films transferred onto glass slides and observed under a epifluorescence microscope). The presence of increasing proportions of TAC produces a shifting on the L_e-to-L_c plateau and hinders the formation of micro (Figure 2C and S3) and nano-domains (seen in the AFM images in Figure S3) of highly packed DPPC. The size of those domains is reduced and, at the highest doses of TAC, DPPC only starts forming domains at substantially higher Π than pure DPPC (see Figure S3). Nevertheless, at around 30 mN/m all the isotherms converge, suggesting that TAC could be expelled out from the interface at a level that does not affect the interfacial behavior of DPPC. Interestingly, Figure 2C and D show that after five compression/expansion cycles, all isotherms resemble pure DPPC and domains start forming again at a lower Π than occurring before the cycles, even in the presence of the highest amount of TAC, suggesting that interfacial dynamics is actually promoting the exclusion of TAC out from the interface.

3.1.2. Static surfactant-driven interfacial vehiculization

Previous experiments revealed that TAC can intercalate between surfactant lipids at the air-liquid interface and that interfacial dynamics may promote the progressive exclusion of TAC from the interface. The next step was to further study whether PS membranes can use the interface to transport incorporated TAC to long distances more efficiently than current drug delivery strategies (e.g. liposomes). To do so, a static vehiculization trough was used. Figure 3A and 3B show the π-time isotherms obtained from the donor and recipient troughs, respectively. Immediately after adding the OE samples in the presence or in the absence of TAC (black arrow in Figure 3A), surface pressure increased until reaching the equilibrium surface pressure (π_eq ≈ 42 mN/m) in the donor compartment, showing no substantial effect of TAC in surfactant interfacial adsorption. In the case of liposomes, only the sample with TAC produced a small increase in surface pressure, indicating that TAC may somehow slightly enhance the adsorption of phospholipids into the air-liquid interface. The π-time isotherms in the recipient trough (Figure 3B) show that only the samples containing OE increase the surface pressure near the equilibrium. This suggests that PS, alone or in the presence of TAC, is actually travelling efficiently along the interfacial bridge.

Figure 3: Interface-assisted diffusion and vehiculization of TAC in static conditions.

(A-D) π-time isotherms comparing the presence of material adsorbed into the air-liquid interface from PS or liposomes (DPPC/POPG 7:3 w/w) in the donor and recipient troughs without (A, C) and with TAC (B, D). An aqueous aliquot (15 μL at 50 mg/mL) of OE or liposomes in combination with 10% TAC (9% TAC + 1% TAC·NBA by mass) was deposited in the donor surface. The surface pressure was continuously monitored in both donor (A) and recipient (B) air-liquid interfaces. Black arrow represents the addition of different formulations on the donor surface. Mean of 3 independent experiments is represented. Bars represent the standard deviation. (G, F) Relative fluorescence emission of TAC·NBA (λ_em=655 nm) measured on both donor (G) and recipient (F) compartments. After each experiment, the donor and recipient interfaces were collected. Mean after 3 independent experiments are represented, with bars indicating the standard deviation. Relative fluorescence was calculated independently with respect to the maximum value on each graph considering all replicates. Paired t-test: (**) p=0.02.

After 30 min, the donor and the recipient interface were carefully collected to measure the presence of TAC·NBA fluorescence (Figures 3C, D). The fluorescent spectra show that, only in the presence of surfactant, TAC·NBA was detected in the recipient trough, confirming that PS can transport TAC over the interface while liposomes are not suitable for a rapid diffusion between the interface-connected compartments.

3.1.3. Dynamic interfacial vehiculization and release

Once confirmed that PS can transport TAC over air-liquid interfaces under static conditions, the next step was to study the possible relevance of interfacial compression-expansion dynamics, using a Lagmuir-Blodget trough as the recipient compartment (dynamic vehiculization trough).

Figures 4A and C illustrates the π-time isotherms of an illustrative whole experiment, including a first adsorption period followed by the changes in pressure occurring during the compression-expansion cycles. Similar to the observations with the static vehiculization trough, when surfactant is deposited into the donor interface (black arrow), the surface pressure rises drastically until a Π_eq of around 42 mN/m. After 35 minutes, the surface pressure in the recipient raised above 30 mN/m, indicating that both donor and recipient interfaces are occupied by surfactant. At this time, 10 compression-expansion cycles were performed in the presence (Figure 4A and B) or in the absence of the interfacial bridge (grey arrow in Figures 4C and D). In the presence of the bridge, the surface pressure at the donor trough fluctuates as a consequence of the compression/expansion dynamics performed at the recipient, evidencing that the connection between both interfaces is transmitting the waves in lateral pressure. During the cycles, a progressive reduction of the exclusion plateau of PS at around 40-45 mN/m is observed in both situations (see Figures 4B and D). This plateau coincides with the equilibrium Π and occurs when Π remains constant during compression or expansion as a consequence of folding and unfolding of interfacial films, respectively. According to the squeeze-out model, the interfacial films fold down during compression and those interfacial components that cannot support such steric forces squeeze-out, counteracting the Π increase [49, 50]. During expansion, the folds that are closely attached to the interface spread out over the interface again, compensating the Π reduction. After this plateau, compressibility of the interfacial film also decreases over the cycles (visually detectable observing the increasing slope). Regardless of the connection between both interfaces, the major reduction of this plateau is observed during the first compression, while in the subsequent cycles it tends to be more gradual. Similar tendency can be detected observing the minimum surface pressures at the end of the expansion steps (see Figure 4E): a sharp reduction during the first cycle is followed by a more gradual reduction of minimal pressure in the following cycles. These observations suggest that interfacial material is being lost during interfacial dynamics. Nevertheless, when the cycles are performed in the presence of the bridge, the minimal surface pressures remain stable from the 6^th cycle, implying that new surfactant coming from the donor trough could be replacing the material lost in the recipient interface during the cycles.

Figure 4: Interfacial dynamics promotes the interfacial trip and release.

(A, B) π-time and π-area isotherms upon compression-expansion cycling of the recipient interface with both compartments connected. (C, D) π-time and π-area isotherms with cycling performed after removing the interfacial bridge connecting donor and recipient compartments (disconnected). (E) Minimal surface pressure after each expansion. Error bars correspond to standard deviation (n=3). Two Way ANOVA (connected/disconnected: p<0.001, cycles: p<0.001, Tukey post-hoc test: (*) p<0.05). (F) Relative fluorescence emission of TAC·NBA (λ_em=655 nm) measured at the surface of the recipient trough in the presence or in the absence of the interfacial bridge. Paired t-test: (**) p=0.012. (G) Relative fluorescence emission of TAC·NBA (λ_em=655 nm) at the surface of the recipient trough before and after cycling or just after the first compression (in different experiments) without the connecting bridge. Relative fluorescence was calculated independently with respect to the maximum value on each graph considering all replicates. Error bars correspond to standard deviation calculated from the measurement of relative fluorescence (n=3). One Way ANOVA (p=0.001), Tukey post-hoc test: (***) p=0.002.

Fluorescence measurements of TAC·NBA in the recipient trough revealed similar results (see Figures 4F and G). The fluorescence observed after performing compression-expansion cycles with both interfaces connected is higher than in the absence of the bridge, indicating the presence of higher amounts of TAC at the interface of the recipient compartment (Figure 4F). Figure 4G shows that fluorescence decays dramatically after the cycles, in accordance to the situation observed in the π-area isotherms. A similar decrease can be observed after the first compression, confirming that the decay in surface pressure is associated with a major exclusion of the drug. To confirm the exclusion of TAC, additional measurements of fluorescence in the subphase with both compartments connected shows that TAC fluorescence is significantly higher after the cycles (see Figure S4). These results suggest that, although TAC is interfacially excluded to the subphase during interfacial dynamics, the interfacial connection allows for a continuous transference of material from the donor to the recipient trough.

3.1.4. Tacrolimus is released from interface-associated 3D structures

Interfacial films derived from the previous experiments were transferred onto mica and glass plates to be further analyzed under atomic force and epifluorescence microscopy as shown in Figure 5. Attending to the “squeeze-out” model [18, 19], three-dimensional structures associated to the interface were expected to appear under compression conditions. Figures 5A and B show the topographical analysis of the films obtained by AFM at different surface pressures before and after performing ten compression/expansion cycles, respectively. Note that the thickness of phospholipid monolayer is around 1.5-2 nm and that the difference between lipid phases is around 0.5-0.8 nm [51, 52]. Figure 5A1 illustrates that at higher Π before the plateau (≈35 mN/m) observed in the π-area isotherm (left panel), the air-liquid interface is saturated with PS. Here, micro- (1-2 μm in length) and nano-domains (70-120 nm in length) of thicker liquid-condensed lipid phases (L_c) can be observed surrounded by thinner liquid-expanded lipid phases (L_e). Some higher structures of around 4 nm height are formed as 3D projections (white arrows in Figure 5A1). At higher pressures in the plateau (see Figures 5A2), these structures increase in height (up to 20 nm) and extension (up to 5 μm). Note that, due to the high lateral pressures, lipid phase coexistence is no longer observable. These results suggest that the interface-associated 3D structures are monolayers and bilayers stacked on top of each other, reaching in some cases up to five bilayers.

Figure 5: Three-dimensional structures associated to surfactant films in the presence of TAC.

(A, B) Topographical images taken under AFM before (A) and after (B) interfacial dynamics. A 3D representation is included on top of some of the images. The graphs on the left panels represent the π-area isotherm with the numbers indicating the pressure at which the images were taken. Graphs below each image represent the height profile of the indicated segment (white dashed lines). White arrows: nanometric projections (4 nm height). (C) Main structures observed in the films transferred from the recipient compartment. Images taken under an epifluorescence microscope from transferences of OE·NBD-PC + TAC·NBA films, before or after applying dynamic cycles with or without the connecting interfacial bridge. White arrows indicate some examples of excluded three-dimensional structures. The histogram was shortened to better differentiate the structures in the images. Red: TAC·NBA; Green: NBD-PC. Bar scale = 50 μm. (D) 3D structures seen under AFM corresponds to the brighter domains under epifluorescence microscopy. The histogram of each fluorescence image was firstly shortened and then contrasted to better differentiate the fluorescent structures. (Red: TAC·NBA; Green: NBD-PC). Bar scale = 25 μm.

Figure 5B1 shows that at low pressures after the cycles (8 mN/m), phase separation occurs. Moreover, nanometric projections of up to 100 nm diameter and 4 nm in height can also be observed (white arrows in Figure 5B1). At higher pressures below the plateau (≈30 mN/m; Figure 5B2), the structures showing several stacked monolayers and bilayers are observed again. However, they appear at lower pressures than observed before applying the interfacial dynamics. These results likely suggest that once the 3D structures are formed during the first compression, some could remain over the expansion periods acting as nucleation sites for extension of 3D projections along the next successive cycles.

The epifluorescence microscopy images (see Figure 5C) revealed highly intense fluorescent spots in all assessed experimental conditions. They correspond to high concentration of fluorescent dyes NBD-PC (green) and TAC·NBA (red) that rearrange forming clusters during compression. After the cycles, the spots are smaller and the areas around completely dark, suggesting complete segregation and exclusion of the fluorescent molecules out from the film. The intensity of fluorescence also decreases over compression and after the cycles. This evidences a possible lateral reorganization of the film during interfacial dynamics, which leads to accumulation of fluorescent dyes in specific locations before being excluded out from the interface. Surprisingly, comparing these images with those obtained by AFM of the same films (see Figure 5D), the fluorescent clusters coincide with the compression-driven three-dimensional structures, suggesting that TAC is firstly accumulated in those specific locations and then excluded from the interface through them.

3.1.5. Readsorption of released material

Once revealed that interfacial dynamics promotes drug release from the interface, we determined whether excluded material could re-adsorb back into the interface or be replenished by new material coming from the reservoir of material at the donor compartment. To do so, after performing the compression-expansion cycles in the presence of the bridge, the surface pressure was further monitored during 30 min with both interfaces either connected or disconnected. Figure 6 represents the π-time isotherms during the whole experiment. When the interfacial bridge is removed (gray arrow in Figure 6A), only a slight increase of up to 10 mN/m is observed in the recipient trough once the cycles were finished, suggesting that the re-adsorption of material already excluded from the interface upon dynamic cycles is limited. Conversely, when the interfaces are connected (Figure 6B), surface pressure rises following similar kinetics to that observed before the cycles but with no initial lapse time. It finally equilibrates with the donor pressure in around 30 min (41-42 mN/m), suggesting the continuous supply of material coming from the donor reservoir.

Figure 6: Material released during compression does not readsorb into the interface.

π-time isotherms were recorded upon compression-expansion cycling and then continued by (A) removal of the interfacial bridge (disconnected), or (B) leaving the bridge after the cycles (connected). Black arrows represent the moment at which surfactant was applied. Grey arrow indicates when the bridge was removed. A representative replica of 3 experiments is shown.

3.1.6. Vehiculization over surfactant-occupied interfaces

All the previous experiments were performed in clean initially opened air-liquid interfaces. In this section, the interfacial vehiculization was performed in interfaces that were previously occupied by surfactant. As illustrated in Figure 7A and B, the first addition of surfactant (grey arrows) produced the typical jump in surface pressure up to 41-42 mN/m. However, after the first 5 minutes it plunged to 10-12 mN/m. This sharp reduction coincides with the beginning of the increase in the recipient compartment before reaching an equilibrium at 10-12 mN/m. This suggests that surfactant leaving the donor surface is not replenished because of the limited amounts of material applied into the donor. At the moment of adding the vehiculizing surfactant under static conditions (black arrows), the surface pressure raised again up to 41-42 mN/m in the donor trough and surfactant started traveling interfacially until equilibrating at the same pressure in the recipient compartment. Under dynamic conditions, the vehiculizing surfactant was added during the fifth expansion, but when surface pressure in the donor trough was still increasing (35 mN/m) to ensure a proper surfactant film covering the interface. This addition of new material results in a notable increase in surface pressure at the donor trough, and a reduction in the previously-observed sharp fluctuations (Figure 7). At the same time, the minimum and maximum pressures of each cycle started to increase at the recipient trough. In parallel, images of each experiment were taken under the epifluorescence microscope. Images in Figure 7C illustrate that the exogenous-like surfactant containing TAC (red) reaches the recipient trough and mixes with the endogenous-like (green) at surface pressures below 45 mN/m. At higher pressures, green fluorescence is not observed while red is clustered in brighter spots. This presumably results from the lateral exclusion of the drug from highly packed domains and the subsequent interfacial exclusion as it happened in the experiments performed with clean interfaces. These results evidence that surfactant merges with the one already occupying the interface and is still able to travel and transport TAC interfacially. In other preliminary experiments (not shown) with higher initial interfacial loads of surfactant we have observed that interfacial dynamics, as promoted by compression-expansion cycling, may be crucial to promote the surface tension gradients that drive surface-assisted spreading of interface-associated materials.

Figure 7: Interfacial vehiculization of TAC along surfactant-coated interfaces.

Surface pressure has been monitored into donor and recipient troughs after covering the interface with NS labelled with BODIPY-PC and later adding OE + TAC·NBA. The second surfactant addition was: (A) after 35 min in static conditions (black arrow), and (B) during the fifth expansion (black arrow) leaving the experiment to continue for five more compression-expansion cycles. (C) Main interfacial structures formed during the diffusion of OE/TAC·NBA mixtures on a NS/BODIPY-PC coated interfaces. Epifluorescence images from the experiments adding surfactant/TAC suspensions in static conditions (static) and during the 5^th expansion (dynamic). The histogram was shortened to better differentiate the structures in the images. Merge is shown (Red: TAC·NBA; Green: NBD-PC). Bar scale = 50 μm.

3.2. Surfing the interface in vivo: a proof of concept

3.2.1. Synergistic effects of surfactant/TAC formulations in preventing LPS-induced inflammation in mice

Figure 8A shows that the numbers of WBC present in bronchoalveolar lavage increased significantly as a consequence of LPS-induced inflammation (LPS; 1.71±0.29 million/mL; mean ± SD, n=4) in comparison with the control group (PBS; 0.21±0.04 million/mL; mean ± SD, n=4). A pretreatment with only TAC (TAC + LPS) or aqueous suspension of OE (OE + LPS) did not prevent the LPS-mediated WBC increase (2.20±0.52 and 2.29±0.21 million/mL, respectively; mean ± SD, n=4). Nevertheless, when mice were pre-treated with TAC vehiculized with OE (OE/TAC + LPS), the numbers of WBC were greatly reduced (0.64±0.12 million/mL; mean ± SD, n=4), revealing that OE/TAC combinations have synergistic effects in preventing the pro-inflammatory effects of endotoxin.

Figure 8: Effects of intra-nasal surfactant-assisted TAC pre-treatment on LPS-induced pro-inflammatory lung alveolar responses in vivo.

Mice were pre-treated as shown and then received LPS an hour later. 5 hours later BAL was collected. PBS; phosphate buffered saline, OE; organic extract, TAC; tacrolimus, LPS; lipo-polysaccharide. (A) White blood cells count in BAL. Mean of four mice is represented for each condition. Four aliquots per mouse BAL were counted in four different squares of 0.04 mm² of a Neubauer chamber. Error bars represent the standard deviation. One Way ANOVA (p<0.001), Tukey post-hoc test: (*) p<0.001. (B) Total protein content in BAL. Mean of four mice is represented for each condition. Error bars represent the standard deviation. One Way ANOVA (p=0.03), Tukey post-hoc test: non-significant, (§) p=0.061.

As shown in Figure 8B, the protein content in the lavage of LPS and TAC + LPS groups (0.70±0.38 and 0.86±0.55 mg/mL, respectively; mean ± SD, n=4) is higher than found in the control (0.22±0.04 mg/mL; mean ± SD, n=4). This evidences that LPS-induced increase in alveolar epithelial permeability cannot be prevented by the delivered doses of TAC alone. In the case of mice pre-treated with surfactant alone (OE + LPS), the total protein content is reduced (0.37±0.16 mg/mL; mean ± SD, n=4). Furthermore, pre-treatment with TAC vehiculized with OE (OE/TAC + LPS) reduced protein content in BAL even more (0.20±0.04 mg/mL; mean ± SD, n=4), suggesting that the combination was beneficial although the effect did not reach significance compared to the TAC + LPS group (p=0.063; One-Way ANOVA with Tukey).

3.2.2. TAC uptake as a proof of surfactant-mediated interfacial distribution

Before evaluating TAC content in BAL cells, we characterized cells obtained from BALs. As expected, Figure S5 shows that the pool of bronchoalveolar cells obtained under control conditions in the PBS and OE/TAC without LPS groups is mainly composed of macrophages 97±2% (mean ± SD; n=3) and 88±9% (mean ± SD; n=3), respectively (see supplementary Fig. S5). However, under LPS-induced inflammatory conditions, the proportion of macrophages decreases (LPS: 9±4%; TAC + LPS: 5±3%; OE + LPS: 13±9%; OE/TAC + LPS: 9±6%; mean ± SD; n=3), as a consequence of an influx of other cells with multi-lobular nuclei, most likely neutrophils based on their morphology. The non-nucleated cells correspond to erythrocytes.

In order to elucidate whether TAC is internalized by BAL cells and if PS enhances this process, the WBC obtained from bronchoalveolar lavages were observed under the fluorescence microscope. Figure 9A shows representative images of the groups of mice pre-treated with TAC in the absence (left) and in the presence (right) of surfactant. The images show that when the drug is combined with surfactant, TAC is internalized into more cells. However, not every type of cell takes up TAC. According to their morphology, the cells that contain TAC appear to be macrophages. The percentage of total cells that take up TAC was also calculated (Figure 9B). The percentage of cells that contain TAC when it is vehiculized by OE (7.9±5.1%; mean ± SD; n=3) is higher than observed when TAC is delivered alone (1.4±1.1%; mean ± SD; n=4) and coincides with the percentage of macrophages (see Figure S5B). In the absence of LPS, the percentage of TAC-containing cells increased (60.0±12.7%; mean ± SD; n=4). This increase may be because the majority of cells in this group are macrophages, suggesting again that macrophages are the main BAL cells taking up TAC. Moreover, TAC fluorescence intensity inside the cells was also measured (Figure 9C). Interestingly, the groups treated with OE/TAC, exposed or not to LPS, showed similar fluorescence intensity, significantly higher than observed in cells of the surfactant-free group. These results suggest that macrophages internalize TAC, and that PS facilitates the process, either by improving the distribution along the respiratory surface or by leveraging the process of surfactant clearance.

Figure 9: TAC content of BAL cells in mice receiving a surfactant-drug combination in vivo before PBS or LPS treatment.

LPS or PBS were given to mice an hour after pre-treatment with TAC or OE/TAC. Cells in BAL were deposited on slides with cytospin and fixed with MetOH. PBS, LPS and OE+LPS groups are not shown as no fluorescence was detected. (A) Merged brightfield and fluorescence images (equal gain processing) show low (top; scale bar = 50 μm) and high magnification (lower; scale bar = 20 μm) views of TAC fluorescence. (B) Percentage of cells that internalise TAC. Bars represent mean of six different images per mouse (n=4 for TAC+LPS; n=3 for OE/TAC+LPS; n=4 for OE/TAC). Error bars represent the standard deviation. Two Way ANOVA (presence of LPS p<0.001; presence of surfactant p=0.327), Tukey post-hoc test: (**) p<0.001. Zoomed bar chart compares the inflamed groups with LPS. Unpair t-test: (§) p=0.051. (C) Fluorescence intensity of cellular TAC·NBA. Mean and standard deviation of cell fluorescence intensity obtained by counting the following average number of fluorescent cells per replicate (n=3 for TAC+LPS; n=2 for OE/TAC+LPS; n=3 for OE/TAC): 24±23 cells for TAC+LPS group, 40±19 cells for OE/TAC + LPS and 84±19 cells for OE/TAC. Fluorescence intensity was measured by ImageJ (Fiji). Two Way ANOVA (presence of LPS p=0.970; presence of surfactant p=0.005), Tukey post-hoc test: (*) p=0.004).

4. DISCUSSION

The feasibility of using PS as a drug carrier has been investigated since the 1990s. Nevertheless, in most cases, the use of PS lipids or lipid/protein mixtures in the context of drug delivery has been proposed as a mere extension of the use of liposomes as carriers for poorly water-soluble drugs. Mechanisms underlying surfactant-assisted drug delivery at the lung’s air-liquid interface are undefined. In this context, the role of breathing dynamics in drug distribution and release, as well as the potential targets, remain unexplored. In biophysical and murine studies, we determined processes by which lipid components of PS enhance hydrophobic drug solubility and delivery to alveolar cells. We showed that compared to drug or PS alone, PS-assisted drug pre-treatment inhibited pro-inflammatory responses in murine models of LPS-induced ALI during spontaneous breathing. These studies indicated that combining PS with the drug was critical in the presence of breathing dynamics. We showed that drug uptake occurred in alveolar cells, and that alveolar permeability and inflammatory cell accumulation were reduced. These findings provide direct evidence of therapeutic drug uptake and protective function in LPS-induced models of ALI. Breathing dynamics seems to enhance this process and also to promote drug release once at the alveolar spaces, targeting alveoli and phagocytic cells.

The development of PS/drug formulations must be carefully analyzed on a case-by-case basis to ensure that the incorporation of drugs into surfactant complexes preserves their interfacial performance and integrity. For instance, compared to the current gold standards in drug delivery, e.g. liposomes and nanoparticles [53]. PS is not a particularly good container for hydrophilic drugs because of the high dynamism of surfactant membranes. Nevertheless, it has been described as an effective system to solubilize hydrophobic drugs such as corticosteroids or antibiotics [13, 17]. Here, we report the feasibility of incorporating the hydrophobic immunosuppressive drug TAC into PS membranes. The experiments performed in CBS showed that the incorporation of TAC, even at the maximum tested proportion of 10% by mass, does not remarkably affect surfactant adsorption properties nor its interfacial behavior under breathing-like dynamic conditions. Thus, TAC is proposed as a good candidate to be vehiculized by PS while preserving PS surface active properties.

As reported in previous studies from our lab [13] and elsewhere [17]. PS resulted an efficient vehicle for surfing the respiratory interface and delivery of drugs such as budesonide, beclomethasone or ciprofloxacin. In contrast, liposomes, one of the most used systems for respiratory delivery of poorly soluble drugs due to versatility, low cost and biocompatibility [3], fail in travelling efficiently over air-liquid interfaces. Distribution of liposomal drugs into the airways could then occur by the extremely slow diffusion of liposomes along the thin but very extended aqueous hypophase coating the whole conductive and respiratory epithelia. However, here we acknowledge a limitation of the study since these experiments were performed in clean interfaces. The presence of pre-existing PS films already occupying the interface might perhaps promote a partial adsorption and fusion with TAC-containing liposomes and enhance a partial intercalation of TAC between surfactant lipids, preceding the spreading of some amount of the drug, something that should be explored in the future. Still, the combined drug/PS suspension likely offers maximal interface-assisted capabilities as a consequence of the interfacial driving action of surfactant proteins. As assessed in the vehiculization trough, we confirmed in this work that PS can efficiently transport TAC long distances over static air-liquid interfaces either clean or already occupied by PS (Figures 3 and 7). However, the lungs are continuously inflating and deflating during the process of breathing. The highly dynamic milieu in the lungs is translated into subsequent repetitive compression and expansion of the respiratory air-liquid interface at the alveoli. We hypothesized that, in a similar way to the interfacial exclusion of unsaturated phospholipids proposed in the so-called “squeeze-out” model [18, 19], such steric forces may promote drug release from the alveolar interface together with surfactant components. By connecting traditional Wilhelmy and Langmuir-Blodgett troughs with interfacial bridges (Figure 1), we have demonstrated that interfacial dynamics actually promotes the loss of interfacial material -which can be recuperated from the subphase-, activates the spreading of new surfactant/drug complexes from donor compartments and endorses the interfacial trip towards recipient surface regions (Figures 4 and 6).

Under limiting conditions of surfactant, the material that is being transferred from the donor to the recipient trough is not replaced, since there is not enough surfactant accumulated in the donor subphase (Figure 7). However, unrestricted amounts of surfactant supplied at the donor trough maintain the interface completely saturated, forming a continuous network of drug-diffusing surfactant between donor and recipient compartments (Figure 4). This highlights the necessity of a continuous network of “fresh” and operative surfactant at the alveolar lining fluid that ensures a constant transference of surfactant into the alveolar interface [54, 55]. In this line, we observed that when the recipient trough has no reservoirs with new surfactant coming from the donor trough (interfaces disconnected), only a little fraction seems to re-adsorb after the interfacial dynamics, at least in a compartment with a relatively large subphase (Figure 6A). Our experiments reveal that the drug can accumulate and be progressively delivered from the interface at particular sites, where three-dimensional structures are formed associated with the interfacial film. These structures, probably formed under the coordination of the hydrophobic proteins SP-B and SP-C during compression (i.e. expiration) [56], are monolayers and bilayers stacked on top of each other, likely enriched in fluid phospholipids [19, 57–59]. This protein-promoted structural reorganization ensures that PS films reach the required low surface tensions to stabilize alveoli at the end of each expiration [18, 60]. Remarkably, most of the material excluded during interfacial dynamics is apparently not replaced. Only when donor and recipient interfaces were maintained connected during and after interfacial dynamics, we observed increase in surface pressure at the recipient compartment, indicating that most excluded material can be replaced by “fresh” surfactant coming from the donor reservoirs (Figure 6). These results are in line with the existence of “used” and less surface-active surfactant pools excluded from the interface [61]. This “used” material has been extensively reported in the literature as small aggregates enriched with unsaturated phospholipids and SP-C [18, 61–64]. Interestingly, epifluorescence images revealed that both small and large 3D structures were full of both the drug and the probe that partitions into the most fluid phospholipids (Figure 5). Quantitation of TAC fluorescence showed that the amount of drug at the interface and tightly adjacent regions decayed after interfacial dynamics from the very first compression (Figure 4G), suggesting that the drug is likely excluded from the interface to the subphase during interfacial dynamics. We in fact detected TAC liberated into the subphase as a consequence of compression-expansion dynamics of the interfacial film (Figure S4). Thus, in a similar manner to unpackable PS components (e.g. fluid unsaturated phospholipids) [18, 49, 50], vehiculized therapeutic passengers such as TAC could be released to the alveolar lining fluid together with PS components in the form of small aggregates through these 3D structures (see Figure 10). In a healthy lung, and in order to ensure a proper alveolar homeostasis, the small aggregates are removed from the alveolar lining fluid by type II alveolar epithelial (ATM) cells for recycling [65], alveolar macrophages for clearance [65–67] and lost through the upper airways via the mucociliary escalator [68]. Consequently, and by leveraging the surfactant cleaning process in alveoli, we propose that the “small aggregates” enriched with the drug could act as a Trojan Horse targeting ATI! cells and phagocytic cells.

Figure 10: A proposed model for the effect of interfacial dynamics on surfactant/drug delivery and release.

(A-D) Proposed model for the interfacial performance of PS/TAC under compression/expansion dynamics. (E) Schematic representation of an inflated (upper panel) and deflated (lower panel) alveolus during expiration (expansion) and inspiration (compression), respectively.

We have proven the concept in mice models of LPS-induced ALI pre-treated with TAC or surfactant/TAC combinations. As hypothesized, TAC was selectively taken up by macrophages in combination with PS and the amount accumulating in BAL cells was considerably higher than observed when the drug was administered alone. At the same time, the increased epithelial permeability, as evident in the increased BAL protein content as well as the number of cells (Figure 8), were reduced in LPS-challenged mice pre-treated with PS/drug formulations. These results reveal that 1) surfactant is an efficient interface-assisted drug delivery system, able to reach the alveolar spaces transporting TAC from the nose, 2) it enhances the availability and uptake of TAC by alveolar macrophages and 3) it promotes a concomitant anti-inflammatory effect possibly via macrophages, at least partially. Nonetheless, it is important to highlight that the effects and uptake by other lung cells and the presence of TAC in other locations (e.g. interstitium or blood stream) were not evaluated. Further studies are needed to have a wider understanding of this strategy and explore the potential to reach the interstitium for treating interstitial lung diseases and lung infections such as tuberculosis, or to cross the air-blood barrier for non-invasive systemic delivery. Additionally, another pertinent question is what could be the contribution of mere drug solubilisation on its distribution, uptake and effects, even if using entities like liposomes that, as observed in our in vitro experiments, are not able to adsorb and spread efficiently along the interface. Although our surfactant-free control formulation shows low uptake and anti-inflammatory effects, it would have been also of interest to compare PS-assisted vehiculization with the canonical vehiculization of a protein-free liposomal preparation delivered in vivo.

Our in vivo experiments, designed to serve as a proof-of-concept of the vehiculization activity of surfactant in vivo, were designed as a prophylactic strategy, instead of a therapeutic treatment. This is likely the most favorable scenario, where a fully open and operative respiratory surface maximizes the chances of surfactant to rapidly spread from the upper towards the distal airspaces. The situation could be different when PS/drug formulations are delivered in a pathological context, where e.g. inflammation and lung injury could compromise lung aeration and interface-assisted surfactant distribution [69]. Still a proper combination of PS and drugs can have an even more synergistic effect, with surfactant enhancing the opening of the smallest airspaces to air and the subsequent efficient distribution of the drug. This is something that should be tested in complementary experiments administering the drug alone or in combination with surfactant after inducing lung injury.

We also observed that the treatment with only surfactant tends to prevent protein leakage but not WBC extravasation. Likely as a consequence of the intrinsic surface active properties of PS, which contributes to maintain the alveolar lining fluid stable preventing the ultra-filtration of fluid coming from the blood stream [70]. Different specific formulations of therapeutic surfactants could be designed for drug delivery depending on the clinical conditions of target lungs (properly aerated lungs or injured lungs requiring restoration of PS inhibition). The proper choice of surfactant doses along with the optimal surfactant/drug proportions could be crucial to allow for their rapid and efficient distribution along the interface. Understanding the efficacy and safety of alveolar drug therapy is urgent now because of the pandemic of viral pneumonia associated with COVID-19, in which the SARS-Cov-2 virus targets lung alveoli to cause severe ALI [71]. ALI leading to the Adult Respiratory Distress Syndrome (ARDS) has killed more than half a million people worldwide and infected >16 million. Several drugs are in patient trials for COVID-19 therapy. Airway delivery of one of these is under serious consideration so that the virus within alveolar cells can be targeted directly. These considerations suggest that PS-assisted drug delivery may play a critical role, not only because of the intrinsic therapeutic properties of PS [72], but also in ensuring that anti-viral and anti-inflammatory therapies reach the site of ALI pathology.

The results presented in this work also illustrate the utility of biophysical models to analyze the elementary processes involved in the interfacial pulmonary drug distribution, which could be tailored to study the behavior of any particular drug or therapeutic agent (e.g. peptides, proteins, RNA, etc.). Biophysical models are a potent tool to design and optimize new surfactant-promoted vehiculization strategies prior to in vivo experimentation, which may reduce ethical impact and costs of using animal models. Important aspects that can be optimized in a case-by-case basis include: 1) the efficient incorporation of therapeutic vehiculizable entities (e.g. drugs, antibodies, siRNA or nanocarriers), 2) the maximal proportion of incorporated drug that can still maintain the integrity of interfacial properties of PS, 3) the ability of any PS/drug combination to be vehiculized via the interface and 4) their interfacial behavior in terms of surface-assisted diffusion and delivery under breathing-like dynamic conditions.

5. CONCLUSIONS

The use of PS as a drug delivery system seems to be highly promising, with 1) PS transporting the drug potentially by surfing the air-liquid interface towards the distal airways, 2) breathing dynamics contributing to drug diffusion and release together with surfactant components, and 3) clearing mechanisms promoting the uptake of surfactant components and drug by alveolar macrophages. The interface-assisted surfactant-supported drug vehiculization should then open novel avenues to treat diseases where drug distribution and targeting to alveolar macrophages or ATII cells could play a key role such as pulmonary infections (e.g. tuberculosis), pulmonary fibrosis, ARDS or other diseases derived from lung injury and inflammation. Understanding the insights of this inhalable respiratory drug delivery strategy becomes beneficial to optimize the versatile distribution, targeting and effectiveness of very different therapeutic agents such as specific drugs, antibodies, siRNA or vaccines. Nevertheless, the extension of this innovative drug delivery approach will require the optimization and production of new surfactants with appropriate adsorption and spreading properties along air-liquid interfaces, good ability to vehiculize therapeutics and enhanced resistance to inactivation by pro-inflammatory components. In parallel, more patient-friendly forms of delivery such as aerosolized formulations should be optimized in substitution of the typical endotracheal administration that sustains current delivery of clinical surfactant in the treatment of newborns. Such considerations, along with the production of a novel generation of synthetic surfactants in large quantities with well controlled and reproducible composition at reasonable cost, will pave the way to the consolidation of this novel but promising drug delivery strategy: the interfacial delivery.

Highlights.

• Pulmonary surfactant is proposed as a drug delivery system with the potential to solubilize and distribute poorly water-soluble drugs over the respiratory air-liquid interface.

• Breathing dynamics promote drug spreading and release.

• Pulmonary surfactant/drug formulations show synergistic anti-inflammatory effects targeting alveolar cells.

ABBREVIATIONS

PS

pulmonary surfactant

NS

native surfactant

OE

organic extract of NS

TAC

tacrolimus

BAL

bronchoalveolar lavage

LPS

lipopolysaccharide

WBC

white blood cells

ATI

alveolar epithelial cells type I

ATII

alveolar epithelial cells type II

AMΦ

alveolar macrophage

i.n.

intranasal

AFM

atomic force microscopy

CBS

captive bubble surfactometer