Preclinical Testing of a New Dry Powder Aerosol Synthetic Lung Surfactant Formulation and Device Combination for the Treatment of Neonatal Respiratory Distress Syndrome

Formulation materials

PLs (DPPC and POPG) and surfactant protein mimic peptide (B-YL) were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL) and CSBio (Menlo Park, CA), respectively. l-leucine was purchased from Sigma-Aldrich Chemical Co. (St Louis, MO). Growth excipients NaCl and MN together with other chemicals were purchased from Fisher Scientific Co. (Hanover Park, IL). Throughout the study, freshly collected deionized water was used.

Formulation production and analytical methods

Multiple batches of the micrometer-sized SLS-EEG MSD2 formulation were produced using the two-fluid spray nozzle-based Buchi S-300 Mini Spray Dryer system (Büchi Labortechnik AG, Flawil, Switzerland). Briefly, the feed dispersions were prepared with a solids composition of DPPC 55%, POPG 7.5%, surfactant protein B peptide mimic B-YL 2.5%, mannitol 9%, sodium chloride 6%, and l-leucine 20% w/w in an ethanol-water (5:95% v/v) cosolvent system, which was sonicated for 40 minutes in a heated water bath (45°C–55°C) with intermittent manual shaking every 10 minutes.^60,64 The resulting PL (DPPC and POPG) and B-YL nominal content of the MSD2 formulation were 62.5% and 2.5%, respectively, compared with the previous NSD1 formulation with corresponding theoretical values of 65% and 5%, respectively. The total solids content of the MSD2 feed dispersion was 0.125% w/v. During spray drying, the feed dispersion was continuously stirred using a magnetic stirrer. Spray drying conditions included the use of a 0.5 mm spray nozzle, drying gas and spray gas flow rates of 30 m³/hr and 1200 L/hr, respectively, drying temperature of ∼80°C, and an outlet temperature of ∼45°C–46°C. Dried powder formulations were collected in a high-efficiency cyclone with a backup bag filter. Immediately postcollection, powder formulations were equilibrated in a glass vial (∼17% relative humidity [RH] at ∼24°C for 30 minutes) and then stored in a sealed desiccator at 0% RH/−20°C until needed for use. ⁶⁴

Validated high-performance liquid chromatography (HPLC) methods were used to quantify DPPC mass in determining aerosol size distribution from cascade impaction and lung delivery efficiency from the in vitro lung chamber (LC) model. The HPLC system consisted of an Alliance 2695 Separation Module with a photodiode array ultraviolent (UV) detector and Empower 2 data acquisition software (Waters Corporation, Milford, MA). The sample injection volume was 100 µL and column temperature was 45°C. The chromatographic separation was achieved using the Waters XSelect CSH C18 XP (2.5 µm, 4.6 × 100 mm; Waters Corporation, Milford, MA). Further details including the descriptions of the gradient elution system and UV absorbance, as well as a separate method using a similar approach for the quantification of B-YL used in content uniformity testing, were previously reported in Momin et al. ⁴⁸

Laser diffraction particle sizing

Powder dispersibility and primary particle size of the SLS-EEG MSD2 formulation were determined using laser diffraction at multiple dispersion pressures. Specifically, components employed included the Sympatec HELOS (Sympatec GmbH, Germany) equipped with a RODOS/M disperser and an ASPIROS sample feeder. A small glass vial was filled with ∼3 mg of spray-dried powder, placed in the sample feeder, and dispersed in the laser beam using the RODOS/M disperser with an R1 lens at dispersion pressures of 0.5 and 4.0 bar. The particle size distribution results (Dv₁₀, Dv_50, and Dv₉₀) were obtained using Sympatec WINDOX 5.10 software (Sympatec GmbH, Germany).

Powder imaging

The particle surface morphology of the SLS-EEG MSD2 formulation was evaluated using a Hitachi SU-70 scanning electron microscope (SEM; Hitachi High-Tech, Tokyo, Japan) at an accelerating voltage of 5.0 kV. A small amount of powder sample was mounted on a metal SEM stub using double-sided carbon tape, and loose particles were removed using compressed air. The samples were sputter coated with platinum grain using a Denton Desk V sputter coater (Denton Vacuum, Moorestown, NJ) before imaging. Aperture, astigmatism (X&Y), and beam alignment were adjusted for optimal imaging, and then the image was captured at 5 kV with a 5 mm working distance.

To capture internal particle structure, a combination of focused ion beam (FIB) cutting and SEM imaging approach was employed through the use of a Scios 2 DualBeam system (ThermoFisher Scientific, Waltham, MA). Once the sample preparation was completed as described above, the specimen was fixed in the sample holder and then inserted into the FIB chamber. Contrast and brightness, astigmatism (X&Y), and focus were adjusted to identify an area of interest using the SEM mode. Then, in FIB mode, the sample was positioned at a 7 mm working distance from the electron beam and tilted at 52° with horizontal. The FIB column was set to 30 kV with ion beam current of 10 pA to mill the particles while minimizing damage to the internal structure, and then the SEM mode was activated to capture the image at 5 kV with a T2 detector setting.

X-ray powder diffraction and moisture sorption testing

The Rigaku Miniflex 6G (Rigaku, Japan) X-ray powder diffraction (XRPD) technique was used to characterize the SLS-EEG MSD2 formulation. Briefly, a small amount of powder was placed on the center of the XRPD sample holder and then the powder surface was gently smoothed using a glass slide. The samples were scanned using a continuous mode in the range of 3°−50° 2θ at a scan speed of 1.5°/min.

To understand the water uptake of the SLS-EEG MSD2 formulation across a wide range of RH values, samples were exposed using the dynamic vapor sorption system (DVS Adventure, Surface Measurement Systems Ltd., UK). Briefly, ∼5mg of powder sample was equilibrated at 0% RH and then exposed to increasing RH from 0% to 95% (sorption cycle) and decreasing RH from 95% to 0% (desorption cycle). At each RH, once the defined equilibrium condition (the rate of mass change [dm/dt] <0.002%) was achieved, the RH was changed to the next level.

Aerosol delivery system and D3 AJ DPI

The iDP-ADS used in this study consisted of an electromechanically controlled timer air source (EM Timer), aerosolization engine, pressure monitoring and control (PMC) unit, and ETT interface connector (Fig. 1). Considering the EM Timer, the inlet pressure regulator was set to provide a 3 L/min flow rate to the device of compressed air (in vitro testing) or pure oxygen (in vivo testing). The timer setting was adjusted to deliver a gas actuation volume (GAV) of 10 mL during in vitro testing and ∼7 mL/kg for the in vivo experiments. In the animal model experiments, a table was used to ensure that for the measured lung compliance of individual animals, the delivery of 7 mL/kg would result in an elastic lung pressure rise (ΔP_E = 1/C*GAV, where C is the lung compliance in milliliters per centimeters of water) ≤ 20 cmH₂O. ⁴⁸ If the ΔP_E was greater than the 20 cmH₂O limit, then the GAV was reduced to meet the pressure limit, at the discretion of the administering clinician. Further details of this procedure, and the corresponding lookup table, were provided in the previous study by Momin et al. ⁴⁸

OPEN IN VIEWER

FIG. 1.

Components of the infant dry powder air-jet aerosol delivery system (iDP-ADS) including (A) setup of the D3 air-jet DPI with connection to the electromechanical (EM) Timer air source, pressure monitoring and control (PMC) unit, and endotracheal tube (ETT) interface connector, (B) inner flow pathway for the D3 air-jet DPI with variable aerosolization chamber, and (C) model of the variable aerosolization chamber components including lower unit level indicators. ID, internal diameter.

The PMC unit enabled the control of respiration, pressure monitoring, and pressure protection during the in vivo experiments and in vitro LC testing. A flow control orifice (0.5 mm diameter) was used in the exhalation port to slow exhalation and better enable the administrator to maintain a positive end expiratory pressure (PEEP) of ∼5 cmH₂O (Fig. 1A). Additional ports on the PMC unit led to an analog pressure gauge (AG Industries, St. Paul, MO) and pressure relief valve, adjusted to ∼25 cmH₂O. For device actuation, at a lung pressure of ∼5 cmH₂O, the administering clinician covered the exhalation port, creating an airtight seal with the lungs, and actuated the device. The exhalation port remained covered for ∼1 second. The administering clinician then uncovered the exhalation port and monitored the lung pressure during exhalation until ∼5 cmH₂O was achieved, at which point the process was repeated. Approximately four device actuations were used, accompanied by a defined rotation of the variable-volume chamber lower unit after each actuation, for each 30 mg device-loaded dose.

A primary component of the iDP-ADS was the infant AJ DPI (Fig. 1B). As previously described, AJ DPIs are designed with geometrically simple elements (small diameter flow passages and smoothly curving aerosolization chamber) and engineered to enable the efficient generation of small-particle aerosols with limited volumes of air.^{55,56,65–67} In previous studies, the aerosolization chamber has been of fixed dimensions and volume, which for the SLS-EEG and other surfactant formulations limited device loading to ≤20 mg to maintain high quality aerosolization,^48,59 in the absence of a cyclic loading mechanism.^52,68 Higher mass loadings either emitted the aerosol too quickly, degrading aerosol quality, or had reduced ED. In this study, we introduce a new variable-volume aerosolization chamber with a movable lower unit, which controls the floor height of the chamber (Fig. 1B, C). For loading of high powder masses, the floor of the aerosolization chamber is positioned at the lowest level, maximizing the available chamber volume. With a first actuation, the geometry is designed such that secondary air currents aerosolize a defined mass of powder on the top of the powder bed while leaving the remaining powder relatively undisturbed. The floor of the variable-volume chamber is then raised by a defined distance using a small turn of the lower unit (Fig. 1C), exposing the next dose of powder for aerosolization with the next actuation. By following an actuate-and-turn sequence, the dose per actuation can potentially be unified while enabling high powder mass loading and maintaining high aerosol quality. As illustrated in Figure 1B, alignment of the positioning tab on the lower unit with the numbers indicating levels of the aerosolization chamber floor was used to guide rotation of the lower unit through four defined levels. During in vitro testing, one actuation per level was employed. During in vivo delivery, one actuation per level was also used; however, the operator could employ additional actuations at each level with minimal powder emissions if needed to maintain animal ventilation.

Aerosol size characterization using cascade impaction

The aerosol emitted from the D3 AJ DPI with the ETT adaptor attached was characterized using cascade impaction with a next-generation impactor (NGI; MSP, TSI Incorporated, Shoreview, MN). The D3 device was connected to the preseparator inlet of the NGI, as previously illustrated by Howe et al., ⁵² using a custom adapter. The NGI was oriented at 90° to horizontal to enable the device to remain in the intended level orientation for use, and the custom adaptor positioned the ETT adapter of the D3 device in the center of and ∼1 cm away from the preseparator inlet. Side openings in the custom adapter enabled the entrainment of room air around the aerosol plume, allowing the NGI to be operated at a constant flow rate of 45 L/min, generated from a downstream vacuum source. The NGI flow rate of 45 L/min was chosen to ensure collection of the aerosol, minimize any effects of settling, and provide appropriate stage cutoff diameters for evaluating small aerosol sizes. Each stage of the NGI was coated with MOLYKOTE® 316 silicone spray (Dow Corning, Midland, MI) to minimize particle bounce and re-entrainment. Before each set of experimental runs, the flow rate was confirmed using a flow sensor (Sensirion SFM3000, Sensirion AG, Stafa, Switzerland) connected to the NGI inlet.

For aerosol size determination, the D3 device was loaded with 30 mg of the MSD2 formulation. Consistent with aerosol delivery to a 1.5 kg infant, 10 mL GAVs were employed at 3L/min, which required an actuation time of ∼0.21 seconds. One actuation was used for each height of the variable-volume chamber device across four levels (three marked as shown in Fig. 1C with a final top level at the maximum floor height). Preliminary development experiments were conducted to determine the correct chamber volume and rate of floor rise to efficiently deliver a high-quality aerosol and maintain a high ED for a 30 mg device-loaded formulation mass and four device actuations, resulting in a target powder mass per actuation of ∼5–10 mg. Three replicate runs for each experiment were performed, and quantitative analysis of DPPC deposition using the HPLC-UV assay was used to assess the aerosol performance characteristics of the SLS-EEG powder formulation. Analysis metrics included ED, MMAD, FPF, and geometric standard deviation. Percent ED (%ED) was calculated as the mass of DPPC in the loaded dose minus the mass of DPPC remaining in the device divided by the initial loaded mass of DPPC (×100 for %ED). Aerosol size calculations were based on the mass of DPPC recovered in the NGI (including preseparator). Based on an airflow rate of 45 L/min, the NGI stage cutoff diameters were determined using the formula specified in USP 35 (Chapter 601, Apparatus 5). The MMAD and FPFs (<5 µm, <2.5 µm and <1 µm) were calculated using Copley Inhaler Data Analysis Software (CITDAS).

Characterization by actuation using SprayVIEW, Spraytec, and emitted mass

A combination of light sheet microscopy (Proveris Scientific SprayVIEW®), laser diffraction particle sizing (Malvern Spraytec®), and gravimetric analysis was used to characterize aerosolization performance on the basis of individual actuations. For each experiment, 30 mg of the MSD2 formulation was loaded into the D3 device, and the device was actuated one time at each level of the variable-volume chamber, denoted as actuations A1–A4, using 10 mL GAVs delivered at 3 L/min from the EM Timer. All experiments were completed in triplicate, denoted as runs R1-R3.

With the SprayVIEW system, the device outlet capillary was placed at the furthest position within the camera field of view, and the laser housing was adjusted to the center of the orifice in the vertical and horizontal directions. To simultaneously record gas flow rate, an inline flow meter (Sensirion SFM3400-AW and RS485-USB/2028) was connected between the EM Timer outlet and D3 device inlet. Image acquisition was set at a frame rate of 500 Hz over a 1-second time acquisition window. As with all in vitro cases, the EM Timer was actuated for 0.21 seconds, which generated the 10 mL GAV at a flow rate of 3 L/min. Analysis of the SprayVIEW data was used to report plume geometry, velocity, and intensity. While a quantitative correlation is not available between SprayVIEW-reported plume intensity and aerosol mass emission, intensity over actuation time was used as a qualitative metric of aerosol emission relative to device actuation flow.

To determine aerosol size distribution for each of the four actuations, the Malvern Spraytec laser diffraction system was employed in the default (open bench) configuration with 1 kHz data acquisition rate and employed a 300 mm lens (0.1–900.0 µm nominal size range). Refractive indices were selected as standard opaque particles (1.50 + 0.50i) sprayed into air (1.00 + 0.00i). Considering the density of the aerosol cloud, necessary system modifications included only considering detector ranges from 16 to “Last” and removal of plastic flanges to prevent cloud recirculation into dead zones. All other parameters remained under manufacturer default values. Concentrated-weighted, time-averaged aerosol polydisperse size distributions (PSDs) and median volumetric diameters (D_v50) values were determined using the Malvern Spraytec 4.00 software for each of the four device actuations (A1–A4).

The emitted formulation mass for each actuation was determined in separate experiments in which the unloaded and loaded D3 device was weighed, using an Ohaus Analytical Plus (Model: AP250D) microbalance (Ohaus Corporation, NJ, USA), and then the loaded device was reweighed after each actuation. Following each actuation, the emitted mass for that actuation was determined as the difference between the previous loaded device mass and current device mass (in mg). %ED per actuation was calculated as the percentage of the emitted mass relative to the original loaded mass (loaded device mass − unloaded device mass). Total emitted mass and %ED were determined as the sum of the emitted masses over the four actuations and the percentage of this sum relative to the original loaded mass, respectively. In method validation experiments, gravimetric analyses of percent emptying matched HPLC quantification (to within 10% relative difference), further validating gravimetrically recorded mass losses for calculating %ED and emitted mass.

In vitro testing of lung delivery efficiency with the LC model

To approximate the lung delivery efficiency achieved in the in vivo animal model experiments, a realistic rabbit LC model was employed, as previously described by Momin et al. ⁴⁸ (Fig. 2). Briefly, upper airway aerosol transport and deposition was captured using a 3D-printed model of the upper TB airways extending from the trachea to outlet generations from approximately G4 to G6, which was extracted from a computed tomography scan of a small mammal with body mass, airway structure, and dimensions very similar to young mature New Zealand White Rabbits. ⁴⁸ Outlet diameters of the upper TB region ranged from 0.8 to 2.3 mm. A majority of the trachea was constructed in flexible silicone material to enable realistic insertion of the ETT with airway sealing. The upper TB model was surrounded by a cylindrical LC, with a total sealed system volume of ∼500 mL, resulting in a realistic measured compliance of 0.45 mL/cmH₂O, consistent with preterm infants experiencing RDS. A filter (Pulmoguard II™, Queset Medical, North Easton, MA) and pressure release valve were used to capture aerosol that did not deposit in the upstream components. To evaluate regional aerosol deposition in the LC model, the inner walls of the airway model were coated with MOLYKOTE 316 silicone spray (Dow Corning, Midland, MI) to minimize particle bounce and simulate an airway fluid coating. The D3 device was loaded with 30 mg of the SLS-EEG formulation and connected to the ETT, which was inserted into the flexible silicone trachea. The LC model was positioned horizontally on a flat surface, similar to the orientation of the rabbits in the in vivo experiments. The AJ DPI was actuated at 3 L/min with 10 mL GAVs one time at each of the four variable-volume chamber levels (Fig. 2), with the LC pressure monitored and released after each actuation. After removing the AJ DPI ETT connection, a final pressure chamber evacuation was performed to capture any remaining suspended aerosol on the filter using a downstream vacuum pump. Deposited powder formulation was recovered from the device, ETT, upper TB model, chamber, and filter using appropriate wash volumes of methanol, and those solutions were analyzed by HPLC-UV to determine the DPPC content. The chamber and filter deposition fractions were grouped together as a lower lung deposition fraction. The upper TB, chamber, and filter deposition fractions were grouped together as a total lung deposition fraction. All deposition fractions were expressed as a percentage of device nominal loaded dose based on the HPLC quantification of DPPC and were calculated from loaded powder mass and previously determined DPPC content.

OPEN IN VIEWER

FIG. 2.

Actuation of the D3-MSD2 device and formulation combination into the in vitro lung chamber (LC) model including an upper tracheobronchial (TB) region consistent with a ∼1.5 kg rabbit. The time point shown is just after the second actuation visualizing a standing cloud of aerosol in the chamber (from actuation A1) with the addition of incoming jets of aerosol flow from the upper TB model outlets associated with A2. G#, airway generation number.

In vivo surgical preparation and instrumentation

All experimental animal procedures were conducted according to the National Institute of Health’s Guide for the Care and Use of Laboratory Animals and approved by the Seattle Children’s Institutional Animal Care and Use Committee. Juvenile female New Zealand White Rabbits (Western Oregon Rabbit Company, Philomath, OR) were tranquilized, sedated, instrumented, monitored, and ventilated according to identical methods established in previously published animal studies.^11,18,69 At the time of in vivo testing, body weight of the D3-MSD2 aerosol group was 1.59 ± 0.23 kg, and body weight of the entire cohort including sham and Curosurf liquid historical controls was 1.56 ± 0.07 kg.

Animals were administered rocuronium bromide to induce neuromuscular blockade during the lung lavage and surfactant administration. Subsequent rocuronium doses were withheld to promote spontaneous inspiratory efforts to assist mandatory ventilator breaths, consistent with the clinical management of preterm infants. Before lung lavage for baseline RDS condition, arterial blood gases (ABGs), lung mechanics, ventilation parameters, and vital signs (i.e., heart rate, blood pressure, and temperature) were acquired at a time point of −60 minutes, as summarized in Figure 3.

OPEN IN VIEWER

FIG. 3.

Time schedule for animal experimental procedures. The sequence of procedures for lung lavage to establish surfactant deficiency and lung injury, treatment groups, time schedule for procedures, surfactant dosing, and physiological measurements is shown. ABG, arterial blood gas; SLS-EEG, synthetic lung surfactant–excipient enhanced growth formulation.

In vivo induction of surfactant deficiency and lung injury

Severe lung injury was induced by a multiple lung insult (hit) model that included hyperoxia, atelectotrauma, extensive surfactant washout (typically 6–10 cycles), and mechanical ventilation.^11,19 For injury induction, the FiO₂ was increased to 1.0, and PEEP was reduced to 0 cmH₂O. The lungs were depleted of surfactant and injured by cyclic alveolar collapse (PEEP of 0 cmH₂O) with repeated lavage using 25 mL/kg of warmed (39°C) normal saline (0.9%), with 5-minute recoveries between lavages until the static lung compliance was 50% of the prelavaged value, SpO₂ of 90%–92% on 1.0 FiO₂ and a PEEP of 5 cmH₂O. ⁶⁹ The animals were deemed surfactant deficient when PaO₂ <100 mmHg was confirmed by two consecutive ABGs 30 minutes apart following lavage. Additional lavages were administered as needed to establish these surfactant depletion and lung injury goals. Following the establishment of surfactant deficiency and severe lung injury, the ventilator rate was initially titrated based on PaCO₂ values of 55–65 mmHg and pH 7.25–7.35, and animals remained on these settings throughout the rest of the experiment. If pH <7.25 at any point, the respiratory rate was increased as needed. Animals with PaCO₂ <40, pH >7.40, and evidence of alveolar overdistension (reduced compliance for the terminal 20% of inspiration, C₂₀/C ratios <0.80) ⁷⁰ had the respiratory rate reduced to prevent pneumothorax.

In vivo experimental protocol

The sample size for the experimental animal group (SLS-EEG) was calculated with a 95% confidence interval, a power of 90%, and a standard deviation (SD) of 0.20; the total recommended sample size was seven (n = 7) for the experimental group based on the mean and SD for PaO₂ from a previous animal study that compared between two different surfactant aerosol delivery systems. ⁷¹ The negative control (sham) and positive control (liquid Curosurf) groups included four and five animals in each group, respectively.

Figure 3 shows a diagram of the animal procedures, study time points during the experiments, and the DDP. Following surfactant washout, animals in the aerosol treatment group received MSD2 aerosol formulation from the iDP-ADS using the D3 device (n = 7). Animals receiving the D3-MSD2 combination aerosol therapy were compared with historical controls previously reported in DiBlasi et al. ¹¹ that received either a clinical dose of intratracheally instilled liquid surfactant (Curosurf at 200 mg/kg; n = 5) or the aerosol delivery process without powder (n = 4) as a (negative sham) control at time point 0. All subjects were supported for 3.5 hours following the initial intervention.

Intratracheal liquid surfactant administration (Curosurf)

As reported in DiBlasi et al., ¹¹ the animals in the liquid Curosurf control group received standard intratracheal administration of porcine-derived Curosurf surfactant (poractant alfa, Chiesi, Parma, Italy). Curosurf comprises ∼96% PLs (80 mg/mL) and 1% surfactant proteins B and C and is only approved for administration by tracheal liquid instillation. The vials of liquid Curosurf (200 mg/kg; 2.5 mL/kg) were drawn up into a syringe and administered based on a standard protocol, described in greater detail elsewhere. ⁹

Dry powder surfactant aerosol delivery

Animals in the D3-MSD2 aerosol group received 60 mg (device-loaded dose) dry powder formulation delivered directly to the ETT. The 60 mg total device-loaded formulation mass was divided into two dose cycles, each delivering 30 mg device-loaded formulation masses and administered 15 minutes apart (t = 0 and 15 minutes). This dose magnitude and delivery timing was selected for consistency with the previous study of DiBlasi et al. ¹¹ where we used the D2-NSD1 combination, in order to enable direct comparison of the in vivo responses and best isolate the in vivo impact of D3-MSD2 improvements in aerosol quality. With the updated MSD2 formulation, containing less B-YL, and a 1.59 kg average rabbit weight for the aerosol group, the initial total device-loaded PL dose was 24 mg-PL/kg, which is consistent with the 24 mg-PL/kg from the previous NSD1 formulation.^11,48 It is noted that each 30 mg device-loaded formulation mass of MSD2 contains only 12 mg-PL/kg. Furthermore, the experimentally determined device %ED of ∼77.4% (NGI) from the ETT adaptor results in a device total emitted mass of 18 mg-PL/kg entering the ETT. Considering the Curosurf dose of 200 mg-PL/kg is also delivered directly to the ETT, we view the AST as delivering approximately one order of magnitude less active therapeutic (PL dose) to the animal.

During aerosol delivery, the D3 AJ DPI was connected directly to the rabbit ETT, enabling precise control over the peak inspiratory pressure (PIP) of 25 ± 2 cmH₂O, PEEP of 5 cmH₂O, V_T of 7 mL/kg, flow of 3 L/min, and actuation time of ∼0.20–0.22 seconds with the iDP-ADS (see Fig. 4). The EM Timer settings for the DPI were determined based on the postlavage compliance values and weight (kg) using a reference calculation chart, which is described in greater detail in the study of Momin et al. ⁴⁸ A total of ∼4 actuations were employed to deliver each 30 mg device-loaded formulation mass. The iDP-ADS employed an adjustable pressure relief valve (∼25 cmH₂O) and an analog pressure manometer, akin to a manual resuscitator, to measure pressure delivery and prevent barotrauma. The powder chamber was manually filled with formulation from previously packaged capsules (Qualicaps Inc, Whitsett, NC) that were stored in Al-Al blister packs, and the DPI was actuated to deliver the aerosol and inflate the lungs, followed by an ∼1-second breath hold, after which exhalation was allowed through the exhalation port. The animal was promptly returned to the ventilator following each of the two dosing cycles (with 30 mg device-loaded formulation mass each) conducted at t = 0 and 15 minutes.

OPEN IN VIEWER

FIG. 4.

In vivo testing setup illustrating aerosol delivery from the iDP-ADS using the D3 air-jet DPI with the ETT interface. iDP-ADS, infant air-jet dry powder-aerosol delivery system; PEEP, positive end expiratory pressure; PIP, peak inspiratory pressure.

Physiological measurements

Figure 3 shows a diagram of the animal procedures, study time points during the experiments, and the DDP. Following lung lavage, gas exchange (ABGs), ventilation parameters, and hemodynamic measurements were obtained at the pretreatment (RDS baseline) condition at −5 minutes. Between −5 and 0 minutes, the initial D3-MSD2 aerosol (30 mg) dose or sham was delivered, or liquid instillation of the total dose of Curosurf was completed. Five-minute time point measurements were then taken to assess the short-term gas exchange effects immediately following therapy due to lung volume derecruitment from the initial ventilator disconnection with the D3-MSD2 aerosol therapy or sham delivery, potential airway obstruction from liquid Curosurf or powder agglomeration, as well as the potential for very rapid response to the AST. The D3-MSD2 aerosol group had two additional measurements at 10 and 15 minutes, the latter of which occurred just before the second 30 mg device-loaded dose. Standard posttreatment measurements were obtained at 30-minute intervals thereafter for 3.5 hours (see Fig. 3).

Respiratory rate, PIP, PEEP, and mean airway pressure were acquired from the ventilator. The pressure–volume relationship was evaluated by performing a brief inspiratory hold and recording the static compliance. The compliance was normalized based on the individual rabbit weights (milliliter per centimeter of water per kilogram). The OI and ventilation efficiency index (VEI) were calculated using ventilator parameters and ABG values to assess disease severity at each time point.^72–74 Animals were euthanized with 100 mg/kg Euthasol following the 3.5-hour postsurfactant measurements and remained ventilated for necropsy.

Statistical analysis

Statistical analysis of the in vitro data for comparing aerosol delivery performance and simulated lung delivery efficiencies was performed using JMP Pro 17 (SAS Institute Inc., Cary, NC). For in vivo data, results were calculated as mean ± standard error of the mean for all continuous physiological variables at each time point for surfactant-treated and control groups. Two-way analysis of variance was used to assess the treatment effects of each group and time point, as well as their interactions. Tukey’s multiple comparisons test compared mean post hoc differences in physiological outcomes at pre and postlavage baseline and at each time point following treatment between the different groups. The criterion for significance was established a priori as p < 0.05 for all comparisons. The two additional measurements at 10 and 15 minutes in the D3-MSD2 aerosol animals were descriptive and not included in the statistical analysis, based on a lack of data in the sham and Curosurf liquid historical controls. ¹¹

Spray drying and powder formulation particle sizing

Over the course of the study, we produced multiple batches of the SLS-EEG MSD2 formulation using the Buchi S-300 Mini Spray Dryer system. The spray rate of five batches of MSD2 powder formulation was ∼2.1 mL/min. The DPPC content (mean ± SD) of these five batches assessed using the previously validated HPLC-UV method was 55.5 ± 2.9%w/w, which was 101.1% of the nominal content (55.0%). The B-YL content (mean ± SD) assessed over the same five batches using a validated HPLC-UV method was 2.5 ± 0.1%w/w, which was 100.7% of the nominal content (2.5%). The coefficient of variation for DPPC and B-YL content was ∼5%, indicating good reproducibility in key components across batches.

Table 1 provides the volumetric particle size of the MSD2 formulation determined using the Sympatec laser diffraction technique at dispersion pressures of 0.5 and 4.0 bar, tested across five different batches. The primary particle size (Dv₅₀) of the MSD2 formulation batches measured at 4.0 bar dispersion pressure was 1.2 µm with coefficient of variation 2.1%, reflecting the consistent size of the formulation for the five batches considered and reproducibility of the powder preparation process. The D_v50 (mean ± SD) of the batches measured at 0.5 bar dispersion pressure was 1.8 ± 0.0 µm, suggesting good dispersibility of the powders. Overall, the spray drying process produces micrometer-sized particles with good dispersibility, which are suitable for the EEG approach and deep lung delivery.

Powder formulation imaging

SEM images of particles composing the MSD2 formulation are displayed in Figure 5, at magnifications of 20k and 50k. Imaging of the external particle structure indicated wrinkled surfaces with relatively spherical particle shapes and geometrical diameters consistent with Sympatec volumetric diameter results at 4.0 bar testing (Fig. 5A). FIB sectioning of particles indicated relatively large voids were present in most particles with particle diameters approximately ≥1 µm. As illustrated in Figure 5B, internal voids had diameters of ∼100–300 nm with between-void spacings of ∼100 nm or greater. Smaller openings, as shown at location 1 in Figure 5B, progressed to larger voids as with location 2 in Figure 5C during successive FIB sectioning.

OPEN IN VIEWER

FIG. 5.

Outer surface and inner structures from FIB/SEM imaging of the SLS-EEG MSD2 particles indicating (A) relatively spherical particles with wrinkled outer surfaces and (B, C) intermediate size inner hollow voids. FIB sectioning revealed intermediate sized void structures (∼100–300 nm) in most particles ≥1 µm upon sufficient sectioning. Location 1 in Panel B is the start of an addition pore that is revealed at location 2 in Panel C with the next FIB cut of ∼200 nm. Location 3 then indicates the start of a deeper void. FIB, focused ion beam; SEM, scanning electron microscope.

XRPD and moisture sorption testing

Figure 6 shows the solid-state properties of the MSD2 powder formulations characterized using XRPD and DVS. The X-ray powder diffractograms of three different batches of the MSD2 formulations (Fig. 6A) showed evidence of a partially crystalline structure with no differences in the peak diffractograms indicating reproducible production of different batches of powders. Characteristic peaks were observed for l-leucine and MN as being in the crystalline state.

OPEN IN VIEWER

FIG. 6.

Physicochemical characterization of the SLS-EEG MSD2 formulation from (A) X-ray powder diffractograms and (B) dynamic vapor sorption (DVS). In Panel b, the particles experience little water mass uptake until >70% relatively humidity (denoted as % P/Po), which is ideal for dry powder EEG aerosol storage and lung targeting.

Figure 6B shows that following an equilibration period at 0% RH, water uptake by the powder at 10%–60% RH (black line) was <9%, indicating relatively low uptake for an unprotected (i.e., unpackaged) spray-dried powder. Above 70% RH, the water uptake increased significantly, indicating the hygroscopic nature of the formulation at elevated RH, which is well suited for the EEG dry powder aerosol application. Low water uptake during storage under ambient conditions and elevated water uptake at higher humidity conditions are important for powder handling and when hygroscopic growth in the humid lung airways is required for EEG particle size increase.

Aerosol size characterization using cascade impaction

Table 2 provides the aerosol particle sizing of the MSD2 formulation delivered with the D3 device (one actuation per level) and a 30 mg device-loaded formulation mass based on cascade impaction with the NGI. The device retention (as measured by DPPC assay) was 22.6% with a preseparator deposition fraction of 5%. The low preseparator fraction led to a FPF <5 µm (FPF_{<5 µm}) of 87.9%, which was more than four times higher than the previously reported FPF_{<5 µm} (19.3%) for the D2-NSD1 combination. ⁴⁸ The mean FPF_{<2.5 µm} was 61.6% after four actuations, suggesting the potential to deliver a substantial amount of powder to the alveolar region. The MMAD of the MSD2 formulation delivered from the D3 device was 2.0 µm, which is lower than the previously reported D2-NSD1 combination of 2.9 µm, ⁴⁸ suggesting more efficient aerosolization and better aerosol quality using the new D3-MSD2 combination.

Characterization by actuation from SprayVIEW, Spraytec, and emitted mass

Figure 7 shows a snapshot of the plume geometry of powder aerosol emitted from the D3 device over four actuations. Mean cone angles of plumes during their maximum signal intensity (∼50 ms from start of actuation and ∼100 mm from tip of the device orifice) for actuations A1–A4 were 26.6°, 29.2°, 30.7°, and 27.1°, respectively, with a 1.9° SD across actuations; corresponding mean cone widths for actuations A1–A4 were 23.7, 26.2, 27.5, and 24.1 mm, respectively, with a 1.8 mm SD across actuations. The plume front velocity for the leading edge of the emitted aerosol cloud up to 100 mm from the orifice for each actuation A1–A4 was 3.0, 1.5, 1.9, and 2.0 mm/ms, respectively, with a 0.6 mm/ms SD across actuations. Plume durations (PDs) measured with the Spraytec analysis for actuations A1–A4 were 319, 315, 313, and 270 ms, respectively, with an SD of 22.8 ms. For comparison, the PD observed in SprayVIEW was 10%–20% longer than with Spraytec. Spraytec PD measurements were expected to be more accurate, as the plume enters and exits the measurement zone with minimum recirculation when the detector tower flanges are removed.

OPEN IN VIEWER

FIG. 7.

Maximum intensity plume images for successive actuations of the spray (taken ∼50 msec after actuation when the leading plume edge has reached ∼100 mm from the tip of the orifice) with the SprayVIEW® system indicating qualitatively consistent emitted dose over each of the four actuations (A1–A4) during a single experimental run (R1) with 30 mg mass loading: (A) A1, (B) A2, (C) A3, and (D) A4. The red vertical line indicates 50 mm from the tip of the orifice (origin; where cone angle lines originate) where the cone angles and widths are taken; the red horizontal line indicates the midline as an angle bisector. Scale is 20 × 10 mm (L × H).

A comparison of simultaneous aerosol emission versus device actuation flow rate over time based on SprayVIEW testing for actuations A1–A4 is provided in Figure 8. SprayVIEW intensity measurements were used to approximate aerosol emission over time and selected as a conservatively longer option compared with the Spraytec measurement of obscuration (100-%Transmission). Flow profiles indicated an airflow actuation duration of ∼0.21 seconds, as expected, as well as an asymptotically decreasing flow through ∼0.3 seconds, likely due to compressible flow effects within the D3 device. Interestingly, for all actuations, a majority of the aerosol plume intensity, approximating powder emission, occurred during the first half of the actuation (∼100 ms). This relationship provides evidence that the D3 device is capable of controlling the amount of emitted mass such that a majority of the aerosol enters the first half of the inhalation volume.

OPEN IN VIEWER

FIG. 8.

Representative graphical output for plume intensity (black line, left axis; SprayVIEW), and gas flow (gray line, right axis; flow meter) for each device actuation of the D3-MSD2 device and formulation combination when delivering a 30 mg loaded mass during a single experimental run (R1) and using one actuation per turn: (A) A1, (B) A2, (C) A3, (D) A4.

Spraytec evaluation of aerosol size distributions for individual actuations A1–A4 over a single run is illustrated in Figure 9. In each case, a majority of the aerosol volume occurs in particles with geometric diameters <2 µm, and very little volume (∼10%) occurred in particles with geometric diameters >10 µm. The volumetric PSD also appears consistent across the four actuations. Mean ± SD Dv₅₀ values evaluated for each actuation A1–A4 resulted in a narrow range of 1.5 ± 0.2 µm to 1.6 ± 0.3 µm. Furthermore, Spraytec laser diffraction FPF values <2.5 µm based on volumetric diameter for actuations 1–4 were 64%, 68%, 65%, and 71%, respectively. In summary, both the mean geometric aerosol size and the particle size distribution were very consistent across the four actuations of the D3 device loaded with 30 mg of the MSD2 formulation and similar to cascade impaction aerodynamic sizing (Table 2).

OPEN IN VIEWER

FIG. 9.

Representative Spraytec® laser diffraction particle size distributions for each device actuation of the D3-MSD2 device and formulation combination when delivering a 30 mg loaded mass during a single experimental run (R1) and using one actuation per turn: (A) A1, (B) A2, (C) A3, and (D) A4. Reported Dv₅₀ values represent the mean (standard deviation; SD) of the data for each actuation number evaluated over three runs (n = 3).

Percent emitted dose (%ED) and emitted mass are provided in Figure 10 for actuations A1–A4 and as the sum of all actuations (“Total”). Relatively high and consistent emitted masses within the targeted range of 5–10 mg/actuation were observed over actuations A1–A3. Emitted mass with A4 was just below the target range (4.1 mg); however, this was viewed as acceptable as the final A4 actuation may be considered a final chamber emptying step. Variability and low emission in actuations A1–A3 were typically compensated for with higher emissions in the subsequent actuations, resulting in low overall variability on total emitted mass. The Total mean ± SD %ED was 76.7 ± 4.6%, which agreed well with the NGI findings (77.4% ± 2.1; Table 2).

OPEN IN VIEWER

FIG. 10.

Percent emitted dose (%ED) and emitted mass (milligrams) for each device actuation (A1–A4) of the D3-MSD2 device and formulation combination when delivering a 30 mg loaded mass and using one actuation per turn. Bars represent mean values evaluated over three runs (n = 3) and error bars are standard deviations. Total represents the net emission of all four actuations.

In vitro testing of lung delivery efficiency with the LC model

Table 3 provides the in vitro regional deposition fractions and estimated rabbit lung delivery efficiencies of the D3-MSD2 combination assessed using the in vitro LC model when administering a device-loaded formulation mass of 30 mg. The ED from the D3 device was 70.0%, with lower lung (filter + LC) and total lung (TB + filter + LC) delivery efficiencies of 40.1% and 56.8%, respectively, indicating a high fraction of the loaded dose was delivered to the simulated rabbit lungs. Although the lung delivery efficiency from D3 device was slightly lower than the previously reported D2-NSD2 combination, ⁴⁸ higher FPF_{<5 µm} in the NGI study suggests significant improvements in aerosol quality, which will be critical for delivering the aerosol directly to the alveolar region with the D3-MSD2 combination in the in vivo experiments.

In vivo animal descriptions and dosing

Following lung lavage, the mean PaO₂ was reduced from ∼460 to 65 mmHg and weight normalized compliance from ∼1.1 to 0.4 mL/(cmH₂O·kg) at baseline (p < 0.001) with increased PaCO₂ and disease severity (OI >15 and VEI ∼0.03) values, all indicative of severely depleted surfactant pools, elevated alveolar surface tension, and severe RDS. Baseline data following induction of lung lavage for surfactant deficiency were not different between groups for body weight, gas exchange, disease severity, and lung compliance.

All the animals in the liquid Curosurf group (historical controls ¹¹ ) experienced some degree of acute airway obstruction from the liquid, resulting in hypoxemia (SpO₂ < 85%), hemodynamic instability (e.g., bradycardia or hypotension), or increased PIP (>25 cmH₂O) by the ventilator during dosing. The D3-MSD2 aerosol and control groups experienced acute hypoxemia (SpO₂ < 85%) upon ventilator disconnection that quickly resolved when positive-pressure actuations were applied for therapy with the iDP-ADS. Following the treatment, the heart rate (p = 0.94) and blood pressure (p = 0.93) did not vary between the groups.

As described, based on an experimental mean rabbit weight of 1.59 kg, the device-loaded dose of 60 mg of the SLS-EEG MSD2 powder formulation contained 24 mg-PL/kg compared with the approximate 200 mg-PL/kg with Curosurf. ⁷⁵ The total treatment duration, which included stabilization, was ∼10 minutes to administer Curosurf. The aerosol delivery required eight actuations for 60 mg delivery of the D3-MSD2 formulation with the D3 device and could be completed in <1 minute. The total therapy delivery time, including ventilator reconnection to enable device reloading and stabilization (i.e., two 30 mg device-loaded dose delivery cycles), was completed in <5 minutes.

In vivo gas exchange

Gas exchange data for the different in vivo groups are shown in Figure 11. Following surfactant depletion, all groups had mean PaO₂ values below 70 mmHg, on FiO₂ of 1.0 at baseline, indicative of severe RDS. As expected, PaO₂ values increased from baseline in the surfactant groups immediately following therapy and remained <100 mmHg in the sham control group throughout the experiment. The mean PaO₂ at the 5-minute time point was not different between the Curosurf and control groups. In contrast, at the 5-minute time point (which included only the first 30 mg device-loaded dose), the D3-MSD2 aerosol group was significantly greater than both the sham control and Curosurf liquid groups, with a PaO₂ of ∼300 mmHg. Following the second (final) device-loaded dose of 30 mg and delivery at 15 minutes, the PaO₂ values between 30 and 150 minutes were significantly greater with aerosol than the Curosurf and sham control groups. As expected, the Curosurf group had higher PaO₂ than the sham control throughout the experiment. Among the surfactant groups, arterial oxygenation was not different at 180 and 210 minutes, and PaO₂ values recovered to 94% of the prelavage with Curosurf liquid. For the D3-MSD2 combination, PaO₂ values exceeded the basal levels at 180 minutes and ended with a final PaO₂ level that was 99% of the prewashout value at 210 minutes.

OPEN IN VIEWER

FIG. 11.

Gas exchange data are shown as mean ± standard error of mean (SEM) for control and surfactant treatment groups during the 210-minute study period. The vertical dashed line represents the time point (0 minutes) following the completion of liquid Curosurf (red circles) administration and the initial 30 mg dose of D3-MSD2 (SLS-EEG) aerosol (green circles) or sham control (blue circles). The mean PaO₂ was higher in the aerosol treatment group than Curosurf and sham groups. Following the second (final) aerosol (30 mg) and sham deliveries at 15 minutes, the PaO₂ values between 30 and 150 minutes were greater with aerosol than the Curosurf and control groups. The Curosurf group had higher PaO₂ than the controls. Among the surfactant groups, arterial oxygenation was not different at 180 and 210 minutes, and PaO₂ recovered to 94% of the prelavage values with liquid Curosurf and 99% with aerosol at 3.5 hours. The liquid Curosurf PaCO₂ values were not different from controls during the experiment. The aerosol group had statistically lower PaCO₂ than the liquid Curosurf and control groups at 30- and 60-minute time points but did not differ between groups following the remainder of the experiment. The aerosol treatment group had greater pH than controls at all time points throughout the experiment. Compared with liquid Curosurf, the aerosol group had greater pH at the 5-, 30-, and 60-minute time points but was not different at the other time points. Two-way ANOVA and post hoc Tukey’s test compared mean differences at each time point between the groups. ^*p < 0.05 for D3-MSD2 aerosol vs. liquid Curosurf. ⁺p < 0.05 for control vs. D3-MSD2 aerosol. ^#p < 0.05 for control vs. liquid Curosurf.

As expected, the mean PaCO₂ increased, and pH decreased among groups due to severe RDS induced by lung lavage (Fig. 11). The PaCO₂ values in controls remained increased throughout the experiment, and although PaCO₂ trended downward in both surfactant groups, the liquid Curosurf values were not different from sham controls during the experiment. The PaCO₂ values with the D3-MSD2 aerosol were statistically lower than liquid Curosurf and controls at 30- and 60-minute time points but did not differ from the Curosurf or sham control groups for the remainder of the experiment. The mean pH values remained between 7.25 and 7.30 in the controls and remained low but were no different from liquid Curosurf, except for the 180 and 210-minute time points, where Curosurf pH was higher than control. The aerosol treatment group had greater pH than controls at all time points throughout the experiment, beginning with the 5-minute blood gas. Compared with liquid Curosurf, the aerosol group had greater pH at the 5-, 30-, and 60-minute time points but was not statistically different at the remaining time points (see Fig. 11), despite having a higher trend at all time points.

In vivo lung compliance and disease severity

Lung compliance and indices of disease severity data for the different groups are shown in Figure 12. The lung compliance improved rapidly in the aerosol group at 5 minutes following the initial 30 mg device-loaded dose and continued to improve following the second 30 mg aerosol delivery at 15 minutes. The D3-MSD2 aerosol group had greater compliance than the liquid Curosurf and control groups from the 5-minute time point until the end of the experiments. Following a small initial increase between 5 and 30 minutes, the compliance values stabilized and then slightly decreased with liquid Curosurf and were not statistically different from sham controls at any of the time points over the 3.5-hour observation period. At the 3.5-hour final time, the compliance in the liquid Curosurf and sham control groups was not improved (0.4 vs. 0.3 mL/[cmH₂O·kg], respectively) from baseline. In contrast, the D3-MSD2 aerosol group compliance nearly doubled from baseline (0.46–0.91 mL/[cmH₂O·kg]) and recovered to 91% of the prewashout value. Furthermore, the D3-MSD2 aerosol group demonstrated an increasing trend in compliance through the final hour of the experiment that continued to the 3.5-hour final time point.

OPEN IN VIEWER

FIG. 12.

Lung compliance, oxygenation index (OI), and ventilation efficiency index (VEI) data are shown as mean ± SEM for control and surfactant treatment groups during the 210-minute study period. The vertical dashed line represents the time point (0 minutes) following the completion of liquid Curosurf (red circles) administration and the initial 30 mg dose of D3-MSD2 (SLS-EEG) aerosol (green circles) or sham control (blue circles). The lung compliance improved rapidly in the aerosol group 5 minutes following the initial 30 mg surfactant dose and continued to improve following the 30 mg aerosol delivery at 15 minutes. The aerosol group had greater compliance than the liquid Curosurf and control groups from the 5-minute time point until the end of the experiments. At the 3.5-hour time, the compliance in the liquid Curosurf and sham control groups was not improved from baseline, but surfactant D3-MSD2 compliance approximately doubled compared with washout baseline and recovered to 91% of the prewashout value. The VEI was higher in the D3-MSD2 aerosol surfactant group than the controls at 30 minutes, and both surfactant treatment groups had higher VEI than the controls from 60 minutes until the 3.5-hour observation period. Two-way ANOVA and post hoc Tukey’s test compared mean differences at each time point between the groups. ^*p < 0.05 for D3-MSD2 aerosol vs. liquid Curosurf. ⁺p < 0.05 for control vs D3-MSD2 aerosol. ^#p < 0.05 for control vs liquid Curosurf.

Surfactant washout produced OI >15 for all cases, consistent with severe respiratory disease. As expected, improvement was not observed for the sham control case over the 3.5-hour observation period, and OI values were higher than the surfactant-treated groups at all time points. In contrast, the surfactant aerosol group showed a rapid reduction in OI (<2) with 80%–91% recovery from baseline (prelavage) that was maintained from the 30-minute time point until the end of the 3.5-hour period. The liquid Curosurf group showed a favorable reduction in OI following treatment with a more gradual recovery, ranging from 28% to 65% between 30 and 180 minutes and 75% at 3.5 hours.

Considering VEI, liquid Curosurf demonstrated an improved trend compared with the sham control throughout the observation period that did not rise to the level of statistical significance. In contrast, the D3-MSD2 aerosol group demonstrated a statically significant improvement versus the sham control at 30 minutes, which continued through the 3.5-hour experiment. Furthermore, the D3-MSD2 aerosol group demonstrated statistically significant improvement versus liquid Curosurf at 60 minutes, which also continued through the remainder of the 3.5-hour observation period.