Inhalable aerosol microparticles with low carrier dosage and high fine particle fraction prepared by spray-freeze-drying

Abstract

Co-suspension drug-loading technology, namely Aerosphere™, can improve fine particle fraction (FPF) and delivered dose content uniformity (DDCU). However, because of its poor drug-loading efficacy, the phospholipid carrier dosage in Aerosphere™ is usually dozens of times greater than that of the drug, resulting in a high material cost and blockage of the actuator. In this study, spray-freeze-drying (SFD) technology was used to prepare inhalable distearoylphosphatidylcholine (DSPC)-based microparticles for pressurized metered-dose inhalers (pMDI). Water-soluble, low-dose formoterol fumarate was used as an indicator to evaluate the aerodynamic performance of the inhalable microparticles. Water-insoluble, high-dose mometasone furoate was used to investigate the effects of drug morphology and drug-loading mode on the drug delivery efficiency of the microparticles. The results demonstrated that DSPC-based microparticles prepared using the co-SFD technology not only achieved higher FPF and more consistent delivered dose than those of drug crystal-only pMDI, but the amount of DSPC was also reduced to approximately 4% of that prepared using the co-suspension technology. This SFD technology may also be used to improve the drug delivery efficiency of other water-insoluble and high-dose drugs.

Graphical abstract

1. Introduction

Most drugs have lower solubility in the propellant hydrofluoroalkane (HFA) than in chlorofluorocarbon (CFC). Pressurized metered-dose inhaler (pMDI) products are mostly suspensions (Sheth et al., 2017), with CFC gradually being replaced by HFA. Traditional suspension pMDIs are affected by interparticle forces, and many of the HFA pMDIs that use microcrystal drugs have poor colloidal stability, which results in rapid creaming or sedimentation. This can lead to dosing variability, because the dose is metered by the volume of suspension and is irrespective of the mass of suspended particles (Cummings, 1999). Dosing uniformity is confounded by the improper timing of actuation when using the pMDIs. Even properly trained patients may experience poor dose reproducibility, particularly with poorly performing formulations (Bell and Newman, 2007; Dolovich, 2000). In addition, there is room for improvement in the amount of drugs that can effectively deposit in the lungs. Most drugs that do not enter the lungs enter the digestive tract, which may lead to undesirable adverse reactions.

A common technical challenge of conventional pMDIs containing micronized drugs is poor drug delivery uniformity and low deposition efficiency in the lungs. The current commercial technology Aerosphere™ disperses porous phospholipid microparticles and micronized drug crystals in HFA propellant to form a homogenous and stable suspension that can achieve consistent dose delivery (Ferguson et al., 2018). The porous phospholipid microparticles are used as carriers and are prepared using PulmoSphere® technology, where the excipient distearoylphosphatidylcholine (DSPC) is used as a surfactant, and perflubron is dispersed in water to form an oil/water emulsion. The oil phase then volatilizes to form pores during spray drying and reduces the density of inhalable microparticles (Chow et al., 2007). Currently, this co-suspension technology has been used in two commercial products, Bevespi Aerosphere® and Breztri Aerosphere® (Israel et al., 2020). The use of DSPC in the Aerosphere™ technology has efficiently solved two performance problems of the traditional pMDIs: it has greatly prolonged the time to maintain the uniformity of the suspension and dispersion after shaking, and improved the accuracy of the delivered dose and the efficiency of drug deposition in the lung (Tamura et al., 2021; Ferguson et al., 2018).

However, because of the low adsorption efficiency of micronized drugs and microparticle carriers, a number of blank inhalable carriers are needed to achieve the desired performance. The mass concentrations of inhalable microparticles are typically greater than 10 times those of micronized drugs (Vehring et al., 2012) (US8324266). The high substance concentration in the HFA suspension increases the probability of actuator orifice clogging, thus increasing the risk of poor therapeutic effect. Moreover, the phospholipid excipient, DSPC, is very expensive and can cost even more than some drugs. The bulk use of DSPC will inevitably increase the manufacturing cost of the co-suspended pMDI and the medical economic burden. As the solvent evaporates during spray drying, the particle collapses and shrinks, reducing the interior porosity of the particles during spray drying (Dolovich, 2000; Heyder et al., 1986), so organic solvents are needed as pore-forming agents to prepare the low-density microparticles during spray drying. However, using organic solvents in the process increases the cost and the risk of environmental pollution (Li and Zhang, 2012).

Therefore, we aim to improve the co-suspension technology by reducing the number of carrier excipients while maintaining physical stability of the suspension and achieving high drug deposition efficiency in the lung. As most studies on inhalable microparticles for pMDI have focused on the co-suspension technology, improvements and innovations in drug-loading modes will improve the inadequacy of co-suspension technology.

In this study, inhalable phospholipid microparticles for pMDI were prepared by the spray-freeze-drying (SFD) process to achieve high aerodynamic performance and suspension stability. This process allows atomized droplets to be frozen to solid microparticles first to maintain the initial morphology, before the solvent sublimes to form pores during the drying process (Niwa et al., 2009; Shahin et al., 2021). The procedure provides a theoretical possibility for preparing low-density microparticles without organic solvent.

Considering the representability of the model drugs, we selected the classic clinical combination of inhalable corticosteroids and β-receptor agonists (O'Byrne et al., 2021). Mometasone furoate (MF) and formoterol fumarate (FF) were used as the model drugs. It is worth noting that MF has low solubility in water. It is difficult to improve the morphology of the MF-microparticles by spray drying. The dose of MF is relatively high, usually hundreds of micrograms per actuation, which is dozens of times greater than that of FF (Tattersfield, 2004). Therefore, achieving efficient drug loading and delivery of MF are of significant challenges.

In this study, we investigated the influence of different preparation methods on the structures of low-density microparticles in the SFD process, as well as the relationship between microparticle structure and aerosol performance, before obtaining the carrier microparticles with a high fine particle fraction (FPF) and the best suspension stability. Further study on the drug-loading mode revealed the cause for the low drug-loading efficiency of the co-suspension technology. Based on these insights, we developed a novel method to improve the drug-loading efficiency and achieve a high FPF for high-dose water-insoluble drugs on the premise of low carrier dosage. This solved the technical problem that limits the application of inhalable microparticles by co-SFD in pMDIs.

2. Materials and methods

2.1. Experimental group design

The formulations in Table 1 (F1–F14) were carefully designed to produce microparticles that can achieve a high FPF and good physical and kinetic stability. Inhalable microparticles were the intermediates used to prepare pressurized metered-dose inhalers (pMDI). The key factors included the solid concentration and the ratio of an oil phase and water in the emulsion. The fewer non-volatile substances contained in the same volume, the lower the density of the microparticles and the smaller the aerodynamic particle sizes are (e.g. F1–F4). When the feedstock was prepared into emulsion form, the oil phase dispersed in the emulsion evaporates, leaving pores in microparticles (e.g. F4–F7). “P1–P17” indicated the formulations of pMDIs in Table 2, which were comprised of inhalable microparticles (called “F”), HFA propellant, and devices components such as the aluminum can, metering valve, and actuator.

Table 1.

Formulation parameters for inhalable microparticles.

Microparticle Formulation	Solid Concentration (% w/w)	Oil Phase (% V/V)	Solid State of MF	DSPC:MF (W/W)
F1	0.80	0	N/A	N/A
F2	1.40	0	N/A	N/A
F3	1.93	0	N/A	N/A
F4	2.68	0	N/A	N/A
F5	2.68	20	N/A	N/A
F6	2.68	30	N/A	N/A
F7	2.68	40	N/A	N/A
F8	2.68	0	Crystal	20:1
F9	2.68	0	Crystal	10:1
F10	2.68	0	Crystal	5:1
F11	2.68	0	Crystal	3:1
F12	2.68	0	Amorphous	3:1
F13	2.68	0	Amorphous	10:7
F14	2.68	0	Amorphous	1:1

Table 2.

Formulation parameters for pMDI formulation.

pMDI Formulation	Microparticle	Microparticle:MF (W/W)	MF Loading Method
P1	F2	N/A	N/A
P2	F3	N/A	N/A
P3	F4	N/A	N/A
P4	F5	N/A	N/A
P5	F6	N/A	N/A
P6	F7	N/A	N/A
P7	F4	5:1	Co-Suspension
P8	F4	10:1	Co-Suspension
P9	F4	20:1	Co-Suspension
P10	F4	30:1	Co-Suspension
P11	F8	N/A	Embedding
P12	F9	N/A	Embedding
P13	F10	N/A	Embedding
P14	F11	N/A	Embedding
P15	F12	N/A	Embedding
P16	F13	N/A	Embedding
P17	F14	N/A	Embedding

It is necessary to find the shortcomings of the co-suspension technology, the rationale of which is utilizing the polarity of phospholipid particles. In the co-suspension formulation, the polarity of the phospholipid particles is similar to that of the HFA, which reduces the tendency of aggregation of the particles. Moreover, the drug crystals associated with phospholipid particles (e.g. F4) spontaneously and formed stable suspensions in propellant HFA (Doty et al., 2017). The formulations of co-suspension were marked as “P7–P10”. MF was dispersed as fine particles in the emulsion feedstock in the embedding drug-loading mode. The MF particles were co-spray-freeze-dried with phospholipid (e.g. DSPC), and such formulations were marked as “F8–F14”. In addition to the co-suspension formulation (P7–P10), the aerodynamic characteristics of all of the inhalable microparticles (F2–F14) could be evaluated only after being transformed into pMDI (P1–P6 and P11–P17); the corresponding relationship was shown in Table 2. We compared the aerodynamic performance in vitro with the brand product Dulera® (Batch No. U006804, Merck, Kenilworth, NJ, USA) made of micronized MF (Batch No. KS-200801, Hunan, China) and FF (Batch No. FF/101/17–18, Guangzhou, China).

2.2. Spray-freeze-drying (SFD)

2.2.1. Preparation of amorphous MF

The feedstock was prepared by dissolving MF in dioxane (Sinopharm, Beijing, China) aqueous solution, followed by stirring to form a solution. Then, the solution was delivered to an atomization nozzle (5 mm, 0.5 mm, Glatt GmbH, Weimar, Germany) with a YZ2515X-A peristaltic pump, which sprayed tiny droplets into the liquid nitrogen. After spraying, the microparticles in the liquid nitrogen were moved to the Virtis Advantage EL-85 Benchtop Freezer (ALT Technologies, East Lyme, USA) and then dried for 36 h.

2.2.2. Preparation of inhalable FF and MF-embedded phospholipid microparticles

CaCl₂ (Batch No. P1902809, Shanghai, China)and FF were dissolved in water and sonicated to obtain a clear solution. Subsequently, distearoylphosphatidylcholine (DSPC, Batch No. 20210501, Shanghai, China) was added to the solution and stirred to disperse it uniformly, before adding octamethylcyclotetrasiloxane (D₄, Alfa, Zhengzhou, China) if required. Finally, micronized crystal or amorphous MF was added, stirred with high-shear stirring (Fluko Equipment, Shanghai, China) for 2 min, and homogenized with high-pressure homogenization (Nano DeBee, BEE International, South Easton, MA, USA) five times at 170 MPa to obtain the feedstock. The feedstock was delivered to the atomization nozzle by a peristaltic pump, which sprayed tiny droplets into liquid nitrogen. After spraying, the microparticles in the liquid nitrogen were moved to a freeze dryer and dried for 36 h (Table 1 (F1–F14)).

*MF: Mometasone furoate.

*FF: Formoterol fumarate.

2.3. Microparticle characterization

2.3.1. Particle size distribution

The inhalable microparticles were measured by laser diffraction using the HELOS (H3838) & RODOS/T4 laser diffractions (R2 lens [0.25/0.45–87.5 μm]; Sympatec GmbH, Clausthal-Zellerfeld, Germany). Size distributions of the samples were determined under a 107-mbar vacuum using 3 bar of dispersion pressure. The measurements were conducted in triplicate (n = 3). The particle size data were expressed as D₁₀, D₅₀, and D₉₀, which represent the equivalent spherical volume diameters at 10%, 50%, and 90% cumulative volume, respectively.

2.3.2. Bulk density determination

Different inhalable microparticles were individually filled in glass bottles with a volumetric scale. The microparticle mass was recorded at 0.5 mL and used to calculate ρ_bulk as the ratio of sample mass and volume.

2.3.3. Morphology assessment using scanning electron microscopy (SEM)

The morphology of the SFD microparticles was observed with the Zeiss Gemini 300 Field Emission Scanning Electron Microscope (Carl Zeiss AG, Jena, Germany) at 3 kV. The microparticles were sprinkled onto a carbon stick tape that was mounted on scanning electron microscopy (SEM) stubs. Excess microparticles were removed by blowing with clean compressed air. Gold was sprayed in the ion sputtering apparatus, and then observed by SEM.

2.3.4. Differential scanning calorimetry (DSC)

Differential scanning calorimetry (DSC) (Waters, Milford, MA, USA) was used to depict the thermal response profiles of SFD microparticles as well as the raw materials. Approximately 1 mg of powders was weighed and loaded into an aluminum crucible and was heated from 30 °C to 300 °C at a constant rate of 10 °C/min.

2.3.5. Powder X-ray diffraction (PXRD)

The polymorphism of SFD microparticles and raw materials was investigated using the Bruker D8 Advance X-Ray Diffractometer (Bruker, Karlsruche, Germany). The tube pressure was 40 KV, the tube current was 40 mA, Cu Kα target ray scanning was used, the divergence slit was 0.6 mm, the sora slit was 4.0°, the conversion range 2θ was 2°–40°, the conversion step was 0.02°, and the speed was 8°/min.

2.4. Preparation of pMDI

The micronized MF (if present) and different phospholipid microparticles prepared by SFD were added individually to 14-mL aluminum cans (Anomatic, Suzhou, China). The cans were crimped using a crimper (PamasolWilliMäder AG, Pfäffikon, Switzerland) after the addition of DF 316 valves (Aptar, Jiangsu, China). Each pMDI (Table 2 (P1–P17)) was filled with 1,1,1,2-Tetrafluoroethane (HFA134a, Suzhou, China) using a pressure filler (Pamasol WilliMäder AG, Pfäffikon, Switzerland).

*MF: Mometasone furoate.

*FF: Formoterol fumarate.

2.5. High performance liquid chromatography (HPLC)

The amount of MF and FF was quantified using HPLC (Ultimate 3000; Thermo Fisher Scientific, Waltham, MA, USA). Briefly, a C18 column (4.6 mm × 300 mm, 5 μm; Waters) was used and the mobile phase composed of 85% methanol and 15% aqueous phosphoric acid (pH 3.5), with a column temperature of 40 °C, detection wavelength of 225 nm, flow rate of 1.5 mL/min, and injection volume of 20 μL. The FF concentration (C_FF) was linear over a range of 0.02–0.4 μg/mL, and the MF concentration (C_MF) was in the range of 0.05–20 μg/mL, showing good linearity with peak area A. The analytical method met the requirement for quantitative analyses.

2.6. pMDI characterization

2.6.1. Physical stability of microparticle suspensions in HFA

Different phospholipid microparticles prepared by SFD were added individually to 30 mL glass bottles (Bormioli Rocco, Province of Parma, Italy). The bottles were crimped after the addition of DF 316 valves. Each sample was prepared by filling with HFA134a. After shaking the bottle, photographs were taken at 0 min and 30 min to record the suspended dispersion status of each sample.

2.6.2. Aerosol performance evaluation

The aerosol performance of the different pMDIs was evaluated with the Next Generation Impactor (NGI; Copley, Nottingham, UK). Different volumes of methanol 0.2% glacial acetic acid (50/50, V/V) were used for dilution to allow a quantifiable drug concentration: 25 mL for the device, adaptor, Micro-Orifice Collector, and NGI Stage 3–5; 50 mL for the induction port (throat); 10 mL for NGI Stage 1, 2, 6, and 7. The sample solutions were analyzed by HPLC. Each formulation was evaluated in triplicate. Fine particle dose (FPD) was the mass of particles with an aerodynamic diameter of less than 5.0 μm as calculated with the assayed MF and FF obtained from HPLC. FPF was defined as the percentage fraction of FPD to the delivered dose. The FPF was calculated by Copley Inhaler Testing Data Analysis Software Version 3.10.

2.6.3. Evaluation of the delivered dose using the variable shaking techniques

In order to evaluate the impact of patient's shaking techniques on delivered dose content uniformity, three shaking modes were employed: mode 1, a standard control shake (20 shakes in 5 s); mode 2, standing for 30 s after standard control shake; and mode 3, a gentle shake (5 shakes in 5 s). The percentage of dose per actuation was calculated using the theoretical content of drug per actuation in the pMDI as a reference.

3. Results

3.1. Evaluation of inhalable microparticle characteristics for structural construction studies

To obtain high FPF and good suspension stability inhalation microparticles, we evaluated the morphology, suspension state, bulk density, and particle size distribution (Fig. 1).

Fig. 1.

SEM (scanning electron microscopy) micrographs of SFD (spray-freeze-drying) microparticles (F1–F7) and the suspension state of pMDIs; F1, F2 and F3 scale bar = 5 μm; F4, F5, F6 and F7 scale bar = 1 μm; A: FPFs of pMDIs (P1–P6).

3.1.1. Morphology

The morphology of the SFD FF microparticles was observed using SEM. All microparticles showed a porous structure (Fig. 1). The solid concentration decreased from 2.68% to 1.4% (Fig. 1(F2–F4)), whereas the surface of the microparticles gradually changed from a stripe-like to flake-like shape, suggesting that the solid concentration of 1.4% was the minimum concentration capable of forming microparticles. When the solid concentration was 0.8% (Fig. 1(F1)), only debris existed in the field of view, and it is difficult to maintain the microparticle morphology under this condition. As the proportion of the oil phase increased (Fig. 1(F4–F7)), the roundness of the microparticles improved. Because the phospholipids were enriched in the continuous phase of emulsions, microparticles formed with higher mechanical strength.

In the co-suspension drug-loading mode study, the surface and porous texture of the inhalable microparticles (Fig. 2 (P7 and P10)) after loading micronized MF crystals became unclear, suggesting that the drug crystals attached to the surface of the microparticles. As the proportion of inhalable microparticles increased, some of the microparticle surfaces still exhibited a porous texture (Fig. 2 (P10)), suggesting that part of the surface was not covered with drug crystals.

Fig. 2.

SEM micrographs of spray-freeze-drying microparticles (F4) and microparticles in co-suspension with pMDI (P7, P10).

F4, P7 scale bar = 1 μm; P10 scale bar = 3 μm.

The embedding drug-loading mode study showed that the microparticles began to exhibit irregular morphology (Fig. 3(F11)). Moreover, when the micronized MF crystals and DSPC were spray-freeze-dried together, drug crystal chimerism was observed in the phospholipid microparticles. The SFD MF (Fig. 3(B)) displayed agglomerate morphology with smooth surfaces indicating that the MF was amorphous and could transform into cuboid-like crystals rapidly in water (Fig. 3(C)). When the inhalable microparticles were prepared using amorphous MF and DSPC as the raw material, cuboid-shaped crystals appeared in the microparticles (Fig. 3(F14)), indicating that crystal transformation occurred during the preparation.

Fig. 3.

SEM micrographs. A: Micronized MF; B: SFD MF; C: Crystal MF F11 and F14: Microparticles A and F11 scale bar = 2 μm; B and F14 scale bar = 1 μm; C scale bar = 5 μm.

3.1.2. Physical stability of microparticle suspensions in HFA134a

Observing the dispersed and suspended state at different times after shaking the bottle (Fig. 1), the microparticle suspensions floated in HFA 134a at solid concentrations of less than 2.68%. The density of microparticles prepared from the feedstock with a solid concentration of 2.68% was closer to the density of HFA 134a, which remained in suspension after standing for 30 min. The inhalable microparticles prepared by different oil phase ratios also showed distinct differences in HFA134a. As increasing oil phase ratios, the microparticles floated gradually, suggesting that the oil phase also decreased the particle density.

3.1.3. Bulk density

The bulk density of all SFD microparticles (Fig. 1(F1–F7)) was less than 30 mg/mL. The bulk density decreased with decreasing solid concentration and increasing oil phase proportion. The bulk density of the F2 particle with a solid concentration of 1.4% and the F7 particle with an oil phase ratio of 40% were less than 10 mg/mL, indicating that both methods could effectively reduce the bulk density.

3.1.4. Particle size distribution measurement by laser diffraction

The size distribution of the SFD particles was measured by laser diffractometry (Table 3). When dispersed at 3 bar, the D₁₀ varied from 0.44 μm to 0.96 μm, the D₅₀ varied from 1.72 μm to 3.43 μm, and the D₉₀ varied from 4.18 μm to 10.03 μm across all formulations (measured by Sympatec). Although the sizes of the inhalable microparticle fluctuated slightly for the different formulations, all inhalable formulations are in the geometrical diameter range, which is suitable for respiratory administration.

Table 3.

Particle size distribution of formulations.

Microparticle Formulation	D10	D50	D90
F1	0.82 (0.02)	2.23 (0.01)	5.39 (0.13)
F2	0.69 (0.01)	2.87 (0.06)	8.97 (0.42)
F3	0.64 (0.04)	1.83 (0.03)	4.51 (0.51)
F4	0.83 (0.01)	3.43 (0.06)	10.03 (0.10)
F5	0.77 (0.01)	2.62 (0.04)	8.00 (0.36)
F6	0.91 (0.01)	2.80 (0.09)	7.44 (0.42)
F7	0.96 (0.02)	2.66 (0.06)	7.40 (0.28)
F8	0.78 (0.01)	3.26 (0.06)	9.77 (0.26)
F9	0.67 (0.01)	3.11 (0.07)	9.76 (0.38)
F10	0.63 (0.01)	2.60 (0.01)	8.34 (0.16)
F11	0.44 (0.01)	2.09 (0.02)	7.23 (0.01)
F12	0.60 (0.01)	2.29 (0.04)	7.59 (0.23)
F13	0.66 (0.01)	2.20 (0.05)	5.92 (0.12)
F14	0.58 (0.01)	1.72 (0.05)	4.18 (0.18)

*MF: Mometasone furoate.

*FF: Formoterol fumarate.

3.2. Aerosol performance evaluation

During the construction of the inhalable microparticle structures, the aerodynamic performance showed a regular alteration following the increase in solid concentration from 1.4% to 2.68% (Table 1(F2, F3, and F4)). The FPF of the corresponding pMDIs (Fig. 1(P1, P2, and P3)) showed a poor increasing trend. As the oil phase increased from 0% to 40% (Table 1(F4, F5, F6, and F7)), the FPF of corresponding pMDIs (Fig. 1(P3, P4, P5, and P6)) showed a decreasing trend, which illustrated that the use of the oil phase as a pore-former in the SFD process to prepare low-density microparticles did not improve the FPF as expected. The formulation (F4) with a solid concentration of 2.68% and an oil phase of 0% had the best aerodynamic performance among the available particles. In the co-suspension drug-loading mode study, the F4 microparticle was used as the carrier. With the carrier microparticle amount increased, the FPF of FF showed no single direction change trend and was approximately 44% (Fig. 4 (Group A, P7, P8, P9, and P10)). However, the FPFs of MF were31.48%, 34.24%, 36.93%, and 41.94%, showing an increasing trend, indicating that increasing the number of carrier microparticles could increase the FPF of MF. In the embedded drug-loading mode study, micronized MF crystals were used as raw material for the preparation of inhalable microparticles. With decreasing amounts of phospholipids (Table 4 (F8, F9, F10, and F11)), the FPF of FF was in the range of 46.51%–48.29% (Fig. 4 (Group B, P11, P12, P13, and P14)) and the FPF of MF was in the range of 42.93%–43.31%. The FPFs of the two drugs were stable, and the correlation between the amount of DSPC and FPF disappeared. However, MF had a lower FPF than FF, and the ΔFPF fluctuated in the range of 3.25%–4.98%, indicating that there was room for improvement in MF-loaded delivery efficiency. When the raw material of crystal MF was changed to the amorphous form (Table 4 (F12, F13, and F14)), the amount of DSPC could be further reduced, and the FPFs of three pMDIs (Fig. 4 (Group C, P15, P16, and P17)) was still similar. Compared to micronized crystal MF as raw material, the FPF of MF increased by about 3%, and the ΔFPF between the two drugs was further decreased, fluctuating in the 0%–2% range.

Fig. 4.

FPFs of different pMDIs.

A: Co-suspension pMDIs;

B: Embedded microparticles with crystal MF as raw material pMDIs;

C: Embedded microparticles with amorphous MF as raw material pMDIs.

Table 4.

FPFs of pMDIs.

pMDI Formulation	FPF%
	FF	MF
P1	44.62 ± 3.96	N/A
P2	46.76 ± 5.20	N/A
P3	47.63 ± 0.50	N/A
P4	44.97 ± 2.50	N/A
P5	39.51 ± 4.93	N/A
P6	35.22 ± 3.22	N/A
P7	42.68 ± 1.99	31.48 ± 0.72
P8	44.97 ± 0.84	34.24 ± 1.32
P9	44.18 ± 0.44	36.93 ± 1.12
P10	45.26 ± 1.39	41.94 ± 2.30
P11	46.51 ± 2.27	43.26 ± 0.86
P12	47.33 ± 3.78	43.14 ± 1.87
P13	47.74 ± 3.29	42.93 ± 1.98
P14	48.29 ± 1.08	43.31 ± 1.73
P15	48.30 ± 4.51	46.53 ± 4.44
P16	47.09 ± 2.02	46.14 ± 3.29
P17	49.10 ± 1.20	48.85 ± 1.37

*MF: Mometasone furoate.

*FF: Formoterol fumarate.

As shown in Fig. 5, pMDI P17, with the lowest amount of DSPC in this study, had a higher FPF than the brand product Dulera®, which was prepared by micronized drugs as inhalable microparticles. Compared to Dulera®, P17 prepared using the crystal transformation embedded method had better aerosol performance, indicating that the drug particles were more easily deposited in the depths of the lungs.

Fig. 5.

FPFs of Dulera® and P17 (n = 3).

3.3. Evaluation of delivered dose using three manual shaking maneuvers

As shown in Fig. 6, with a standard control shake (mode 1), the delivered doses from P17 and Dulera® were all close to the theoretical doses and had good dose uniformity. However, in the delayed actuation mode after shaking (mode 2) and the gentle shake (mode 3), the delivered dose content uniformity (DDCU) of the brand product Dulera® became worse. The delivered doses for MF ranged from 85.8% to 127.1% of the targeted dose under mode 2, and from 93.4% to 127.2% under mode 3; and the delivered doses for FF ranged from 47.5% to 99.1% under mode 2, and from 109.7% to 158.4% under mode 3. Some doses of Dulera® even exceeded the acceptable range mentioned in the drug label, whereas the delivered dose content uniformity and accuracy of P17 still showed excellent performance. Comparatively, the dose delivery for the drug crystal-only Dulera® showed increased variability.

Fig. 6.

DDCU results for Dulera® and P17 under different shaking modes.

3.4. Solid state characterization

DSC and PXRD were used to study the solid state of SFD microparticles (Fig. S9). Before the SFD process, the micronized MF was in crystal form as characteristic peaks at 9.24°, 12.99°, and15.97° (Chen et al., 2005) were detected. The micronized FF was in crystal form as characteristic peaks at 5.53°, 9.22°, 15.10°, and 15.87° were detected. DSPC was also in crystal form as characteristic peaks at 5.93°, 20.02°, 22.27°, and 22.94°were detected.

After the SFD process, the characteristic peaks of raw materials in six formulations related to the construction of microparticle structures (Table 1 (F2–F7)) disappeared, and the characteristic peaks of DSPC disappeared, indicating that DSPC transformed crystal into an amorphous state after SFD process. The characteristic peaks of FF also disappeared, possibly because its proportion in the phospholipid microparticles was too low (6.6%), leading to coverage of the characteristic peaks, so its crystal form could not be detected. Regarding embedded MF (Table 1 (F8–F14)), F8 and F9 did not show the characteristic peaks of MF, when the concentration of MF in phospholipid microparticles was low. With the increasing concentration of micronized MF, the characteristic peaks of MF appeared gradually (Fig. S9 (F10–F14)). The sample made of MF solution by SFD did not show the characteristic peaks of MF, indicating that amorphous MF was prepared. Using amorphous MF as raw material to prepare embedded inhalable phospholipid microparticles (Fig. S9 (F12–F14)), all samples had characteristic peaks. For example, the characteristic peaks of MF in F13 at 9.21°, 12.96°, and 15.95° were detected, consistent with the micronized MF, indicating that the amorphous MF underwent crystal transformation during the preparation process.

In Fig. S10, there were two overlapping endothermic peaks (melting point) of MF at about 228.83 °C and 234.68 °C, and an exothermic peak at about 239.07 °C. Complex thermal changes near the melting point indicated that the degradation reaction occurred during the melting process of MF (Chen et al., 2005). The overlapping endothermic peaks of FF were about 116.11 °C and 133.55 °C (Jarring et al., 2006). The endothermic peaks of DSPC were about 127.97 °C and 182.51 °C, whereas the endothermic peak of blank phospholipid microparticles after SFD was about 93.71 °C, and a wider endothermic peak was about 239.31 °C, indicating that the SFD process changed the molecular arrangement of DSPC and changed the thermal phenomenon. An exothermic peak of SFD MF was about 170 °C, which was caused by amorphous crystallization, and an endothermic peak was about 224 °C, which was close to the position of the melting endothermic peak of micronized MF. There were two endothermic peaks in the six formulations (Table 1 (F2–F7)) related to the construction of microparticles' structure and in the four formulations (Table 1 (F8–F11)) related to the microparticles embedded with crystal MF, one of which was about 91 °C (87.58 °C–93.51 °C) and the other was about 237 °C (234.88 °C–241.28 °C). These two peaks are consistent with the positions of the endothermic peaks of the blank phospholipid microparticles after SFD, which was caused by the thermal phenomenon of the DSPC matrix. The endothermic and exothermic peaks of MF crystals and the endothermic peak of FF in the six formulations did not appear, possibly due to the small proportion of drugs in the microparticles. The position of the endothermic peaks of the microparticles made of amorphous MF (Table 1 (F12–F14)) became irregular. It is possible that the proportion of MF in the microparticles increased and gradually exhibited its associated thermal phenomenon. The intensity of endothermic or exothermic peaks of MF and DSPC varied with the concentration ratio and showed overlapping peaks with different final intensities.

4. Discussion

The aerodynamic performance of inhalable aerosol microparticles affects the amount of drug that can deposit in the lung, and the aerodynamic particle size is an important indicator in predicting drug deposition. The calculation formula is as follows: (Pilcer and Amighi, 2010):

$$d_{\mathit{ae}}=d_{\mathit{geo}}\sqrt{\frac{{\rho }_p}{{\rho }_0\bullet {\rm X}}}$$ (1)

Where d_ae is the particle aerodynamic diameter, d_geo is the geometric diameter, ρ_p is the density of drug particles, ρ₀ is the unit density, and Χ is the shape factor. According to the aerodynamic particle size evaluation formula, in addition to reducing the geometric particle size of inhalable particles, lower density can also reduce the aerodynamic particle size (Dolovich, 2000; Heyder et al., 1986). For pMDI, the density of the inhalable microparticles also affected its suspension stability in the propellant HFA. Therefore, controlling microparticles' density is a feasible method to achieving the technical goal.

4.1. Effects of different methods of reducing density on microparticle structure

The co-suspension technology allows the oil phase to evaporate during spray drying to form pores of the microparticles and thus reduce their density. While during SFD, the atomized droplets were first frozen to prevent the collapse and contraction of microparticles, and then the solvent sublimes to increase the porosity of the microparticles and reduce their density. To conduct in-depth research on controlling the above two methods of adjusting density, the model drugs should have minimum impact on the structure of microparticles. Due to the low dose of FF, its proportion in the microparticles was only about 6.6% (w/w). In addition, FF could be dissolved in the feedstock and evenly dispersed in the microparticle matrix. Therefore, FF would have little impact on the microparticle structure, and the Lip/FF microparticles were expected to have similar aerodynamic characteristics as that of blank inhalable phospholipid microparticles.

The bulk density of microparticles decreased significantly with the decrease in solid concentration. As the oil phase ratio increased, the bulk density also decreased. However, considering the physical stability of a suspension, F3 (bulk density: 21.5 mg/cm³) and F5 (bulk density: 21.9 mg/cm³) had a similar bulk density but had a considerable difference in suspension stability. Standing for 30 min after standard shake, F3 floated well in HFA 134a, whereas F5 only floated slightly in HFA 134a. F2 (bulk density: 8.78 mg/cm³) and F7 (bulk density: 7.21 mg/cm³) showed similar differences in suspension stability, and F2 floated significantly more. The bulk density is related to the microparticle porosity (Liao et al., 2019), and the pores include both the external pores and internal pores of microparticles. The buoyancy of microparticles in HFA 134a depended on the volume of HFA displaced. The closed pores inside the microparticles can generate buoyancy, and the external pores cannot generate buoyancy effectively because the HFA would enter them. Based on the results of bulk density and suspension stability, it can inferr that the pore structure obtained by water evaporation was different from that obtained by oil phase evaporation in the emulsion. The latter had more pores connected with the outside and rose more slowly in HFA 134a. The results of electron microscopy (Fig. 1) showed that there were more pores on the surface of the microparticles caused by the oil phase and that the particles with solvent-induced pores had a flake structure, as well as relatively few surface and external pores. The aerodynamic characteristics of the corresponding pMDIs (Fig. 1) provide further evidence for the change in density. As the proportion of the oil phase increased, the FPF decreased gradually because the HFA in pMDIs entered further into the external pores, which increased the microparticle density in the initial stage after actuation and reduced the FPF. These results further showed that the production of using the oil phase to cause pores in the spray thermal drying process could not achieve the expected objective in the SFD process.

The microparticles (F4) obtained by the method of adjusting microparticle density by solid concentration had not only a high FPF but also the best suspension stability, which was expected to achieve consistent delivered dose.

4.2. High efficiency drug-loading and delivery method

In the co-suspension drug-loading mode, drug-loading depended on the adsorption strength of the carrier microparticles and micronized drugs. For specific carrier excipients, the adsorption strength is related to the drug's molecular polarity. With HFA as a continuous phase in pMDI, the closer that the polarity of the drug particles is to that of the phospholipid particles and the further the polarity of drug particles is from that of HFA, the stronger the force of adsorption between the two microparticles (Doty et al., 2017). This means that different drug molecules may have different adsorption efficiencies with DSPC. MF was adsorbed on the surface of carrier microparticles (Fig. 3) so that the upper limit of the amount of MF absorbed by the carrier microparticles was limited to the surface areas of the carrier microparticles. As shown in the SEM image (Fig. 3 (P10)), the drug could not cover the surface of carrier microparticles completely, suggesting that the adsorption efficiency between the carriers and MF was low in co-suspension mode. Therefore, part of the MF must not have been loaded on the phospholipid microparticles, and the aerosol performance of free micronized MF might be different from that of the loaded MF. Previous studies on co-suspension have adsorbed multiple drugs on the surface of carrier microparticles simultaneously, which makes it difficult to show the difference in aerodynamic performance between the micronized drug and carrier microparticles and to evaluate the degree of drug-loading by the carrier (Doty et al., 2017; Shetty et al., 2021). Therefore, in our co-suspension study, a low dosage of water-soluble FF was added with DSPC to prepare the carrier, namely lip/FF-microparticles. Since FPF (FF) could roughly represent the aerodynamic performance of the carrier (due to its low impact on the microparticle structure), △FPF, which was obtained by subtracting FPF(MF) from FPF(FF), could be used to indicate the adequacy of MF particles loaded by the carrier. The more complete the adsorption was, the closer the △FPF was to zero. The results of aerosol performance showed that △FPF decreased significantly with the increase of Lip/FF-microparticles, indicating that the sufficiency of MF being loaded depended on the amount of Lip/FF-microparticles. When the Lip/FF-microparticles were 30 times that of MF (Table2 (P10)), the △FPF was reduced to approximately 3%, suggesting that the amount of carrier microparticles at this ratio could meet the loading requirements of micronized MF.

In contrast to the co-suspension mode where drug crystals were absorbed onto the surface of the carrier microparticles, in the drug embedding mode, drug particles were embedded inside the carrier microparticles, allowing more drug particles to be bonded to the carriers. The binding of drug crystals and carriers in the drug embedding mode, which depended on the matrix strength of the microparticles, was also different from that in co-suspension mode; the former had higher binding strength and was less affected by the polarity of drug molecules. Therefore, micronized MF and DSPC were homogenized together to prepare drug-embedding microparticles by SFD. The results of △FPF (Fig. 4 (Group B)) suggested that the correlation between FPF (MF) and the carrier dosage was reduced. When the ratio of DSPC and MF decreased to 3:1 in the drug embedding mode (Table 2(P14)), the FPF(MF) of P14 remained similar to that of P10 at the ratio of 30:1 in co-suspension mode, but the amount of DSPC was only approximately 10% that of P10. However, △FPF was as high as 5%, indicating that there remained a certain amount of micronized MF that was not loaded effectively by carrier microparticles; therefore, this drug-loading method requires further improvements.

To achieve the goal of efficient embedding, the targeted particle size of raw MF particles should be much smaller than the inhalable microparticles obtained by co-SFD. The ideal particle size of raw MF should be 100 nm or less. Because MF is hard, it is difficult to obtain microparticles with a small particle size by mechanical crushing processes (such as jet milling). Although MF is insoluble in water, bottom-up manufacturing processes could be applied to meet the objectives. Solution-mediated phase transformation (SMPT) provides a feasible technical solution. As the solubility of amorphism is higher than that of stable crystal, amorphous substances dissolve partially with the help of solvent and then precipitate to form a stable crystal. The solubility difference between different crystal forms is the force of transformation (Greco and Bogner, 2012). In this study, amorphous MF prepared by SFD was used as the raw material of the embedding inhalable microparticles. The transient molecular dispersion state of MF during the crystal transformation process during feedstock preparation can be used to increase its contact adequacy with DSPC and further improve the drug-loading efficiency. Compared to P14, P17 obtained the higher FPFs of the two drugs, and FPF(MF) was very close to FPF(FF), even though the DSPC:MF ratio continued to decrease to 1:1. The results indicated that the crystal transformation embedding method could improve the drug-loading efficiency of water-insoluble drugs with phospholipid carriers in the co-SFD process. Compared to the co-suspension mode, the dosage of loaded drug by the crystal transformation embedding mode can increase by about 15-fold.

4.3. Performance of pMDIs with different preparation principles

To evaluate the differences in inhalation characteristics between phospholipid microparticle pMDI and the traditional drug crystal-only pMDI, P17, which had the least DSPC in this study, and the brand-name drug Dulera®, which uses a micronized drug in the form of inhalable microparticles, were selected for the comparative study. In addition to the significant improvement of FPF, the phospholipid microparticle pMDI (P17) also had important advantages in DDCU.

The phospholipid microparticles prepared in this study had good suspension stability, which provided a physical property for achieving consistent delivered dose of pMDI. Common handling errors in patients, such as failure to shake the pMDI sufficiently before taking a dose and delay between shaking and actuation, could result in the delivered drug being higher or lower than the targeted dose (Crompton, 1982). The severity of this problem depends on the rate of aggregation and sedimentation of the suspension system (Doty et al., 2017). To evaluate the impact of these common handling errors on DDCU, the phospholipid microparticle pMDI and the drug crystal-only pMDI were subjected to different shaking modes prior to obtaining the delivered dose.

When the modified shaking modes of either a delay from shake to actuation (mode 2) or a gentle shake (mode 3) were employed, the drug crystal-only pMDI showed marked variability, indicating that the micronized drug particles aggregated in a short time, resulting in uneven concentration. In modes 2 and 3, the mean delivered dose for MF was close to the targeted dose, while the mean delivered dose for FF varied greatly. In these cases, the mean delivered dose for FF was 76.2% lower than the targeted dose under mode 2 and 126.9% higher than the targeted dose under mode 3, respectively. These results demonstrate that the suspension behavior of different micronized drug crystals in HFA is very different, and the same continuous phase is difficult to meet the requirements of multiple drug molecules in parallel. Phospholipid inhalable microparticles prepared by SFD can adjust the particle density according to the density of the specific propellant to achieve optimal suspension performance. The crystal transformation embedded method enables the delivery of multiple drugs to be more robust and consistent under challenging shaking conditions. All single delivered doses of MF and FF were within the acceptable range, and the ratio of the two drugs was extremely stable and close to the targeted value, thus meeting the clinical requirements.

The polarity of the DSPC microparticles is similar to that of the HFA, which reduces the tendency of microparticles to aggregate. Moreover, surface irregularities can significantly decrease the interparticle aggregate strength (Weiler et al., 2010). The phospholipid microparticles in this study displayed corrugated surfaces, which also significantly improved the dispersibility of the particles and enabled the pMDI to obtain a high drug delivery efficiency and a consistent delivered dose.

5. Conclusion

The drug-embedding phospholipid microparticles were prepared by utilizing the trans-crystallization of drugs in the amorphous form during the preparation of feedstock. This preparation method broke the two technical bottlenecks that limit the application of inhalable microparticles by co-SFD in pMDIs. One is the low drug-loading efficiency caused by the poor combination of carrier and insoluble drug crystals during co-SFD, while the other is the physical instability of the inhalable microparticles obtained when the drug solution is used during co-SFD. Phospholipid microparticle pMDI prepared using the SFD technology not only achieved consistent dose delivery and an FPF which is greater than that of drug crystal-only pMDI (Dulera®), but it also reduced the percentage of DSPC to approximately 4% compared to that used by the co-suspension technology. In addition, the crystal transformation embedding mode can increase the amount of drug-loading by approximately 15-fold compared to the co-suspension mode. This delivery technology may also be used to prepare other high-dose drugs for inhalation.

Credit author statement

Quan Xi: Conceptualization, Methodology, Investigation, Writing-original draft, Funding acquisition, Resources.

Jiaying Miao: Data curation, Investigation, Formal analysis.

Zhen Cao: Visualization, Investigation, Formal analysis.

Hao Wang: Methodology, Writing-reviewing, Editing, Visualization, Supervision, Funding acquisition.

Declaration of Competing Interest

No potential conflict of interest was reported by the authors.

Appendix A. Supplementary data

Supplementary material 1

Supplementary material 2

Data availability

The data that has been used is confidential.