Ethoxy acetalated dextran nanoparticles for drug delivery: A comparative study of formulation methods

Abstract

Dextran-based polymers, such as ethoxy acetalated dextran (Ace-DEX), are increasingly becoming the focus of research as they offer great potential for the development of polymer-based nanoparticles as drug delivery vehicles. Their major advantages are the facile synthesis, straightforward particle preparation and the pH-dependent degradation of the particles that can be fine-tuned by the degree of acetalation of the polymer. In this study we have shown that Ace-DEX can not only compete against the commonly used and FDA-approved polymer poly(lactic-co-glycolic acid) (PLGA), but even has the potential to outperform it in its encapsulation properties, e.g., for the herein used anti-inflammatory leukotriene biosynthesis inhibitor BRP-187. We used three different methods (microfluidics, batch nanoprecipitation and emulsion solvent evaporation) for the preparation of BRP-187-loaded Ace-DEX nanoparticles to investigate the influence of the formulation technique on the physicochemical properties of the particles. Finally, we evaluated which production method offers the greatest potential for achieving the demands for a successful translation from research into pharmaceutical production by fulfilling the basic requirements, such as reaching a high loading capacity of the particles and excellent reproducibility while being simple and affordable.

Graphical abstract

1. Introduction

The clinical translation of polymer-based nanoparticles (NPs) has been moderately slow thus far, despite the enormous scientific output regarding their formulation and characterization. Although the challenges to reach market approval are multifold, optimizing particle production to minimize process variability is one of the main concerns (Ferrari et al., 2018; Mulhopt et al., 2018; Wang et al., 2021). Several production methods of solid polymer NPs are well established in basic laboratory research, e.g., batch nanoprecipitation (BP) and emulsion (EM)-based methods. However, the translation of NP-based drug products to clinical scale is still considered a bottleneck and is not up to speed with the increased interest in their biomedical application (Gonella et al., 2022; Yang et al., 2022). In this regard, technologies that offer continuous production with well-controlled process parameters are essential. Within the last decades, the utilization of microfluidic technology was established (Karnik et al., 2008; Ma et al., 2017). This technology became the “must-have” for controlled formulation and has been brought to many applications where good manufacturing practice (GMP) is imperative (Webb et al., 2020). Indeed, microfluidics (MF) seems to solve many problems in terms of reproducible production of drug-loaded nanocarriers (Shrimal et al., 2020; Valencia et al., 2012). Up to now, a special focus has been on the microfluidic assisted formulation of biocompatible and biodegradable polyesters such as poly(lactic acid) PLA, poly(lactic-co-glycolic acid) (PLGA) or poly(caprolactone) (PCL) as indicated by the high scientific output (>22,000 publications). These polyesters are well-known and have been used for decades to produce NPs (Gupta et al., 2021). Previously, we demonstrated that a microfluidics-assisted nanoprecipitation of PLGA not only produced defined and reproducible particles, but also substantially improved the drug loading capacity (>3× higher than before) (Behnke et al., 2020).

Interestingly, only a few studies describe the formulation of dextran-based polymers with microfluidics, although such materials, e.g., acetalated dextrans are being considered a promising alternative to the commonly used polyesters (Ma et al., 2019). Wang et al. highlighted the therapeutic potential and formulation methods of the acetalated dextran-based carriers recently in their review (Wang et al., 2021). The tunable acetalated dextrans which are mainly employed in the field of drug delivery are methoxy acetalated dextran (Ac-DEX) and ethoxy acetalated dextran (Ace-DEX) among others, and are derived from the FDA-approved dextran (Bachelder et al., 2017; Cohen et al., 2010; Deirram et al., 2019). With their pH-dependent (bio)degradability, these materials offer an interesting advantage in drug delivery (Binauld and Stenzel, 2013; Kauffman et al., 2012). Ace-DEX behaves similarly to the more known Ac-DEX; however, the difference in degradation behavior is crucial. While Ac-DEX degrades to methanol and acetone (Ac), Ace-DEX degrades to ethanol and acetone (Kauffman et al., 2012). In particular with regard to translation and potentially high doses that might have to be applied in vivo, Ace-DEX offers advantages due to its lower toxicity (Kauffman et al., 2012). Dextran-based polymers have already been formulated into NPs with various drug molecules showing promising carrier properties (Fu et al., 2018; Graham-Gurysh et al., 2020; Kretzer et al., 2021). Hence, in some studies, Ace-DEX is considered to be one of the polymer systems that will be used as nanocarrier material in clinical trials in the near future (Wang et al., 2021). We also recently demonstrated that drug-loaded Ace-DEX particles revealed an increased drug encapsulation performance compared to PLGA particles (Kretzer et al., 2021). In this previous study, the active pharmaceutical ingredient (API) BRP-187 (4-(4-chlorophenyl)-5-[4-(quinoline-2-ylmethoxy)phenyl] isoxazol-3-carboxylic acid) was encapsulated as the active ingredient. BRP-187 belongs to the class of novel dual inhibitors that target the 5-lipoxygenase-activating protein (FLAP) and microsomal prostaglandin E₂ synthase-1 (mPGES-1) (Koeberle and Werz, 2015, Koeberle and Werz, 2018). Such dual inhibitors of FLAP/mPGES-1 belong to a recently established class of anti-inflammatory drugs known as Lipid Mediator Class Switch Inducers (LMCSIs) (Kretzer et al., 2022; Levy et al., 2001; Serhan et al., 2014). They have shown superior in vitro efficiency when compared to the conventional non-steroidal anti-inflammatory drugs (NSAIDs), mainly by inhibiting the biosynthesis of pro-inflammatory prostaglandins and leukotrienes, while at the same time promoting the biosynthesis of the pro-resolving lipid mediators during inflammation (Gerstmeier et al., 2019). BRP-187 represents a promising anti-inflammatory compound; however, it strongly binds plasma protein thereby reducing its in vivo bioactivity (Garscha et al., 2016). Therefore, BRP-187 requires the utilization of bioavailability enhancement techniques, such as formulation into polymer-based NPs, to prevent plasma protein binding (Shkodra-Pula et al., 2020a). An initial study used batch nanoprecipitation yielding loading capacities in the lower range of 2 to 3%. In order to enable in vivo studies of Ace-DEX particles and novel drugs, however, a robust formulation with a higher drug load is desired. Optimal formulation conditions must be found to enable the clinical application of a new polymer-based drug delivery system. The formulation requires to be applicable in the industry and provide stable carrier formulation with sufficient drug loading while meeting the regulatory and economic requirements. First machine learning approaches have already been developed to guide formulations, e.g. in terms of miscibility and solubility of APIs with polymers. However, experimental research is necessary to confirm the AI-generated results, and also provide new data to assist further in silico predictions. Therefore, we herein studied the most commonly used formulation methods in academia and industry, i.e., batch nanoprecipitation and emulsion solvent evaporation, and compare them to microfluidics for their feasibility in preparing stable Ace-DEX particles loaded with BRP-187 as a representative for its drug class (Vauthier and Bouchemal, 2009; Zielinska et al., 2020).

Microfluidics is a formulation technique that manipulates fluids within a small-scale environment (Shkodra-Pula et al., 2020b; Strook, 2008; Whitesides, 2006). In principle, fluids are injected into miniaturized reactors—known as micromixers or chips—composed of micro-sized channels with geometries that ensure vigorous mixing (Shkodra-Pula et al., 2020b). The prevailing idea about the feasibility of microfluidics is that this approach holds great potential to streamline high throughput production at low energy input while using very small volumes of materials (Valencia et al., 2012). Miniaturization enables a continuous process because the lateral dimension is reduced, and thus, the mixing efficiency is increased due to the higher surface-to-volume ratio and the lower effective volume (Lu et al., 2018; Xin et al., 2020).

Nanoprecipitation is a bottom-up formulation method that allows the arrangement of molecules based on their thermodynamic affinities in a multiphase system, in which solutes arrange to minimize their free energy and reach thermodynamic stability (J. Hitanga et al., 2015; Sinha et al., 2013). However, if the conditions in such a multiphase system are not thermodynamically favorable, there is a high tendency for particle aggregation and/or solute precipitation. This tendency is somewhat reduced in top-down formulation methods that use external forces (e.g., high-power ultrasonication), which break down particle aggregates and force the system into thermodynamic equilibrium.

The emulsion method is also a bottom-up process and can be used in two variants, a single emulsion (as in this study) or a double emulsion (Nava-Arzaluz et al., 2012). The polymer and the active ingredient are dissolved in a volatile, non-water-miscible organic solvent. The properties of the particles can already be influenced by the choice of solvent. The second phase is water containing a stabilizer, such as the surfactant poly(vinyl alcohol) (PVA) used in our study. After the addition of the aqueous phase to the organic phase, an emulsion with defined nanodroplets can be produced by using ultrasound or a homogenizer, for example. The NPs are then hardened by the subsequent evaporation of the solvent.

The techniques were compared with respect to the final particle size, polydispersity index (PDI), and stability depending on the initial drug loading. Two different drug concentrations were tested: 3% and 10% (w/w) referred to the polymer weight. Furthermore, the inter-day reproducibility for each formulation method was analyzed and the adaptability of the formulation protocols for scale-up production in GMP-environment was assessed.

2. Materials and methods

2.1. Materials

BRP-187 (4-(4-Chlorophenyl)-5-[4-(quinoline-2-ylmethoxy)phenyl] isoxazol-3-carboxylic acid) was synthesized based on established procedures (Banoglu et al., 2016). Acetone (>99%, extra pure) was purchased from Acros Organics. Ethyl acetate (ROTISOLV® ≥99.9%, GC Ultra Grade) was purchased from Carl Roth GmbH. Poly(vinyl alcohol) (PVA) (Mowiol 4–88, M_w 31,000 g mol⁻¹), dimethylsulfoxide (DMSO, anhydrous ≥99.9%), proteinase K from Tritirachium and all other materials were received from Sigma Aldrich unless otherwise stated. Ace-DEX was synthesized based on a previously published protocol (Mn 13,700 g mol⁻¹, Р= 1.41, DS_cyclic = 1.17 and DS_acyclic = 0.87) (Kauffman et al., 2012). Purified water was used in all stages of NP preparation, purification, and characterization.

2.2. Nanoparticle formulation

2.2.1. Microfluidics

Two low-pressure syringe pumps (neMESYS 290 N, Cetoni, Korbußen, Germany) with volume capacities of 10 mL (used for the dispensing the polymer/drug solution) and 25 mL (used for the dispensing of the aqueous PVA solution) were connected through polytetrafluoroethylene (PTFE) tubing (ID: 0.5 mm, OD: 1 mm) to a herringbone chip (Zeonor COP, microfluidic ChipShop, Jena, Germany). The chip consisted of three staggered herringbone fields of 600 μm width. Each herringbone field had two inlet channels (200 μm depth × 300 μm width) and one outlet channel (600 μm). The herringbone bone chip was selected because it has a precise control over the size and shape of the particles due to the specific geometry that generates a high level of shear force and mixing, as well as a low energy requirement and high efficiency as well as low cost due to the simple design and the fact that it can be manufactured with inexpensive materials.

The formulation parameters were based on a previously published protocol established for PLGA particles (Behnke et al., 2020). For all formulations, Ace-DEX solutions of 15 mg mL⁻¹ were prepared in an organic solvent, either acetone or a 1:9 mixture of ethyl acetate (EtOAc) to acetone. The required amount of drug (10 mg mL⁻¹ in DMSO) was added to the polymer phase. Next, 0.5 mL of the polymer solution (or polymer/drug solution) were co-injected with 4 mL of the aqueous solution of 0.3% (w/v) PVA through the herringbone chip (flow rate ratio (FRR) 4:1) and a total flow rate (TFR) of 2000:500 μL min⁻¹. The NPs were collected via the outlet channel into a glass vial and mixed with 10 μL 0.01% triethylamine (TEA). A minimum of three to a maximum of five NP formulations were prepared consecutively and stirred on a magnetic stirring plate (MIX 15 eco stirrer, 2mag AG, Munich, Germany) at 800 rpm for 2 to 3 h. Next, the NP formulations were pooled and purified via centrifugation (section 2.2.4). Particle size distribution of the NP formulations was measured via DLS before and after pooling (section 2.3).

2.2.2. Batch nanoprecipitation

Ace-DEX was used to encapsulate BRP-187 at a 3% and 10% (w/w) theoretical loading. Ace-DEX (75 mg) was dissolved in acetone (5 mL), and BRP-187 was dissolved in DMSO to make a stock solution of 10 mg mL⁻¹. The polymer solution was mixed with 225 μL or 750 μL drug solution for the 3% or 10% loading, respectively. In a 100 mL glass bottle, 40 mL of 0.3% (w/v) aqueous PVA solution with 100 μL TEA 0.01% (v/v) was prepared. Next, the organic (polymer/drug) solution was transferred into a 5 mL syringe with a 21 G × 4^3/4 (0.8 × 120 mm) cannula. The syringe was mounted in the syringe pump (Aladdin AL1000–220, World Precision Instruments, Friedberg, Germany) and the canula was bent 90° and immersed in the aqueous phase, touching the glass wall. The organic phase was precipitated into the aqueous phase at a flow rate of 2 mL min⁻¹, while stirring on a magnetic stirrer at 800 rpm, using a 7 × 25 mm stirring rod. To remove acetone from the formulation, the NP dispersion was left stirring at 800 rpm under a fume hood for 24 h. All formulation steps were performed at room temperature.

2.2.3. Emulsion-evaporation

For each formulation approach, two times Ace-DEX (15 mg) was dissolved in ethyl acetate (1 mL) and mixed with 45 μL or 150 μL of BRP-187 solution (10 mg mL⁻¹ prepared in DMSO) for 3% or 10% (w/w) theoretical loading, respectively. The polymer and drug solution were mixed in a 15 mL falcon. Before sonication, 2 mL of 1.5% (w/v) PVA solution were added to the organic phase. Emulsification was achieved using an ultrasound probe (Hielscher Sonotrode S26d2, Hielscher Ultrasonics, Teltow, Germany) with Ø 2 mm and approx. 120 mm in length and an ultrasonic generator (UP200ST, Hielscher Ultrasound Technology, Teltow, Germany). The ultrasonication settings were as follows: 100% cycle at 100 W power, 20% amplitude for 20 s, with the falcon immersed in ice. The emulsion was immediately transferred into 8 mL of pure water and 30 μL TEA 0.01% (v/v). To remove EtOAc from the formulation, the NP dispersion was stirred at 800 rpm under a fume hood for minimum 2 h. After the evaporation of the EtOAc, the samples were measured via DLS and the two 15 mg approaches were pooled to form a 30 mg batch. This pooled sample was subsequently measured again via DLS and purified by centrifugation (section 2.2.4). All formulation steps were performed at room temperature.

2.2.4. Nanoparticle purification

After formulation and solvent-evaporation, the NPs were centrifuged to remove excess surfactant and/or unencapsulated drug. NP dispersions were transferred into 50 mL falcons and centrifuged on a fixed rotor speed of 16,639 ×g for 60 min and 120 min, respectively, for the emulsion approaches at 20 °C (Centrifuge 5804 R, Eppendorf, Wesseling-Berzdorf, Germany). The supernatant was removed, and the NP pellets were redispersed in 5 mL pure water with 100 μL TEA 0.01% (v/v) (for batch nanoprecipitation), 3.2 mL pure water with 128 μL TEA 0.01% (v/v)) (emulsion-evaporation) or 0.8 mL of pure water with 32 μL 0.01% (v/v) TEA for each formulation, i.e. maximum 4 mL H₂O and 160 μL 0.01% (v/v) TEA for microfluidics. The purified NPs were vortexed for 10 to 20 s and sonicated for 30 min in an ultrasound water-bath. The redispersed NPs were stored at 4 °C for 24 to 48 h to allow adequate rehydration before they were finally lyophilized and stored at 4 °C. The yield was determined via the following formula:

$$\mathit{Yield}=\frac{\mathit{mass\; of}\mathit{NP}\mathit{recovered}-\mathit{mass\; of\; drug\; recovered}-\mathit{mass\; of}\mathit{PVA}\mathit{recovered}}{\mathit{mass\; of\; polymer\; introduced\; to\; the\; formulation}}\times 100$$

2.3. Nanoparticle characterization

2.3.1. Dynamic light scattering

Dynamic light scattering (DLS) and electrophoretic light scattering (ELS) were used to characterize the size distributions and zeta-potentials (ZP, surface potential) of the NPs, respectively, utilizing the Zetasizer Ultra (Malvern Panalytical GmbH, Malvern, United Kingdom). The Zetasizer operates with a laser wavelength of 633 nm and the particles were measured in polystyrene microcuvettes (Brand GmbH + Co KG, Wertheim, Germany) at 25 °C with a back-scattering angle of 174.7°. The hydrodynamic diameter (d_H) and PDI of the NPs reported here correspond to the intensity-weighed particle size distribution. d_H and PDI of the particles prepared from all three methods were characterized at three points: After evaporation of the organic solvent, after centrifugation and subsequent resuspension in water, and after lyophilization and subsequent resuspension in water. The zeta-potential was characterized after centrifugation and subsequent resuspension in water, using the DTS1070 capillary cuvette. The NPs were diluted 1:100 with pure water and measured five times, 15 runs each with a run time of 1.68 s, for d_H and PDI, and measured three times for determination of ZP. The degradation behavior of drug-loaded particles was investigated by incubating the particles with a 2 mg mL⁻¹ proteinase K solution (in water) at a mass ratio of 1:49, monitoring the mean count rate (kcps) over a period of 26 h. For this purpose, the attenuator was kept constant at eight and the particles were measured in 838 measurements, each with 15 runs and a respective run duration of 1.68 s.

2.3.2. Scanning electron microscopy (SEM)

A Sigma VP Field Emission Scanning Electron Microscope (Carl-Zeiss AG, Oberkochen, Germany) was used to obtain the particle images. The micrographs were acquired with the InLens detector at a 6 kV acceleration voltage. 10 μL of the 1:5 diluted NP dispersions were pipetted on mica substrates and air-dried. Before the measurement, samples were coated with a thin layer of platinum (4 nm) via sputter coating (CCU-010 HV, Safematic).

2.3.3. UV-VIS spectroscopy

Encapsulation efficiency (EE) and loading capacity (LC) of the BRP-187-loaded NPs were determined

utilizing a UV-VIS plate reader (Infinite M200 Pro plate reader, Tecan Group Ltd. Männedorf, Switzerland). Initially, NP suspensions were lyophilized into aliquots of 300 to 500 μL. The dried NPs were subsequently dissolved in UV-grade DMSO in a volume equal to the lyophilized aliquots. The UV absorbance of 100 μL of particle solution was measured undiluted and diluted (1:2, 1:4 and if necessary up to 1:8) at λ = 316 nm, with 3 × 3 multiple reads per well and a 2000 μm well border using a Hellma Quartz 96-well plate (Hellma, Müllheim, Germany). The concentration of BRP-187 in the NPs was determined based on a calibration curve of BRP-187 obtained from a concentration range of 1.95 to 250 μg mL⁻¹. The EE and LC values were calculated using the following formulas:

$$\mathit{LC}=\frac{\mathit{mass\; of\; drug\; recovered}}{\mathit{mass\; of\; particle\; recovered}}\times 100$$

$$\mathit{EE}=\frac{\mathit{LC}}{\mathit{LC}\mathit{theoretical}}\times 100$$

The residual amount of PVA (%, w/w) in the lyophilized NPs was also quantified via UV-VIS spectroscopy according to a previously published protocol (Spek et al., 2015). The concentration of the resuspended NPs was thereby between 4.0 and 11.1 mg mL⁻¹.

2.4. Evaluation of NPs in human leukocytes for their efficiency to inhibit 5-lipoxygenase (5-LOX) product formation

2.4.1. Cell isolation and cell culture

Leukocyte concentrates were prepared from peripheral blood obtained from healthy human adult donors (females and males between 18 and 65 years) that received no anti-inflammatory treatment for the last ten days (Institute of Transfusion Medicine, University Hospital Jena, Germany). The approval for the protocol was given by the ethical committee of the University Hospital Jena and all methods were performed in accordance with the relevant guidelines and regulations according to a standard protocol described earlier (Jordan et al., 2020). To isolate polymorphonuclear leukocytes (PMNL), the leukocyte concentrates were mixed with dextran (dextran from Leuconostoc spp. M_W ∼ 40,000 g mol⁻¹, Sigma Aldrich, Taufkirchen, Germany) for sedimentation of erythrocytes; the supernatant was centrifuged on lymphocyte separation medium (Histopaque®-1077, Sigma Aldrich). Contaminating erythrocytes in the pelleted PMNL fraction were removed by hypotonic lysis using water. The pelleted PMNL fraction was subsequently washed twice in ice-cold phosphate-buffered saline pH 7.4 (PBS) and finally resuspended in PBS.

2.4.2. Determination of 5-LOX product formation in PMNL

For evaluation of the effects on 5-LOX product formation in human PMNL, cells (5 × 10⁶ in 1 mL) were pre-incubated in PBS containing 0.1% glucose and 1 mM CaCl₂ with vehicle, differential generated NPs (encapsulated with BRP-187 or without at a concentration of 1 μM) or free 1 μM BRP-187 for 15 min at 37 °C (Kretzer et al., 2021). Cells were then stimulated with 2.5 μM Ca²⁺-ionophore A23187 (Cayman, Ann Arbor, USA) for 10 min, and then the incubation was stopped with 1 mL ice-cold methanol containing 200 ng mL⁻¹ prostaglandin B₁ as internal standard. Samples were subjected to solid phase extraction and formed 5-LOX products were separated and analyzed by RP-HPLC as described (Werz et al., 2002).

2.5. Statistics

Results are expressed as mean ± S.E.M. of each independent experiment, where n represents the indicated numbers from separate donors. Statistical analysis and graphs were made by using GraphPad Prism 9 software (San Diego, CA, USA). A p-value ≤0.05 is a criterion for statistical significance.

3. Results and discussion

Although the formulation of Ace-DEX particles using various formulation methods has already been described in the literature; to the best of our knowledge there are no published studies in which microfluidics, not only by utilizing a herringbone chip (as applied here) but in general, have been used (Kretzer et al., 2021; Wang et al., 2021). Accordingly, there is a lack of studies comparing different formulation techniques—including microfluidics—for the preparation of drug-loaded Ace-DEX NPs. Yet, the introduction of the API into the formulation process of a polymer represents a crucial point and can become a knockout criterion for the particular polymer or technique. Depending on the properties of the API and the polymer chains they interact with each other by attractive forces or non-specific interactions and form an even more condensed/packed particle—a condition that can be applied to produce very small particles—or it can lead to instabilities resulting in crushing out of the drug or aggregation of the particles (Mackenzie et al., 2015; Reisch et al., 2015). The interaction of the drug and polymer is thereby influenced by the formulation method and the applied processing conditions, e.g. the used solvent, polymer, and drug concentration as well as the concentration of the surfactant (Cortes et al., 2021; Mackenzie et al., 2015; Muljajew et al., 2021).

Herein, three different formulation methods, i.e. (1) microfluidics (w/o EtOAc and w EtOAc), (2) batch-nanoprecipitation, and (3) emulsion, were selected for the preparation of Ace-DEX[BRP-187] particles to study their influence on the particle formation (Table 1). To ensure the comparability of the methods, the key parameters were kept constant. These parameters include, first and foremost, the concentration of the polymer phase as well as the applied drug feed. Accordingly, a polymer concentration of 15 mg mL⁻¹, a PVA content of 0.3% (w/v) referred to water and a drug stock solution in a concentration of 10 mg mL⁻¹ were used in all methods to test two drug loadings (3% (w/w) and 10% (w/w)) next to blank formulations (Table 1 and Table SI 1). The stock solution of the drug was prepared in DMSO because the hydrophobic BRP-187 is only soluble in selected solvents such as DMSO, dimethylformamide and dimethylacetamide. Due to its lower overall toxicity, DMSO was selected from the suitable solvents (Shkodra-Pula et al., 2020a). Additionally, EtOAc was used as a co-solvent with a proportion of 10% for the microfluidic approach to investigate whether it reveals a positive effect on the formulation, such as a less disruptive workflow due to less clogging. The formulation factors that were selected for the experimental screening phase were considered optimal based on investigations from preliminary working protocols (Behnke et al., 2020; Shkodra-Pula et al., 2019; Shkodra-Pula et al., 2020a).

Table 1.

Results of the main characteristics of the particles based on the preparation method (MF = microfluidics, BP = batch precipitation and emulsion with subsequent solvent-evaporation).

	#	Solvent	Drug feed [%]	Size [nm]	PDI	ζ [mV]	LC [%]	API precipitatesa
MF Ac	P1	Ac	0	182	0.06	−27.1	/	−
	P2	Ac	3	212	0.12	−35.1	1.51	−
	P3	Ac	10	229	0.16	−32.7	7.80	−
MF EtOAc:Ac	P4	EtOAc:Ac 1:9	0	193	0.09	−29.1	/	−
	P5	EtOAc:Ac 1:9	3	248	0.13	−32.6	2.18	+
	P6	EtOAc:Ac 1:9	10	280	0.17	−30.9	9.61	+
BP	P7	Ac	0	166	0.05	−15.9	/	−
	P8	Ac	3	202	0.06	−22.9	2.50	−
	P9	Ac	10	242	0.08	−31.4	9.07	−
Emulsion	P10	EtOAc	0	117	0.07	−2.4	/	−
	P11	EtOAc	3	111	0.18	−5.4	/	++
	P12	EtOAc	10	396	0.49	−24.9	/	++

Ac: Acetone, EtOAc: Ethyl acetate, Size (d_H) in nm, polydispersity index (PDI), zeta-potential (ζ), drug loading capacity (LC) in % referred to NP mass.

^aSEM analysis with respect to visible BRP-187 precipitates (− (no), + (few), ++ (many)).

The final particle suspensions were examined by DLS in terms of particle size, size distribution and via ELS to obtain the zeta-potential (Fig. 1, Table 1 and Table SI 1). A detailed SEM analysis was performed to investigate whether the drug BRP-187, which tends to precipitate at high drug feeds, is also present as free precipitates. Such drug precipitates are often invisible in the DLS measurements, which makes intensive SEM measurements necessary (Fig. 2 and Fig. SI 2). Furthermore, the residual PVA content and the API quantity were determined for all NPs by UV‐Vis measurements to determine the loading capacity (Table 1 and Table S1).

Fig. 1.

Particle attributes, including the hydrodynamic diameter (d_H) (A), polydispersity index (PDI) (B) and loading capacity (LC) (C) of NPs prepared with microfluidic-assisted nanoprecipitation (MF), batch nanoprecipitation (BP) and emulsion-evaporation (EM).

Fig. 2.

SEM images depict the morphology of the loaded nanoparticles depending on their preparation method. The left row shows the formulations with a drug feed of 3% (w/w) (P2, P5, P8, and P11) and the right row shows those with a 10% (w/w) drug feed (P3, P6, P9, and P12).

The formulation of unloaded Ace-DEX particles was a prerequisite for this comparative study and is in general possible as it was described in previous studies for the two standard methods emulsion and nanoprecipitation (Kretzer et al., 2021; Wang et al., 2021). However, besides the new microfluidic approach also the standard techniques were tested with the intended protocols for the preparation of blank Ace-DEX particles and in all methods the key parameters were kept constant.

The blank formulations of the Ace-DEX without BRP-187 (P1, P4, P7, and P10, Table 1 and Fig. SI 1) resulted in defined and monomodally distributed Ace-DEX particles with nanoparticle sizes ranging from 100 to 200 nm with PDI values below 0.1 for all methods. The particle sizes are in the expected range for the formulation using emulsion or nanoprecipitation as also described by other studies (Fu et al., 2018; Kretzer et al., 2021). Since microfluidics is in principle a method of controlled nanoprecipitation, particle sizes of those formulations are, as expected, in the range of those obtained by BP. The zeta-potential (ζ) of the blank Ace-DEX particles was in the negative range between −15.9 mV and − 29.1 mV for the precipitation methods (P1, P4, P7) and close to neutral for the emulsion method (P10, ζ = −2.4 mV). The reduction of the zeta-potential for P10 was caused by the higher concentration of the PVA surfactant used during the initial formulation. PVA interacts with the polymer chains at the particle surface and thus affects its surface potential (Cortes et al., 2021; Rebanda et al., 2022; Shkodra-Pula et al., 2019).

All methods were suitable to obtain well-defined blank Ace-DEX particles with the applied protocol. However, significant formulation-dependent differences were observed when the API (BRP-187) was introduced into the formulations (Fig. 1 and Table 1 (P2, P5, P8, and P11)). A drug feed of 3% (w/w) referred to polymer increased the particle sizes by 30 to 55 nm for the precipitation-based methods, i.e. MF and BP (P2, P5, and P8), whereas the same was not observed for the emulsion-based formulation (P11) (Table 1, Fig. 1A). However, the PDI increased to a higher factor for the NPs prepared by emulsion (an increase by 2.6 compared to an increase by a factor of 1.2 to 2 for the precipitation techniques) (Table 1). This higher PDI value of 0.18 already indicated more irregularities in the particle formation with the emulsion method (Table 1). It became even more pronounced when the drug feed was increased to 10% (w/w) (Fig. 1B). The high drug feed resulted in undefined, multimodal distributed particles with an average size of about 400 nm and a PDI value of 0.49 (P12), ultimately revealing that this method was not suitable for the formulation of stable Ace-DEX[BRP-187] particles. The precipitation methods also showed a shift towards larger particle sizes and broader particle distributions, however, the PDI increased only slightly and remained in the target range of below 0.2.

A closer look at the SEM images of the NPs prepared by the emulsion technique revealed that the non-encapsulated drug represents a large fraction of the sample, even more when the drug feed is increased from 3 to 10% (w/w) (Table 1 (P10 vs. P11 and P12) and Fig. 2). Considering the content of BRP-187 in the samples determined by UV-VIS spectroscopy, it was even more evident that high proportion of the total sample was due to the free drug precipitates (Table SI 1 (P11 a-c and P12 a-c)). Due to their size, the drug precipitates sediment faster than the NPs and are, therefore, less likely to be removed during purification compared to small NPs. The emulsion method is susceptible to interference, since the particle formation depends on several factors, i.e., the proportions of the solvent and anti-solvent, the total volume processed as well as the API dissolved in DMSO, which ultimately disturbs the production of stable particles. Evidence of DMSO disturbance was specifically observed when 10% drug was fed in the formulation, as no stable emulsions were formed (Table 1, Table SI 1, and Fig. 2). Although DMSO was a good solvent for the highly hydrophobic drug, it is still water-miscible, and as such, it is not favored for use in an emulsion-based method, considering that this method requires non-miscible solvents.

In these formulations, the use of EtOAc in the formulation favored the occurrence of drug precipitates. This effect was also observed in microfluidic formulations with ethyl acetate (MF EtOAc:Ac), albeit less marked as it was only used as a co-solvent up to 10% for the formulations P4-P6 and Fig. SI 2. If the three precipitation methods are compared in detail, the MF with EtOAc led to the largest (280 nm) and most widely distributed particles (PDI = 0.17), which, however, were within the target range of below 0.2 (Table 1 and Table SI 1). Ethyl acetate, even if typically used in emulsion formulations, is partially miscible with water up to about 10% ((v/v) at 25 °C). It is therefore often used in microfluidics because it affects the interfacial tension. This, like the viscosity, is increased in comparison to acetone and, thus, in theory, favors droplet formation. However, since it was only used to a small extent in our study and the mixing speeds within the chip are high, the actual formation of droplets is restricted. Nevertheless, this proportion of EtOAc leads to an influence on the interaction of organic to aqueous phase and, thus, also on the precipitation of the particles. Overall, compared to pure acetone, the precipitation does not occur instantaneously and quantitatively when the two phases first meet within the chip, but rather is slowed down. In particular at a high polymer concentration of 15 mg mL⁻¹, this leads to a lower strain on the very fine microfluidic channels, since a smaller number of particles precipitate per time unit compared to pure acetone. However, only if it is used to a limited amount of 10% (w/w), since it also favors the occurrence of free drug precipitates, at least for the BRP-187 used in this study and the formation of ill-defined structures, as has also been observed with the emulsion formulations (Fig. SI 1). This is due to the fact that the proportion of EtOAc in the organic phase leads to an altered interaction of the API with water. The hydrophobic active ingredient already gets in contact with the water to some extent, which leads to a partial precipitation of the pure drug even before the polymers precipitate as particles, since this takes place at a slower speed compared to pure acetone. Even if this can only be noted in isolated spots on the SEM images, it is important to assess the role of the non-encapsulated drug in nanoformulations, which are to be considered for clinical translations to further optimize the down-streaming purification steps process, if necessary. For microfluidic formulations where pure acetone was used (P1 to P3), the particles were even more defined. This can be attributed to faster precipitation of the particles (Table 1 and Table SI 1). Furthermore, SEM investigations revealed that these NPs were also more homogeneous and that free drug and undefined structures were not observed. The overall most defined and tendentially smallest particles, and also the most homogeneous according to the SEM investigations, were produced via BP (Table 1 and Fig. 2). This might be surprising at first glance since microfluidics in particular can offer lower batch-to-batch variability as a crucial advantage, but it has already been described in the literature that BP can achieve comparable or even better results if the protocol is optimized appropriately (Hamdallah et al., 2020).

Alongside the morphology of the particles (size and PDI), the final drug loading in the particles is considered a critical particle attribute. The emulsion method with the protocol used here was unsuitable for the encapsulation of BRP-187. As a consequence, these formulations were not considered any further. Meanwhile, for all three precipitation methods, a high drug loading between 7.8% up to 9.6% (w/w) BRP-187 relative to Ace-DEX was achieved (P3, P6, and P9, Fig. 2). In the case of the MF with EtOAc, however, occasionally free drug precipitates were observed in the SEM images, thus it is likely that the encapsulated drug was lower than 9.6% (w/w) (P6) as determined by UV‐Vis spectroscopy. Using BP, LC values in the range of 9.1% (w/w) (P9) were consistently achieved and were noticeably higher than the values obtained for the similar MF variant with pure acetone (7.8%, P3).

When evaluating the two best-performing methods (BP and MF) in terms of interday reproducibility, costs and formulation efforts, general applicability and scale-up potential, BP stood out. The interday reproducibility is important since even very slight differences in size in the nanometer-scale induce changes in the surface energy of the NPs, thereby influencing numerous factors, i.e. solubility, stability, and adsorption (Baer, 2018). Therefore, each individual formulation was prepared and carried out independently on different days. An advantage of the MF is that it is in general less operator-dependent. All formulation parameters are controlled by a software and the formulation itself is then carried out without the active involvement of the operator during the production. BP on a laboratory scale bears the risk that the reproducibility is reduced, for instance already due to the different positioning of the needle tip during the production process. However, this does not make the manufacturing method worse and less precise per se but underlines the necessity of precisely elaborated formulation protocols. Particularly given the production under GMP conditions and in large scale, parameters such as the immersion depth and the distance of the needle tip from the center of the vessel would have to be precisely defined. However, it has been shown here that reliable manufacturing is possible with BP, which even surpasses the MF. Nevertheless, it needs to be considered that microfluidic is a continuous process and can therefore be scaled up more easily (Valencia et al., 2012). The system we used has two pumps and allows a throughput of up to 10.8 g of polymer per day. However, the throughput can be increased if several pumps are connected in parallel. The neMESYS Base 120 system can operate with up to eight pumps. This process is also known as parallelization (Forigua et al., 2021; Valencia et al., 2012). In comparison, BP can also be upscaled, but the size of the reactors cannot straightforwardly be increased at will and may require several reactors at the same time to reach the desired scale. Another aspect to be considered are the costs of the formulation (Forigua et al., 2021; Valencia et al., 2012). For small scales for microfluidic assisted formulations, regularly new chips must be purchased leading to a higher initial investment in the equipment, which makes the production more expensive overall compared to BP, which uses low-cost needles and syringes. However, this point can also be relativized if chemically inert chips, such as glass- or stainless-steel chips, are established. In addition, with regard to upscaling, it must be considered that the initial acquisition costs for BP also increase significantly with the utilization of large reactors. Depending on the scale and available MF chips, the evaluation of the costs can therefore be different. Another hurdle that MF faces for the production of NPs, in particular for large and solid polymer-based particles, is the problem of clogging within the chip, which can cause the formulation to abort, resulting in a loss of material and associated costs. Therefore, it is necessary to find a suitable chip for each system initially. Based on this study and the excellent results that were achieved for the BP method, it must be noted that applying microfluidics with disposable chips for this drug-polymer system seems to become obsolete.

3.1. Efficiency of the generated NPs to inhibit 5-LOX product formation

In order to demonstrate that the encapsulated drug can be efficiently delivered to the cells of interest, human PMNL were used as primary immune cells to assess FLAP-dependent 5-LOX product formation. In agreement with our previous data (Garscha et al., 2016; Shkodra-Pula et al., 2020a), free BRP-187 at 1 μM efficiently inhibited 5-LOX product formation in isolated PMNL (Fig. 3). BRP-187 encapsulated in Ace-DEX particles and prepared by different formulation techniques, efficiently suppressed 5-LOX product formation in human PMNL compared to the free drug, implying that the efficiency of BRP-187 is not altered by formulation techniques (Fig. 3). Furthermore, particles containing 3% and 10% (w/w) BRP-187 showed similar inhibition potencies independent of the drug feed, whereas NPs without the drug (blank) had no inhibitory effect on 5-LOX product formation (Fig. 3). The efficient inhibition of 5-LOX is also an evidence that the particles released the drug as expected. The enzymatic degradation of the particles containing BRP-187 was investigated and proofs the ability of the particles to release BRP-187 in a relevant time frame (t₅₀ = 43 min, Fig. SI 3). In addition, the pH-dependent degradation and, thus, the release of the drug was already shown in a previous study (Shkodra-Pula et al., 2020a).

Fig. 3.

Inhibition of 5-LOX product formation by encapsulated BRP-187 in differentially generated NPs. PMNL were preincubated with PBS (control), free 1 μM BRP-187 or with differential generated NPs (encapsulated with BRP-187; respective 1 μM or without), for 15 min at 37 °C and then stimulated with 2.5 μM A23187. After 10 min, the reaction was stopped, and 5-LOX products were extracted via solid phase extraction (SPE) and analyzed by RP-HPLC. Values are given as 5-LOX products (LTB₄, trans-LTB₄, epi-trans-LTB_4, and 5-HETE) in percentage of control. For statistical analysis matched one-way ANOVA with Tukey's multiple comparisons test was used; n.s. = not significant.

4. Conclusion

Ace-DEX was investigated as carrier material for the compound BRP-187 as a model drug representing a new class of anti-inflammatory APIs with high lipophilicity, and was found to be superior in comparison to a previous study with PLGA as carrier material with respect to drug loading. While the earlier attempt to increase the drug loading of PLGA particles was highly associated with the occurrence of free drug precipitates, this was not observed for the herein used Ace-DEX polymer when acetone was used as solvent. With batch nanoprecipitation as the simplest technique a final drug load of 9.1% was achieved. The particles formulated with the developed protocol were well-defined and stable compared to those of microfluidics or emulsion. As hoped, the nanoparticles with a 10% drug feed showed a comparable efficacy profile compared to the particles with a 3% drug feed. This is particularly important for in vivo translation, as the required effect concentrations can be achieved with a lower dose of nanoparticles. Using this protocol and Ace-DEX as carrier material for BRP-187 encapsulation eliminated additional purification steps, which are associated with increased usage of material, time and costs. These are crucial facts in terms of industrial production and clinical translation.

Conclusively, Ace-DEX represents a very promising polymer for upcoming applications, especially also for the encapsulation of difficult to encapsulate APIs. In the future, we plan to translate the presented protocol for batch nanoprecipitation into a GMP conform production and utilize different APIs.

Author contribution

M.B.: Concept research idea and experimental design, Formulation and DLS characterization of nanoparticles, Quantification of the drug loading, SEM measurements, Preparation of manuscript, Correction of manuscript. P.K.: Synthesis of Ace-DEX, Correction of manuscript. P.D.: In-vitro determination of enzyme inhibition efficiency, Preparation of manuscript, Correction of manuscript. B.S.: Concept research idea and experimental design, Preparation of manuscript, Correction of manuscript. B.B.S.: Formulation and DLS characterization of nanoparticles, Quantification of the drug loading, Correction of manuscript. J.A.C.: Synthesis of BRP-187, Correction of manuscript. S.St.: Supervision of SEM measurements, Correction of manuscript. P.M.J.: In-vitro determination of enzyme inhibition efficiency, Preparation of manuscript, Correction of manuscript. S.Sch.: Correction of manuscript. S.H.: Supervision of SEM measurements, Correction of manuscript. A.V.: Concept research idea and experimental design, Preparation of manuscript, Correction of manuscript, Supervision. O.W.: Correction of manuscript. U.S.S.: Concept research idea and experimental design, Correction of manuscript, Supervision.

Declaration of Competing Interest

The authors declare no conflict of interest.

Appendix A. Supplementary data

Supplementary material

Ethoxy acetalated dextran nanoparticles for drug delivery: A comparative study of formulation methods—SI

Data availability

Data will be made available on request.