Preparation of Uniform Nonspherical PLGA Particles in the Nanosize Range by Using a Two-Step Premix Membrane Emulsification Technique

Abstract

The shape of polymeric particulate systems is a critical determinant of their in vivo fate, and nonspherical nanoparticles are promising candidates for advanced drug delivery. However, spherical-shaped particles are usually produced by conventional emulsion-based methods due to the spontaneous minimization of interfacial tension. In our previous work, uniform nonspherical poly­(lactic-co-glycolic acid) (PLGA) particles in the micro-size range were successfully prepared by a premix membrane emulsification technique with phosphate-buffered saline (PBS) as a deformation inducer. However, it is difficult to produce nanosized nonspherical particles because of the inherent conflict between the reduction of droplet size and emulsion deformation. To overcome this limitation, a two-step premix membrane emulsification strategy was developed in the present study to temporally and spatially decouple emulsion formation from shape deformation. Through optimization employing one-variable-at-a-time (OVAT) and orthogonal experimental design, the influence of key formulation factors, including PLGA concentration, poly­(vinyl alcohol) (PVA) concentration, PBS concentration, and oil phase volume, on the shape of the particles was systematically evaluated. Under the optimal conditions, uniform nanorods with minor axes of 331.11 ± 85.90 nm were fabricated, and the rod fabrication yield reached nearly 100%. Paclitaxel, a classical chemotherapeutic drug, was successfully encapsulated into the obtained nanorods without altering their anisotropic morphology. Furthermore, the drug-loaded nanorods exhibited enhanced in vitro anti-tumor efficacy against A549 lung cancer cells. This methodology provides a facile route for preparing uniform nonspherical PLGA particles in the nanosize range, offering promising applications as drug carriers in nanomedicine.

1. Introduction

With the development of biomedicine, biodegradable polymeric particulate systems have been extensively applied as drug carriers and tissue engineering scaffolds in biomedical fields, owing to their exceptional biocompatibility, excellent biodegradability, and controlled drug release profiles. − Among various polymers, poly­(lactic-co-glycolic acid) (PLGA), which has received US Food and Drug Administration (FDA) approval for injectable drug delivery systems, stands out as one of the most extensively investigated biodegradable polymers to date. Accumulating evidence demonstrates that the physicochemical properties of polymeric particles, such as size, surface charge, and shape, significantly influence their interactions with cells and tissues and eventually determine their fates in vivo. ,

Recent research has highlighted particle shape as a pivotal design parameter. It is recognized that nonspherical particles exhibit superior performance compared to their spherical counterparts, including reduced phagocytic clearance, prolonged systemic circulation, enhanced specific targeting ability via increased contact with cells, improved vascular margination through the edge effect, and augmented tumor tissue penetration. These advantages underscore the potential of nonspherical nanoparticles as promising platforms for advanced drug delivery systems in biomedicine.

Inspired by these unique advantages, significant efforts have been devoted to developing shape-engineered polymeric particles, and various methods for preparing nonspherical polymeric particles have been proposed in recent years, including particle replication in nonwetting templates (PRINT), 3D printing, microfluidic fabrication, film stretching, template assembly, and so on. While these methods enable precise morphological control and even the production of complex geometries such as bullet-shaped and discoidal particles, they often suffer from complex preparation processes, specialized instrument requirements, high costs, and scalability challenges, thereby restricting fundamental investigations into geometry-property relationships in polymeric particles and ultimately limiting the practical implementation of shape-specific polymeric particulates in biomedical fields. In contrast, emulsion-based methods offer distinct advantages in terms of scale-up feasibility and cost-effectiveness. , Nevertheless, conventional emulsion-based methods inherently favor spherical morphologies due to the minimization of interfacial energy during emulsification.

Recent advancements in emulsion , have demonstrated that the incorporation of some small guest molecules, such as tris base or sodium tripolyphosphate, can induce the deformation of emulsions during the solidification process. On the basis of this strategy, our group and others have successfully prepared rod-shaped PLGA particles via a modified emulsion solvent evaporation method. , However, these conventional emulsion approaches typically yield polydisperse particles when preparing nonspherical particles, raising concerns regarding batch reproducibility and restricting the practical implementation of these nanoparticles in clinical applications. Therefore, there remains a challenge in the development of a simple, controllable, versatile, and high-throughput method that can produce nonspherical polymeric particles in the nanosize range with uniform size and shape.

Premix membrane emulsification has emerged as a robust platform for manufacturing uniform polymeric particles. Our prior work established a shape-manipulation strategy combining the premix membrane emulsification technique with phosphate-buffered saline (PBS)-induced deformation, enabling the fabrication of uniform nonspherical PLGA microparticles. However, previous studies have indicated that when the size of these microparticles approaches or exceeds the dimensions of the microvasculature, pulmonary capillary trapping or retention in other organs may occur, thereby increasing the risk of adverse events, such as pulmonary embolism, and limiting their application in systemic administration. In contrast, controlling the size of nonspherical particles within the nanosize range is generally more favorable for intravenous systemic administration and may facilitate particle extravasation and transport through tissues and tumor interstitium. ,, In addition, at the same particle mass or polymer amount, reducing the particle size generally increases the total specific surface area of the system, thereby providing a more potentially available interfacial area for subsequent surface functionalization, ligand conjugation, and multifunctional design. , Therefore, the development of nonspherical nanoscale systems (<1 μm) with uniform size and well-controlled morphology is of great significance for further optimizing their drug delivery performance in biomedical fields like diagnosis and cancer therapy. ,, However, preparing uniform nonspherical PLGA nanoparticles remains considerably more challenging than preparing their micron-sized counterparts. Attempting nanoscale deformation in our prior work presented a challenge due to a critical contradiction: droplet size reduction (demanding higher energy input) compromises its deformation due to interfacial instability.

Herein, this study addresses these critical limitations by developing a sequential two-step premix membrane emulsification strategy that spatiotemporally decouples nanoemulsion formation from shape deformation processes. During the preparation, an orthogonal experimental design was employed for systematic parameter optimization. We first optimized a one-step premix membrane emulsification method using one-variable-at-a-time (OVAT) methodology and then engineered a novel two-step method to prepare uniform nonspherical PLGA nanoparticles using orthogonal experimental optimization. Critical formulation parameters, including PLGA concentration, poly­(vinyl alcohol) (PVA) concentration, PBS concentration, and oil phase volume, were rigorously evaluated through an orthogonal experimental design. Comparative analyses elucidate the mechanistic advantages of two-step processing over conventional one-step methods, while demonstrating the superior efficiency of orthogonal design versus traditional OVAT approaches. Finally, paclitaxel (PTX) was employed as the model drug to validate the feasibility and applicability of this method for the preparation of nonspherical drug-loaded nanorods. The in vitro anti-tumor performance of PTX-loaded nanorods against A549 lung cancer cells was further investigated.

2. Materials and Methods

2.1. Materials

PLGA with a d,l-lactide/glycolide mol ratio of 75/25 (Mw 13.6 kDa) was purchased from Daigang Biological Engineering Co., Ltd. (Shandong, China). Dichloromethane (DCM, AR), PBS, and 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) were purchased from Solarbio Biological Technology Co., Ltd. (Beijing, China). PVA (PVA-217, polymerization degree 1700, 88–89% alcoholysis degree) was purchased from Kuraray Co., Ltd. (Japan). Paclitaxel (PTX, degree of purity ≥99%, BR) was purchased from Dalian Meilun Biotechnology Co., Ltd. (China). Acetonitrile (chromatographic grade) was purchased from Chengdu Kelong Chemical Co., Ltd. (China). A custom-designed premix membrane emulsification system (FM0210/500M) equipped with a 2.8 μm pore SPG membrane was provided by Senhui Microsphere Tech Co., Ltd. (China). Dulbecco’s modified Eagle’s medium (DMEM) was purchased from Gibco (USA). Fetal bovine serum (FBS) and the penicillin–streptomycin complex were purchased from Shanghai XP Biomed Ltd. (China). The human nonsmall cell lung cancer cell line (A549) was purchased from the National Collection of Authenticated Cell Cultures.

2.2. One-Step Preparation and the Optimization of Blank Nonspherical Particles

The one-step method was adapted from our previous microparticle fabrication method, employing the OVAT methodology to investigate the possibility of fabricating uniform nanosized nonspherical PLGA particles. Briefly, as shown in Figure , 0.75 g of PLGA was dissolved in 12 mL of DCM to form the oil phase. Then, a certain amount of surfactant PVA and deformation inducer PBS were dissolved in 450 mL of deionized (DI) water as the water phase. Then, the oil phase was added dropwise to the aqueous phase under high-shear homogenization (19,200 rpm, 2 min), followed by triple extrusion through a premix membrane emulsification instrument (N₂ pressure: 0.8–1.2 MPa) to obtain emulsion droplets with uniform size. Afterward, the uniform emulsion was stirred magnetically at 700 rpm for 4 h for DCM evaporation and particle solidification. Then, PLGA nanoparticles were collected by centrifugation, washed five times with DI water, lyophilized, and stored at −20 °C until further use.

1.

One-step membrane emulsification protocol for the engineering of nonspherical PLGA nanoparticles.

2.3. Two-Step Preparation and Orthogonal Optimization of Nonspherical Particles

To resolve the contradiction between size reduction and deformation, a sequential two-step premix membrane emulsification strategy was developed to spatiotemporally decouple nanoemulsion formation from shape deformation processes, as shown in Figure . A similar PLGA oil phase was homogenized with a PVA-I solution (1.25% w/v, 75 mL) and extruded through the premix membrane emulsification equipment as above to obtain the primary emulsion. Then, the primary emulsion was transferred to a PVA-II solution containing PBS for droplet deformation. The other steps were similar to those of the one-step preparation method.

2.

Two-step membrane emulsification protocol for the engineering of nonspherical PLGA nanoparticles.

An orthogonal array L₉(3⁴) was employed for optimization in the two-step method. As shown in Table , four critical parameters, including PLGA concentration (A: 31.25, 62.50, 93.75 mg/mL), oil volume (B: 8, 4, 2 mL), PVA-II concentration (C: 1.25%, 1.75%, 2.25% w/v), and PBS concentration (D: 2.5, 5.0, 7.5 mM) were evaluated to identify the optimal parameters for nanorods.

1. Orthogonal Test Design.

Level	PLGA concentration (mg/mL)	Oil volume (mL)	PVA-II concentration (%)	PBS concentration (mM)
1	31.25	8	1.25	2.5
2	62.50	4	1.75	5.0
3	93.75	2	2.25	7.5

2.4. Two-Step Preparation of PTX-Loaded PLGA Nanorods and Nanospheres

To prepare PTX-loaded PLGA nanorods, 15 mg of PTX was added to the oil phase to prepare PTX-loaded PLGA nanorods through the optimal formulation. The preparation method was the same as that described in Section . In addition, PTX-loaded spherical PLGA nanoparticles were also prepared using the same procedure as nonspherical nanoparticles, with the exception that PBS was omitted from the PVA-II solution.

2.5. Morphology Characterization

The shape and surface morphology of PLGA particles were observed using a JSM-6700F (JEOL, Japan) scanning electron microscope (SEM). The major axes (length of particles, L) and minor axes (width of particles, W) of the particles were measured from ≥100 particles per batch through their SEM micrographs via ImageJ 1.48. The aspect ratio (AR) was calculated as AR = L/W. Larger AR values indicate a greater deformation degree. The rod yield was defined as the percentage of particles with AR ≥ 1.5 to indicate the deformation efficiency. All data were described as mean ± standard deviation (SD).

2.6. Deformation Factor Analysis

In the two-step method of orthogonal experiments, the impact of each process factor on the deformation degree was evaluated through mean (K) and range (R) analysis, which are described as follows

$$K_i=S_i/\mathrm{s}(i=1,2,3)$$ 1

$$R=\max (K_i)-\min (K_i)$$ 2

where S _i indicates the sum of the AR values at level i, and s is the number of occurrences of each level in one factor column in the orthogonal table (s = 3 in our experiments). Larger R values indicate greater factor significance. The optimal parameter combinations for uniform nanorods with higher AR could be obtained through K and R analysis.

2.7. Drug-Loading Efficiency and Encapsulation Efficiency

The PTX drug-loading efficiency (LE) and encapsulation efficiency (EE) of PLGA particles were detected by high-performance liquid chromatography (HPLC), as described previously. Briefly, 2 mg of the freeze-dried drug-loaded particles were sufficiently dissolved in 1 mL of acetonitrile under sonication for 30 min, followed by filtration through a 0.22 μm membrane filter prior to HPLC analysis. The PTX concentration was then quantified by an HPLC system (Agilent, USA) equipped with a Kromasil C18 column (4.6 mm × 150 mm, 5 μm) at a detection wavelength of 227 nm. The mobile phase consisted of distilled water and acetonitrile (40:60, v/v). The injection volume and flow rate were 20 μL and 1 mL/min, respectively. The column temperature was maintained at 30 ± 1 °C.

Then, the drug-loading efficiency (LE) and encapsulation efficiency (EE) of PLGA particles were calculated as follows

$$\mathrm{LE}\%=\frac{\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}\mathrm{o}\mathrm{f}\mathrm{P}\mathrm{T}\mathrm{X}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{a}\mathrm{p}\mathrm{s}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{n}\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{t}\mathrm{i}\mathrm{c}\mathrm{l}\mathrm{e}\mathrm{s}}{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}\mathrm{o}\mathrm{f}\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{t}\mathrm{i}\mathrm{c}\mathrm{l}\mathrm{e}\mathrm{s}}\times 100\%$$ 3

$$\mathrm{E}\mathrm{E}\%=\frac{\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}\mathrm{o}\mathrm{f}\mathrm{P}\mathrm{T}\mathrm{X}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{a}\mathrm{p}\mathrm{s}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{n}\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{t}\mathrm{i}\mathrm{c}\mathrm{l}\mathrm{e}\mathrm{s}}{\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l}\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}\mathrm{o}\mathrm{f}\mathrm{P}\mathrm{T}\mathrm{X}\mathrm{a}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{d}}\times 100\%$$ 4

2.8. In Vitro Cytotoxicity of PLGA Particles against A549 Cells

The MTT assay was employed to evaluate the cytotoxicity of the different formulations against A549 cells. A549 cells were cultured in DMEM supplemented with 10% FBS under a 5% CO₂ atmosphere at 37 °C. In brief, A549 cells were seeded into each well of 96-well plates at a density of 8 × 10³ cells/well and incubated for 24 h. Then, the medium was removed, and 100 μL of fresh medium containing PTX, spherical or nonspherical PLGA particles (with equivalent PTX concentrations of 0.01, 0.1, 1, and 10 μg/mL) was added (n = 4). Fresh medium was used as a negative control, and cell-free culture medium was used as a blank control. After incubation for 48 h, the medium was removed and replaced with 100 μL of fresh medium containing 0.5 mg/mL MTT, and again incubated for 4 h. Then, the medium was removed and replaced with 150 μL of DMSO. Optical density (OD) values were measured at 570 nm using a microplate reader (Varioskan LUX, Thermo, USA).

Cell viability was calculated as follows

$$\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{v}\mathrm{i}\mathrm{a}\mathrm{b}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{y}(\%)=\frac{{\mathrm{O}\mathrm{D}}_{\mathrm{g}\mathrm{r}\mathrm{o}\mathrm{u}\mathrm{p}\mathrm{s}}-{\mathrm{O}\mathrm{D}}_{\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{k}}}{{\mathrm{O}\mathrm{D}}_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}-{\mathrm{O}\mathrm{D}}_{\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{k}}}\times 100\%$$ 5

3. Results

3.1. Optimization of PLGA Nanorods via a One-Step Method

3.1.1. Influence of PBS Concentration

Our previous research demonstrated that PBS, a widely used inorganic salt, could induce deformation using an emulsion solvent evaporation method. Therefore, the impact of PBS concentrations (2.5 mM, 5 mM, and 10 mM) on the morphology of PLGA nanoparticles was first investigated. As shown in Figure , the AR of nanoparticles initially increased from 3.31 ± 2.32 (2.5 mM) to 4.85 ± 2.30 (5 mM) with an increase of PBS concentration, followed by a sharp decline to 1.51 ± 0.49 at 10 mM (Table S1). These results indicate that the optimal PBS concentration for nanorods was 5 mM, which was employed for subsequent studies.

3.

Influence of PBS concentration. SEM images of PLGA particles fabricated with different PBS concentrations of (A) 2.5 mM, (B) 5 mM, and (C) 10 mM. Scale bar: 1 μm.

3.1.2. Influence of PLGA Concentration

A previous study demonstrated that PLGA concentration can affect particle size due to its influence on the viscosity of the oil phase. Therefore, the influence of PLGA concentration on particle deformation was investigated. As shown in Figure and Table S1, lower concentrations of PLGA (25, 31.25 mg/mL) yield nanorods with comparable AR (3.75 ± 1.64 vs 3.88 ± 1.27). However, when the PLGA concentration was increased to a higher concentration of 62.50 mg/mL, the fabricated particles changed into irregular polygonal morphologies (AR = 2.15 ± 0.75). The results indicate that a lower concentration of PLGA was beneficial to deformation. On the other hand, the PLGA concentration could affect the particle yield per batch. Hence, a 31.25 mg/mL PLGA concentration was selected for subsequent investigation because of its optimal balance between deformation efficiency (AR = 3.88 ± 1.27) and particle yield.

4.

Influence of PLGA concentration. SEM images of PLGA particles fabricated with different PLGA concentrations of (A) 25 mg/mL, (B) 31.25 mg/mL, and (C) 62.50 mg/mL. Scale bar: 1 μm.

3.1.3. Investigation of PVA Concentration

PVA is widely used as a surfactant to stabilize the interface between the oil and water phases. Therefore, the influence of PVA concentration on particle deformation was investigated. As shown in Figure and Table S1, lower PVA concentrations of 0.5% and 1.0% produced rod-shaped nanoparticles with an AR of around 5.3. While a higher PVA of 2.0% achieved nanorods with higher AR (8.14 ± 4.66) despite decreased rod yield (64.04%) which means increased polydispersity. This result reflects the contradiction: reduced interfacial tension by increasing the PVA concentration could enhance deformation but promote droplet coalescence. Thus, in order to achieve a higher AR, a PVA concentration of 2.0% was chosen in the subsequent optimization.

5.

Influence of PVA concentration. SEM images of PLGA particles fabricated with different PVA concentrations of (A) 0.5%, (B) 1.0%, and (C) 2.0%. Scale bar: 1 μm.

3.1.4. Investigation of the Oil Phase Volume

As shown in Figure , the oil phase volume did have an impact on particle deformation. At 2 mL, slightly deformed particles (AR ≈ 2) with polygonal morphologies and porous structures were formed. Increasing the oil phase volume to 4 mL yielded many long filaments, a few polygons, and spheres (AR = 10.19 ± 5.14). An increase in the oil phase volume to 8 mL produced mixed morphologies (some worm-like particles, some spindle-shaped rods, and some spherical particles). Thus, a larger oil phase volume favored the production of particles with higher AR (like filaments), but increased polydispersity and decreased rod yield were observed (Table S1).

6.

Influence of the oil phase volume. SEM images of PLGA particles fabricated with oil phase volumes of (A) 2 mL, (B) 4 mL, and (C) 8 mL. Scale bar: 1 μm.

3.2. Two-Step Orthogonal Optimization

Although the one-step method could be used to prepare nanorods with smaller sizes, their shapes and sizes were not very uniform. For example, when optimizing the oil volume, 4 mL of the organic phase volume could fabricate nanorods with the highest AR of 10.19, yet the standard deviation was 5.14, and the rod yield was 72.38%, leading to an increased polydispersity. Thus, a two-step method was developed for preparing nonspherical PLGA nanoparticles, simultaneously enabling greater deformation, smaller particle size, better uniformity, and higher rod yield.

The orthogonal experimental design of a four-factor three-level case (L₉(3⁴)) was employed in a two-step method to comprehensively optimize the preparation process of uniform nonspherical nanoparticles. Table lists the 9 experimental factor combinations according to the orthogonal arrays with the help of the software “Orthogonality Experimental Assistant”. Figure A-I corresponds to the SEM images of particles fabricated by orthogonal test 1-9 in Table . The ARs of the particles, K ₁ , K ₂ , K ₃, and R, were calculated and are shown in Table . By comparing the R value and the K value of each factor, the primary and secondary influencing factors in the formulation, and the optimal level combination were determined.

2. Orthogonal Array Analysis of the Influence of Factors on the Aspect Ratio (AR).

Test number	PLGA concentration (mg/mL)	Oil phase volume (mL)	PVA-II concentration (%)	PBS concentration (mM)	Aspect ratio (AR)
1	31.25	8	1.25	2.5	7.40 ± 1.95
2	31.25	4	1.75	5.0	1.25 ± 0.12
3	31.25	2	2.25	7.5	1.12 ± 0.07
4	62.50	8	1.75	7.5	1.31 ± 0.19
5	62.50	4	2.25	2.5	1.24 ± 0.12
6	62.50	2	1.25	5.0	1.07 ± 0.04
7	93.75	8	2.25	5.0	1.67 ± 0.36
8	93.75	4	1.25	7.5	1.16 ± 0.07
9	93.75	2	1.75	2.5	1.10 ± 0.06
K 1	3.26	3.46	3.21	3.25
K 2	1.21	1.22	1.22	1.33
K 3	1.31	1.10	1.34	1.20
R	2.05	2.36	1.99	2.05
Optimal levels	A1	B1	C1	D1

7.

SEM images of PLGA particles fabricated in the orthogonal experiment. (A–I) shows the morphologies of the particles prepared in the orthogonal experiment 1-9 in Table . Scale bar: 1 μm.

3.2.1. Range Analysis of the Influencing Factors

As shown in Figure and Table , differently shaped nanoparticles with AR varying from 1.07 (test number 6) to 7.40 (test number 1) could be fabricated under different conditions. Specifically, nanoparticles in test numbers 3, 6, and 9 were fabricated in a spherical shape and smooth surface with an AR of approximately 1 (Figure C,F,I). Nanoparticles in test numbers 2, 5, and 8 were nearly spherical with porous structures (Figure B,E,H). The nanoparticles in test numbers 1, 4, and 7 were all deformed to varying degrees (Figure A,D,G). In detail, PLGA nanoparticles in test number 1 were stretched into long nanorods with an AR of 7.40 ± 1.95 (Figure A and Table ), which showed the largest degree of deformation. Orthogonal tests 4 and 7 fabricate nanoparticles with irregular polygonal morphologies (Figure D,G) and ARs of 1.31 and 1.67 (Table ), respectively. The values of K and R are shown in Table .

The significance of each factor on nanoparticle deformation can be clearly observed by comparing the R value for each factor. As shown in Table , the orthogonal design identified oil phase volume (R = 2.36) as the dominant deformation determinant, followed by PLGA concentration (R = 2.05), PBS concentration (R = 2.05), and PVA-II concentration (R = 1.99).

3.2.2. Effect of a Single Parameter on Deformation and Optimal Parameters

To gain insight into the influence of a single parameter on particle deformation, the main effect plots for AR versus different factors/levels were analyzed. As shown in Figure , the K value of AR was first decreased from 3.26 to 1.21 as the PLGA concentration increased, but then slightly increased to 1.31 at the highest PLGA concentration of 93.75 mg/mL. Similar to the profile of PLGA concentration, with an increase of PVA-II concentration from 1.25% to 1.75%, the K value was decreased from 3.21 to 1.22, but then increased to 1.34 at a PVA concentration of 2.25%. Increasing the PBS concentration from 2.5 mM to 7.5 mM decreased the K value of AR from 3.25 to 1.20. In terms of the oil phase volume, 2 mL and 4 mL yielded nonspherical nanoparticles with similar K values of 1.10 and 1.22, respectively; nevertheless, 8 mL oil phase volume fabricated nanoparticles with the highest K value of 3.46. Finally, the optimal level combination of the four factors was determined as A₁B₁C₁D₁, that is, PLGA concentration of 31.25 mg/mL, oil phase volume of 8 mL, PVA-II concentration of 1.25%, and PBS concentration of 2.5 mM, which corresponded to orthogonal test 1. These optimal parameters yielded more uniform nanorods (AR = 7.40 ± 1.95) with a narrow distribution of major and minor axes (Figure S1) and a high rod yield of 100.00% (Table ), representing a 27.62% improvement in the rod fabrication yield compared with the one-step method. Moreover, the minor axis of the nanorods in test 1 was 331.11 ± 85.90 nm.

8.

Response plots for the average effect of the parameters on the K values of AR for the L₉(3⁴) orthogonal design.

3. Major and Minor Axes, and Rod Yield of PLGA Particles Fabricated in Orthogonal Test.

Test number	Major axes (nm)	Minor axes (nm)	Rod yield (%)
1	2362.24 ± 596.07	331.11 ± 85.90	100.00
2	930.48 ± 289.96	743.46 ± 212.68	0.00
3	471.87 ± 249.19	428.39 ± 241.79	0.00
4	1022.93 ± 312.82	784.06 ± 224.2	13.73
5	1363.19 ± 244.81	1101.39 ± 200.35	6.00
6	1020.01 ± 378.76	961.93 ± 374.32	0.00
7	1514.92 ± 383.45	914.76 ± 184.44	68.63
8	1139.85 ± 291.41	1037.54 ± 267.97	0.00
9	1604.63 ± 283.66	1384.03 ± 210.87	0.00

3.2.3. Surface Morphology Control

When optimizing the shape of the nanoparticles by the orthogonal experiment design, it is observed that the orthogonal conditions could simultaneously modulate the surface morphology of the nanoparticles (Figure ). It can be observed that the spherical particles exhibited either smooth (tests 3, 6, and 9) or porous (tests 2, 5, and 8) surfaces, while the deformed nanoparticles maintained smooth surfaces (tests 1, 4, and 7). Then, we further adopted range analysis to investigate the influence of the four parameters on the surface morphology. To simplify the definition of surface morphology, nanoparticles with porous surfaces were scored as 1, and nanoparticles with a smooth surface were scored as 0. Table S2 shows the influence of different factors on the surface morphologies. Range analysis identified the oil phase volume as the key driver for porosity generation (Table S2), and 4 mL of the oil phase volume was preferable to fabricate porous nanoparticles. This demonstrates that the oil phase volume enabled dual control over the shapes and surface morphologies.

3.3. Characterization of PTX-Loaded PLGA Nanorods and Nanospheres

As shown in Figure , PTX-loaded PLGA nanospheres (PTX-NSs) and nanorods (PTX-NRs) were successfully prepared through optimal formulation. The encapsulation of PTX did not interfere with the rod shape. The LE and EE of nonspherical PLGA nanoparticles were 4.75 ± 0.22% and 79.11 ± 3.70%, respectively, while those of spherical PLGA particles were 3.32 ± 0.13% and 55.28 ± 2.22%, respectively.

9.

SEM images of PTX-loaded PLGA nanospheres (PTX-NSs) and nanorods (PTX-NRs).

3.4. In Vitro Cytotoxicity of PTX-Loaded PLGA Nanorods and Nanospheres

The cytotoxicity of PTX-NRs and PTX-NSs was evaluated on the A549 cell line across a range of concentrations (0.01, 0.1, 1, and 10 μg/mL). As shown in Figure , exposure to free PTX, rod-shaped PTX-NRs, or spherical PTX-NSs resulted in a dose-dependent reduction in cell viability. At most concentrations, both PTX-loaded nanoparticles (PTX-NRs and PTX-NSs) exhibited higher cytotoxicity against A549 cells than free PTX, highlighting the advantage of the nanocarrier system. Specifically, compared with the PTX-NSs group, treatment with PTX-NRs caused a higher cytotoxicity at the tested concentration, reaching statistical significance at a concentration of 1 μg/mL.

10.

Cell viability following treatment with free PTX, PTX-NSs, and PTX-NRs at varying PTX concentrations.

4. Discussion

An analysis of the literature in the past 10 years showed that only about 0.7% (median) of the injected dose of nanoparticles was delivered to the tumor site, which greatly weakens the advantages of nanotechnology. On the basis of this, modifying the physicochemical properties of nanoparticles to enhance their accumulation in tumors has been a hot topic in recent years, among which the shape of drug carriers has received extensive attention. , More and more evidence has shown that nonspherical particles have unique advantages compared with their spherical counterparts. , For example, Chauhan et al. demonstrated that nanorods surpassed nanospheres in tumor penetration capabilities with a penetration speed 4.1 times faster and a penetration volume 1.7 times greater. However, one remaining challenge is to develop a polymeric nanoparticle fabrication technique that simultaneously enables control over shape and uniformity, nanoscale dimensions, simple processes, and high throughput.

Emulsion techniques can have high throughput and simple processes, but often fail to control shape and uniformity. To address these shortcomings, our previous study developed a PBS deformation inducer combined with the premix membrane emulsification technique (one-step method) to successfully prepare uniform nonspherical PLGA particles. Nevertheless, only micron-scale nonspherical particles were obtained. Thus, in the present research, an SPG membrane with a smaller pore size (2.8 μm) was chosen in the premix emulsification technique to explore the possibility of preparing uniform nonspherical PLGA nanoparticles. The whole preparation process could be divided into three stages: (1) formation of spherical droplets, (2) deformation of droplets by stirring, and (3) solidification of nonspherical particles. The deformation of spherical droplets can be explained by the action of fluid mechanics. Briefly, the droplet deformation can be determined by two dimensionless concepts, Ca and M, which can be calculated as follows

$$\mathit{Ca}=\gamma {\eta }_s\alpha /\mathrm{\Gamma }$$ 6

$$M={\eta }_d/{\eta }_s$$ 7

where γ is the shear rate, η _d and η _s are the viscosities of the dispersed phase and the continuous phase, α is the droplet size, and Γ is the interfacial tension between the dispersed and continuous phases. A higher Ca value or lower M value in the system indicates a more serious deformation.

In the present research, two different methods (one-step method/two-step method) were used to test the possibility of preparing nonspherical nanoparticles by optimizing vital influencing factors, including the PBS concentration, PLGA concentration, PVA concentration, and oil phase volume.

First, the OVAT methodology was adopted in optimizing the one-step preparation system, similar to our previous research. The results show that a PLGA concentration lower than 31.25 mg/mL was better for droplet deformation (Figure ) because a low PLGA concentration indicates a low viscosity in the disperse phase, which means a low M value and high deformation. Moreover, a medium PBS concentration of 5 mM was beneficial for deformation, which was similar to our previous research. Our previous research measured the PBS-induced changes in the interfacial tension Γ. It was found that when PBS concentration was lower than 5 mM, the interfacial tension decreased sharply with an increase of PBS, thus increasing the Ca value and promoting deformation. At PBS concentrations higher than 5 mM, the interfacial tension decreased slowly with an increase of PBS. Nevertheless, a higher concentration of PBS may increase the electrolytes in the emulsion, which can form a double electric layer to stabilize PLGA droplets and inhibit further deformation. At last, due to the competition between decreased Γ and increased electrolytes, a medium PBS concentration of 5 mM was beneficial for deformation. Similarly, as a surfactant agent, PVA can simultaneously decrease the interfacial tension Γ and increase the viscosity of water phase η _s , both of which can increase the Ca value and decrease the M value. Theoretically, a higher PVA can result in a smaller emulsion size and heavier deformation. As shown in Figure , the particles indeed became smaller with higher AR at higher PVA concentrations, but many irregular polygonal and spherical particles appeared at the same time and the rod yield sharply decreased (Table S1). This abnormal phenomenon was probably due to the unstable emulsion and secondary droplet rupture at higher PVA concentrations. In order to fabricate nanorods with a higher AR, a higher concentration of PVA was ultimately chosen as the water phase for further optimization. When optimizing the oil phase volume, it could be seen that a higher oil volume was favorable for deformation. At a smaller oil volume (Figure A), the solvent of the droplets evaporated too quickly, leading to the premature solidification of the droplets before they can be adequately stretched, so the produced particles exhibited a porous structure and irregular shape. In contrast, a larger oil volume can afford enough time for droplet deformation, but higher polydispersity could occur at the largest oil volume due to the unstable interfacial condition in the dense emulsion, which can cause coalescence of droplets. Thus, at the end of the OVAT optimization of the one-step method, nanorods with higher AR (AR ≈ 10) could be prepared, but the rod yield (72.38%) and uniformity were still unsatisfactory.

The difficulty to prepare uniform nanorods by the one-step method might be due to the contradiction between the requirement of smaller particles in nanosize and the demand for severe deformation. As indicated in previous research by our group and Omolola Eniola-Adefeso’s group, , the requirement of smaller particles means a smaller α value, resulting in a smaller Ca value, which will cause weaker deformation. The breakthrough came through the temporal decoupling of the emulsification process and deformation process in the two-step protocol. This strategic separation allows the independent optimization of α (via extrusion through SPG membranes) and Ca (through PBS/PVA modulation in the second continuous phase), effectively resolving the critical size-deformation trade-off (Figure ). In this way, the value of α remained unchanged, but the value of surface tension Γ was further reduced in the second continuous phase, so that the value of Ca increased, and the deformation was more efficient. Thus, changing from the one-step method to the two-step method can balance the contradiction between α and Γ by changing the PVA concentrations in the two water phases.

When optimizing the two-step method for nanorods, an orthogonal experimental design was employed to investigate the effect of formulation factors on particle deformation. By analyzing the K values, the influence of the four parameters on deformation can be obtained. Similar to the influence in the one-step method, the lower PLGA concentration of 31.25 mg/mL was favorable for particle deformation. In contrast to the one-step method, the optimum PBS concentration in the two-step method was 2.5 mM, lower than 5 mM in the one-step method, which can be attributed to the higher PVA-II concentration (>1.25%) used in the two-step method. Both PBS and PVA can reduce the interfacial tension. Thus, for a higher PVA-II concentration in the two-step method, a smaller amount of PBS was enough to reduce the interfacial tension to promote deformation. The optimal PVA concentration was the lowest, which was different from that in the one-step method because the lowest PVA-II concentration in the two-step method (1.25%) was already higher than that in the one-step method (0.5% and 1.0%). Moreover, the inverse relationship between PVA-II concentration and AR reduction likely originated from the surfactant-mediated secondary rupture of deformed droplets at elevated PVA levels, which was a possible mechanism consistent with the increased polydispersity observed in one-step formulations under high PVA conditions. At last, similar to the pattern in the one-step method, a larger oil volume was beneficial for deformation. Moreover, the range analysis showed that the oil phase volume with the highest R value was the dominant deformation determinant. In the premix emulsification technique, a larger oil volume tends to generate slightly larger emulsion droplets, which corresponds to a relatively higher Ca. Beyond modulating the initial droplet size, the oil phase volume may also affect the droplet solidification kinetics by altering the solvent evaporation rate and total interfacial area of the droplets. Previous studies have demonstrated that the rate of droplet solidification and solvent removal determines the time window during which droplets undergo stretching and deformation, thereby shaping the final particle morphology. , Therefore, the impact of the oil phase volume on particle morphology may result from the combined effects of droplet size, solidification kinetics, and hydrodynamic droplet deformation, rather than being controlled by a single factor.

Compared with the one-step method, particles optimized by the two-step method were more uniform with smaller nanorods, which was due to the elimination of the contradiction between the smaller droplet size and higher deformation. Notably, the two-step method achieved 100% rod yield with minor axes lower than 500 nm, which means a 27.62% improvement over the one-step method. Therefore, the two-step premix membrane emulsification technique was preferable to fabricate uniform nonspherical nanoparticles. Figure S2 visually illustrates the mechanism underlying the improved deformation achieved by the two-step strategy. In the one-step process, droplet formation and deformation occur simultaneously within the same continuous phase. Consequently, the formulation conditions required to produce uniform nanoscale droplets simultaneously define the hydrodynamic environment for the droplet deformation. According to the definition of the capillary number Ca = γη _s α/Γ, decreasing the droplet size inevitably reduces the characteristic droplet dimension (α) and thus lowers Ca. ,, Therefore, in the one-step process, producing small droplets and achieving efficient deformation are inherently conflicting objectives. Conversely, modifying the continuous phase to increase Ca may simultaneously affect droplet formation and emulsion stability. In contrast, as shown in Figure S2, the two-step process temporally decouples droplet formation from droplet deformation. In the first step, premix membrane emulsification is used solely to generate monodisperse nanoscale droplets, thereby fixing the droplet size (α) prior to deformation. The preformed droplets are subsequently transferred into a second continuous phase containing PBS and PVA, where the interfacial properties and rheological conditions can be independently tuned without affecting droplet formation, thereby providing a more favorable environment for droplet deformation. During this stage, the droplet size (α) remains essentially unchanged, while PBS and PVA further reduce the interfacial tension (Γ), thereby increasing Ca and creating a more favorable hydrodynamic environment for droplet deformation. , Accordingly, compared with the optimized one-step process, the two-step process achieved a 100% rod-shaped particle yield and produced more uniform nanorods. These results suggest that the temporal separation of droplet formation and droplet deformation may prolong the effective deformation window and enable more effective regulation of the key hydrodynamic parameters governing droplet deformation.

On the other hand, when there are more than two factors involved in one experiment, orthogonal experimental design is better than OVAT optimization because it can powerfully and effectively achieve an optimal combination of factors in a relatively small number of tests. , It should be noted that several optimal factor levels appeared at the boundary of the investigated range, suggesting that further improvement could be achieved by expanding the experimental space. To systematically and efficiently investigate a broader optimal experimental space, artificial intelligence (AI)- or machine-learning-assisted experimental optimization strategies can be adopted in future research.

Moreover, when optimizing the shape of nanoparticles by orthogonal experiment design, it is very interesting to find that the surface morphology of the particles can also be adjusted (Figure ). Range analysis can clearly show that the oil phase volume had the main influence on the surface morphology, and a medium volume of 4 mL was preferable to fabricate porous nanoparticles (Table S2). The porous structures observed in Figure B,E, and H may be associated with solvent evaporation and mass transfer during droplet solidification. Previous studies have demonstrated that during the solidification process, the evaporation rate of dichloromethane affects solvent exchange and water diffusion within the droplets, thereby altering the internal morphology of the resulting particles. In addition, the continuous evaporation of dichloromethane gradually increases the polymer concentration within the droplets. Once the critical condition for phase separation is reached, a polymer-rich phase and a polymer-lean phase emerge. The polymer-lean phase progressively expands and coalesces, eventually forming porous structures after complete particle solidification and solvent removal. Therefore, we speculate that the porous particles observed in this study may result from the combined effects of solvent evaporation kinetics, intradroplet mass transfer, polymer solidification, and localized liquid–liquid phase separation.

After optimization, the chemotherapeutic drug PTX was successfully encapsulated into PLGA nanorods through this method, demonstrating the feasibility of the two-step premix membrane emulsification technique for producing uniform drug-loaded nanorods. Furthermore, PTX-NRs exhibited enhanced anti-tumor efficacy against A549 cells, indicating the great potential of nonspherical nanorods as promising PTX delivery systems.

In summary, by using the two-step premix membrane emulsification technique, uniform nonspherical nanoparticles can be successfully fabricated. This method has obvious advantages over the methods described in the literature due to its simple operation process, high yields of rod-shaped particles, controllable particle size, and good uniformity. It is of significant importance that the simple and controllable preparation method developed here will allow more investigators to have quick and inexpensive access to nonspherical drug delivery systems and thus rapidly advance their potential application in nanomedicine.

Nevertheless, the present study has several limitations. First, direct characterization of the droplet deformation process was not implemented. Nanoscale emulsion droplets are too small, making real-time observation of their dynamic deformation infeasible with conventional optical microscopy. Moreover, our analysis of the deformation mechanism mainly relies on the Ca and M values. Nevertheless, it cannot be neglected that the droplet size may influence deformation not only through Ca but also via alterations in the droplet specific surface area, interfacial mass transfer, and solvent evaporation kinetics. Furthermore, the detailed mechanism governing the formation of porous particles remains to be elucidated. In future research, we will employ high-resolution in situ characterization techniques, such as high-resolution confocal laser scanning microscopy and cryoelectron microscopy, to systematically investigate the dynamic evolution of emulsion droplets, so as to obtain more direct evidence to clarify the mechanisms behind droplet deformation and pore generation.

5. Conclusion

In this study, the premix membrane emulsification technique was used to fabricate and optimize uniform nonspherical PLGA particles in the nanosize range. Both the one-step method and the two-step method were optimized to prepare nanorods. However, the two-step method was more favorable to prepare smaller and uniform nanorods due to the spatiotemporal decoupling of emulsification and deformation processes. Moreover, the orthogonal experimental design could provide robustness to parameter screening for optimizing deformation. Specifically, the shape, size, and surface morphology of these particles could be regulated by careful manipulation of key process parameters, including PLGA concentration, PBS concentration, PVA concentration, and oil phase volume in our system. In general, this study provides a simple and controllable technique for the preparation of uniform nonspherical particles at the nanoscale, with subsequent potential applications in drug delivery for tumor therapy.

Any additional information required to reanalyze the data reported in this paper is available from the corresponding author upon request.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c07438.

• Particle dimensions, aspect ratios, and rod yields from one-variable-at-a-time (OVAT) experiments (Table S1); orthogonal array analysis of the effects of process factors on particle surface morphology (Table S2); statistical distributions of the major axis, minor axis, and aspect ratio of the optimized nanorods prepared by the two-step method (Figure S1); schematic illustration of the proposed droplet deformation mechanism underlying the one-step and two-step premix membrane emulsification strategies (Figure S2) (PDF)

⊥.

Z.L., L.L., and Y.Z. contributed equally to this work. Conceptualization: J.W., G.W., and Q.F. Methodology: Z.L., L.L., Y.Z., X.W., and B.F. Investigation: Z.L., L.L., Y.Z., and Q.F. Software: Q.H., X.W., and H.L. Writingoriginal draft: Z.L., L.L., M.X., and Q.F. Writingreview and editing: Y.Z., J.W., G.W., and Q.F. Funding acquisition: L.L., M.X., X.W., and Q.F. Resources: G.W., J.W., and Q.F. Supervision: G.W., and Q.F. All authors have reviewed the manuscript.

The authors declare no competing financial interest.