Release mechanisms of PLGA microparticles prepared using a microfluidics device or a beaker

Abstract

The aim of this study was to better understand the release mechanisms of poly(lactic-co-glycolic acid) (PLGA) microparticles prepared via emulsification - solvent extraction/evaporation using a “classical beaker” vs. a “microfluidics device”. Ibuprofen-loaded microparticles were studied by optical microscopy, SEM, X-ray powder diffraction, X-ray μCT and drug release measurements from single microparticles in well agitated phosphate buffer pH 7.4 or agarose gel (mimicking living tissue). The use of a microfluidics device facilitated the preparation of microparticles with a less broad size distribution. However, in addition to the microparticle size, the inner system structure was found to be also of utmost importance for the resulting drug release kinetics in this case. Interestingly, even microparticles with similar size, composition and inner & outer structure exhibited a broad spectrum of individual drug release patterns. This was true, irrespective of the type of preparation method and experimental release set-up, and could be explained as follows: The investigated microparticles were characterized by a continuous inner pore network and an initially smooth & non-porous surface. Drug release set on as soon as: (i) the pore network got direct access to the release medium (e.g., due to a “weak point” in the PLGA surface layer), or (ii) substantial system swelling started (after a lag-time of several days). Importantly, each microparticle had its own, specific structure, which determined “its way” to release the drug. Furthermore, the experimental conditions were found to be of key importance: The presence of a surrounding agarose gel protected the microparticles from damage caused by convective fluid flow, and hindered microparticle swelling, thus, slowing down drug release.

Graphical abstract

1. Introduction

Poly(lactic-co-glycolic acid) (PLGA)-based microparticles offer an interesting potential for parenteral controlled drug delivery (Park et al., 2021; Lagreca et al., 2020; Park, 2025), because: (i) A large spectrum of drug release rates can be provided during periods ranging from a few days to several weeks. (ii) PLGA is completely biodegradable and biocompatible. (iii) A variety of techniques can be used to produce them, also at industrial scale. (iv) They can be relatively easily administered to the patient. Several drug products have been commercially available for decades (Park et al., 2019). However, the development of PLGA microparticles for controlled drug delivery is highly challenging and generally based on time-consuming and cost-intensive series of trial-and-error experiments. This can in great part be attributed to the complexity of the underlying drug release mechanisms and often poor understanding of how the composition and morphology of the microparticles determines drug release.

Different physico-chemical phenomena can be involved in the control of drug release from PLGA-based microparticles (Fredenberg et al., 2011), such as water penetration into the system upon contact with the release medium, drug particle dissolution, limited drug solubility effects, diffusion of dissolved drug molecules/ions through the microparticles, hydrolytic cleavage of the ester bonds, polymer swelling (Gasmi et al., 2015a, Gasmi et al., 2015b), microparticle disintegration, the potential creation of local drops in micro-pH (Brunner et al., 1999; Fu et al., 2000; Liu et al., 2008), autocatalytic effects (Ford Versypt et al., 2013), the closure of initially existing surface pores (Kang and Schwendeman, 2007), plasticizing effects of water (Blasi et al., 2005) as well as polymer-drug interactions (Blasi et al., 2007). Depending on the type of drug and PLGA (e.g., its polymer molecular weight and type of end groups), system composition (e.g., drug loading and potential presence of other excipients) and type of preparation technique (which can strongly impact the microparticles' inner and outer morphology), the relative importance of these mechanisms can differ. In addition, their contribution to the control of drug release can change over time.

Generally, the resulting drug release patterns from PLGA-based microparticles are mono-, bi-, or tri-phasic (Tamani et al., 2021; Yoo and Won, 2020). The initial burst phase is often attributable to drug, which has direct access to the system's surface. The second drug release phase (with a more or less constant drug release rate) might be explained by the local swelling of surface-near regions (undergoing more rapid degradation due to the direct contact with the release medium), rendering drug located in these regions much more mobile (Tamani et al., 2019). The third (and final) drug release phase has been explained in several cases by a substantial overall microparticle swelling: The system is transformed into a highly swollen polymer gel (Gasmi et al., 2016). This fundamental overall system swelling starts as soon as the PLGA network has reached a critical polymer molecular weight: Initially, the PLGA chains are relatively long and hydrophobic. Thus, the amounts of water, which can penetrate into the system are limited. However, upon each ester bond cleavage, two new hydrophilic groups are created (a -OH group and a -CCOH group). Hence, the microparticles become more and more hydrophilic over time. In addition, the degree of polymer chain entanglement decreases (rendering the networks mechanically less stable) and more and more water-soluble degradation products are generated, creating a steadily increasing osmotic pressure inside the microparticles. At a certain time point, the microparticles are sufficiently hydrophilic and mechanically fragile so that the osmotic pressure can attract substantial amounts of water into the system, fundamentally changing the conditions for the release of the remaining drug: The latter becomes much more mobile. It has also been reported that the presence of a surrounding hydrogel (mimicking living tissue) (Li et al., 2021; Sun et al., 2017; Ye et al., 2012) can sterically hinder this substantial system swelling, resulting in slower drug release (Lefol et al., 2023). Please note that in certain cases, all drug is already released in the first and/or second drug release phase(s), leading to mono- or bi-phasic release patterns. Also, if no drug has “early access” to the microparticles' surface, the initial burst phase might be negligible. The relative importance of the different drug release phases can strongly depend on the microparticles' (inner and outer) morphology. The latter is to a large extent determined by the preparation method.

A variety of techniques can be used to prepare PLGA-based microparticles for controlled drug delivery (Wischke and Schwnedeman, 2008), including for instance emulsification – solvent extraction/evaporation (O' Donnell and McGinity, 1997; Otte et al., 2022; Sharifi et al., 2020; Zhang and Bodmeier, 2022), spray-drying (Wan and Yang, 2016; Otte and Park, 2022), the use of supercritical CO₂ or milling. In the case of “emulsification – solvent extraction/evaporation”, simple or multiple emulsions can be used to encapsulate the drug. The emulsions can be created using different equipment, including a “classical beaker” or microfluidics device (Keohane et al., 2014; Yonet-Tanyeri et al., 2022). The polymer is dissolved in an organic phase. Upon contact with an outer aqueous phase, the organic solvent partially partitions into the water and evaporates once it reaches the water-air interface (Katou et al., 2008). More organic solvent follows and at a certain time point, its amount in the organic phase is insufficient to dissolve the PLGA, which starts precipitating: The microparticles are forming. The conditions under which the droplets are created during emulsification are decisive for their size, which determines the size of the resulting microparticles. In addition, the conditions under which the solvent is extracted into the outer aqueous phase are critically important for the rate at which the polymer precipitates and, thus, for the inner and outer morphology of the resulting microparticles. Since the flow of the liquids is better controlled in a microfluidics device compared to a “classical beaker”, the size distribution of the generated droplets is narrower, leading to a more homogeneous microparticle size. Since the latter can be expected to impact drug release (e.g., determining the lengths of diffusion pathways), more reproducible drug release patterns might be expected when using a microfluidics-assisted technique to prepare PLGA microparticles compared to a “classical beaker method” (Berkland et al., 2002). Furthermore, the inner and outer microparticle morphology might show less inter-particle variability when the systems are formed using a microfluidics device compared to a “classical beaker”, resulting in more homogenous drug release patterns.

The aim of this study was to prepare ibuprofen-loaded PLGA microparticles using the “classical beaker method” and a “microfluidics-assisted technique” and thoroughly characterize the systems. Importantly, drug release was studied from single microparticles and intentionally similarly sized systems were selected: The idea was not to study the effects of the microparticle size (or width of its distribution) on drug release, but to investigate whether a narrow microparticle size distribution can guarantee highly reproducible drug release patterns (or whether the microparticle morphology also plays a key role). Ibuprofen release was monitored in well agitated phosphate buffer pH 7.4 as well as upon inclusion into agarose gel (to mimic living tissue). The observed release kinetics were explained based on a thorough characterization of the microparticles using optical microscopy, SEM, X-ray powder diffraction and X-ray μCT.

2. Materials and methods

2.1. Materials

Poly (D,L lactic-co-glycolic acid) (PLGA, 50:50 lactic acid:glycolic acid; Resomer RG 503H; Evonik, Darmstadt, Germany); ibuprofen (BASF, Ludwigshafen, Germany); polyvinyl alcohol (Mowiol 4–88; Sigma-Aldrich, Steinheim, Germany); agarose (genetic analysis grade), potassium dihydrogen orthophosphate and sodium hydroxide (Acros Organics, Geel, Belgium); acetonitrile and dichloromethane (VWR, Fontenay-sous-Bois, France).

2.2. Microparticle preparation

Ibuprofen-loaded microparticles were prepared using an oil-in-water (O/W) solvent extraction/evaporation technique using two types of equipment:

Beaker method: The organic phase was a solution of PLGA and ibuprofen in 4 mL dichloromethane (850 mg/150 mg, 800 mg/200 mg, 600 mg/400 mg, 500 mg/500 mg polymer/drug for 15, 22, 36 and 48 % practical drug loading). This organic phase was emulsified into 2.5 L of an aqueous solution of poly(vinyl) alcohol (PVA) (0.25 % w/w) under stirring for 30 min (900 rpm, Eurostar power-b; Ika, Staufen, Germany). Upon dichloromethane extraction/evaporation, the polymer precipitated, trapping the drug. The formed microparticles were hardened by adding 2.5 L of the same outer aqueous PVA solution and further stirring for 4 h at 700 rpm (Eurostar power-b). Microparticles were separated by filtration (Nylon filter, 0.45 μm, 13 mm; GE Healthcare Life Sciences Whatman, Kent, UK), washed with demineralized water, and freeze-dried for 3 d (Christ Alpha 2–4 LSC+; Martin Christ, Osterode, Germany). Drug-free microparticles were prepared accordingly, without adding ibuprofen.

Microfluidics-assisted technique: 200 mg/50 mg, 175 mg/75 mg, 150 mg/100 mg, 125 mg/125 mg PLGA/ibuprofen were dissolved in 10 mL dichloromethane (for 15, 22, 36 and 48 % practical drug loading). This organic phase was emulsified into an aqueous PVA solution (1 % w/w) in an OB1 MK3+ microfluidics device, equipped with the ESI – 3.05.02 Smart Interface software (Elveflow, Paris, France). A “T-junction” was used, the flow rates were as follows: 20 μL/min (aqueous phase), 10 μL/min (organic phase). The emulsion was recovered in 2 L aqueous PVA solution (0.25 % w/w), followed by stirring for 2 h at 100 rpm. Microparticles were separated by filtration, washed with demineralized water, and freeze-dried for 3 d (Christ Alpha 2–4 LSC+). Drug-free microparticles were prepared accordingly, without adding ibuprofen.

2.3. Optical microscopy

Microscopic pictures of microparticles were taken using a Nikon SMZ-U apparatus (Nikon, Tokyo, Japan), equipped with an AxioCam ICc1 camera and the Axiovision Zeiss Software (Carl Zeiss, Jena, Germany).

2.4. Practical drug loading

The practical drug loading was determined by dissolving approximately 5 mg microparticle samples in 5 mL acetonitrile, followed by filtration (PVDF syringe filters, 0.22 μm; GE Healthcare). The drug content of the solution was determined by HPLC-UV analysis using an Ultimate 3000 Series HPLC apparatus, equipped with an LPG 3400 SD/RS pump, an autosampler (WPS-3000 SL) and a UV–Vis detector (VWD-3400RS) (Thermo Fisher Scientific, Waltham, USA). The mobile phase was a 67:33 (v:v) mixture of phosphate buffer pH 6.8 (50 mM KH₂PO₄ and 22.5 mM NaOH) and acetonitrile. Ten μL samples were injected into a C18 reversed-phase column (Gemini 5 μm; 110 Å; 150 × 4.6 mm; Phenomenex, Le Pecq, France). The detection wavelength was 225 nm, the flow rate was 1 mL/min.

2.5. In vitro drug release

Ibuprofen release into phosphate buffer pH 7.4 (USP 42) was measured from single microparticles in 96-well standard microplates (Carl Roth, Karlsruhe, Germany) using the following two experimental setups:

In well-agitated bulk fluid: Microparticles were introduced into the wells (1 microparticle per well). Each well was filled with 200 μL phosphate buffer pH 7.4 and sealed with aluminum foil. The microplate was horizontally shaken at 80 rpm and kept at 37 °C (GFL 3033; Gesellschaft fuer Labortechnik, Burgwedel, Germany). At predetermined time points, the entire bulk fluid was carefully replaced by fresh release medium using a Hamilton syringe (Microlite #710, 100 μL; Hamilton, Bonaduz, Switzerland). The withdrawn samples were analyzed for their drug contents by HPLC-UV analysis, as described in section 2.4. (injecting 20 μL samples). Perfect sink conditions were provided throughout the observation periods.

In agarose gel: 0.5 % w/w agarose gel was prepared by dissolving the polysaccharide in phosphate buffer pH 7.4 (USP 42) at 135 °C under stirring for 1 h (400 rpm, IKA RCT standard; IKA-Werke, Staufen im Breigsau, Germany). The solution was poured into the wells (50 μL per well), followed by cooling to room temperature (causing sol-to-gel-transition). A single microparticle was placed in the middle of the gel in each well. Another 50 μL agarose solution was carefully poured into each well (at a temperature close to gel formation), followed by cooling to room temperature. One hundred μL phosphate buffer pH 7.4 were added on top of each gel. The wells were sealed with aluminum foil and horizontally shaken at 80 rpm and 37 °C (GFL 3033). At predetermined time points, a Hamilton syringe was used to carefully replace the entire bulk fluid by fresh release medium. The withdrawn samples were analyzed for their drug contents by HPLC-UV analysis, as described above for the well-agitated bulk fluid method. Perfect sink conditions were provided in the phosphate buffer pH 7.4 (to which the agarose gel was exposed) throughout the observation periods.

2.6. Microparticle swelling

Microparticles were treated as for the in vitro drug release measurements described in section 2.5. At predetermined time points, pictures of the microparticles were taken using an Axiovision Zeiss Scope-A1 microscope, equipped with an AxioCam ICc1 camera (in the case of microparticles prepared with the beaker method) or an Axiocam 305 colour camera (in the case of microparticles prepared with the microfluidics method) (there was no specific scientific reason for the difference in camera, the latter was “upgraded” at a certain time point). The diameters of the microparticles were determined using the Axiovision Zeiss Software.

2.7. X ray powder diffraction

X-ray powder diffraction analyses were performed with a Panalytical X'Pert Pro diffractometer, equipped with a Cu X-ray tube (λ = 1.54 Å) and Hilgenberg glass capillaries (diameter 0.7 mm) in transmission mode with the X'Celerator detector. The diffractograms were recorded from 4 to 60° (2θ, 0.0167° steps).

2.8. Scanning electronic microscopy (SEM)

The microparticles' internal and external morphology before exposure to the release medium was studied using a JEOL Field Emission Scanning Electron Microscope (JSM-7800F, Tokyo, Japan). Samples were fixed with a ribbon carbon double-sided adhesive tape and sputter coated with a thin chrome layer. Cross-sections of the microparticles were obtained after inclusion into water-based glue (UHU, Bolton Group, Buehl, Germany). After drying for 48 h, slices of microparticles were manually cut with a razor blade.

2.9. X-ray microcomputed tomography (X-ray μCT)

X-ray μCT analysis was performed using a SkyScan 1172 micro-CT scanner (Bruker, Kontich, Belgium) to characterize the inner structure of the microparticles (before exposure to the release medium) in a non-invasive manner. Samples were placed on a polystyrene support. A camera array with 4000 × 2672 was used, resulting in a resolution of 2.26 μm. The rotation step was 0.25 degrees.

3. Results and discussion

The aim of this study was to prepare ibuprofen-loaded PLGA microparticles by emulsification – solvent extraction/evaporation using the classical “beaker method” and a “microfluidics-assisted” technique. The drug loading and microparticle size were intentionally kept similar. The key properties of the systems were determined, in particular the resulting drug release kinetics under different conditions: upon exposure to a well agitated bulk fluid and upon inclusion into an agarose gel.

Fig. 1 illustrates the two manufacturing processes. Briefly, the polymer and drug were dissolved in dichloromethane. This organic phase was emulsified into an outer aqueous poly(vinyl alcohol) (PVA) solution. When using the classical “beaker method”, the emulsion droplets were created in the beaker. The size distribution of the generated droplets is generally rather broad, because the fluid velocities and shear forces vary within the beaker. In contrast, when using the “microfluidics-assisted technique”, the emulsion droplets were created under conditions which were better controlled (e.g., in terms of fluid velocities and shear forces). Consequently, the size distribution of the generated droplets was much narrower. Upon contact with water, parts of the dichloromethane are extracted into the aqueous phase. Dichloromethane reaching the air-liquid interface evaporates. Further dichloromethane follows and its concentration in the organic phase decreases. At a certain time point, the amount of organic solvent is no more sufficient to dissolve the PLGA, and the polymer starts to precipitate, encapsulating the drug. Importantly, not only the conditions for the generation of the dichloromethane droplets are different in the investigated set-ups (“beaker vs. microfluidics”), but also the conditions for dichloromethane extraction into the outer aqueous phase and subsequent evaporation. Consequently, not only the size, but also the morphology of the microparticles might substantially differ. To provide similar microparticle sizes for the two preparation techniques, the PLGA concentration in the organic phase had to be adjusted: It was roughly 10-times higher when using the “beaker method”. This resulted in smaller drug losses during production (microparticle hardening being faster) at lower drug loadings. At higher drug loadings, drug loss was similar, potentially due to saturation effects of ibuprofen in the outer phase. To always provide similar practical drug loadings, the theoretical drug content was adjusted accordingly, if needed (please see section 2.2).

Fig. 1.

Schematic presentations of the investigated microparticle preparation techniques, using a: (A) classical beaker, or (B) microfluidics-assisted device.

3.1. Key properties of the microparticles

Fig. 2 shows optical microscopy pictures of ensembles of ibuprofen-loaded PLGA microparticles prepared using the: (A) “classical beaker method” or (B) “microfluidics-assisted technique”. The drug loading was 15 % in these cases. As expected, the size distribution of the microparticles prepared with the “beaker method” was much broader compared to those produced with the “microfluidics technique”. However, in this study, potential size effects were intentionally avoided by manual selection of similarly sized microparticles for both types of preparation techniques. Furthermore, please note that some of the microparticles were transparent, while others were opaque (the investigation of this aspect was beyond the scope of this study).

Fig. 2.

Optical microscopy pictures of ibuprofen-loaded microparticles prepared with the “Beaker method” or “Microfluidics-assisted technique” (before exposure to release medium, 15 % drug loading).

The inner and outer structure of microparticles loaded with 15 % ibuprofen (before exposure to the release medium) is illustrated in Fig. 3: SEM pictures of surfaces and cross-sections are shown. As can be seen, the surfaces were smooth and non-porous in all cases. In contrast, numerous tiny pores were visible within the systems, irrespective of the type of preparation technique. The pores were uniformly distributed throughout the microspheres, forming a continuous network. X-ray μCT confirmed this inner microparticle structure: The top row Fig. 4 shows images of ibuprofen-loaded microparticles, which were prepared with the “beaker method” or “microfluidics-assisted method” (before exposure to release medium). X-ray μCT offers the advantage of virtual cross-sectioning, thereby avoiding potential artefacts that can arise from physical cutting during sample preparation for SEM analysis. Video 1 shows a series of virtual cross-sections through this type of drug-loaded PLGA microparticles (before exposure to the release medium). Clearly, tiny pores are homogeneously distributed throughout the spheres. Video 2 shows a series of virtual cross-sections through ibuprofen-loaded microparticles prepared with the “microfluidics-assisted technique” (before exposure to the release medium). Also in this case, tiny pores were visible throughout the spheres. For reasons of comparison, also drug-free microparticles were characterized by X-ray μCT: The bottom row in Fig. 4 shows pictures of particles prepared with the “classical beaker method” and “microfluidics-assisted technique”. Interestingly, drug-free microparticles were non-porous, irrespective of the type of preparation technique. This might at least in part be explained by differences in the solubility of the PLGA in “pure dichloromethane” versus “dichloromethane containing a high concentration of dissolved ibuprofen” (Lefol et al., 2023). The kinetics of polymer precipitation during microparticle formation can be expected to strongly depend on the solubility of PLGA in the organic phase.

Fig. 3.

SEM pictures of surfaces and cross-sections of ibuprofen-loaded microparticles prepared with the “Beaker method” or “Microfluidics-assisted method” (before exposure to release medium, 15 % drug loading). The degree of magnification increases from the top to the bottom.

Fig. 4.

X-μCT images (virtual cross-sections) of microparticles loaded with 15 % ibuprofen prepared with the “Beaker method” or “Microfluidics-assisted method” (before exposure to release medium) (top row). For reasons of comparison, also drug-free microparticles are shown (bottom row).

Fig. 5 shows the X-ray diffraction patterns of PLGA microparticles loaded with 15 % ibuprofen, prepared with the “classical beaker method” and “microfluidics-assisted technique”. For reasons of comparison, also the X-ray diffraction patterns of the raw materials (ibuprofen and PLGA) are illustrated. Clearly, the drug was initially crystalline and the polymer amorphous. In none of the PLGA microparticles diffraction peaks were visible, suggesting that the drug is either completely dissolved in the polymer, or partly dissolved & partly dispersed in the form of amorphous particles. At high drug loadings (e.g., 48 %), diffraction peaks from crystalline ibuprofen were visible (Lefol et al., 2023).

Fig. 5.

X-ray diffraction patterns of PLGA microparticles loaded with 15 % ibuprofen, prepared with the “Beaker method” or “Microfluidics-assisted method” (before exposure to the release medium). For reasons of comparison, also the X-ray diffraction patterns of the raw materials are shown (PLGA, ibuprofen).

Thus, the inner and outer structure of the microparticles were similar, irrespective of the type of preparation method: using a microfluidics device or a “classical” beaker.

3.2. Drug release kinetics

Fig. 6 shows the observed ibuprofen release kinetics from PLGA microparticles loaded with 15 % ibuprofen, prepared with the “classical beaker method” or “microfluidics-assisted technique” upon: (A) exposure to well agitated phosphate buffer pH 7.4, or (B) inclusion into an agarose gel. Drug release is shown from single microparticles, their initial diameter is indicated on the right-hand sides of the diagrams. In the bottom row of Fig. 6 the swelling kinetics of the same microparticles are illustrated. The same symbols in the top and bottom rows describe the same microparticles. Please note that the swelling kinetics were estimated based on the “apparent” average diameter of the microparticles observed by optical microscopy. In the case of the “agitated bulk fluid set-up”, microparticle swelling was not hindered. However, in the “agarose set-up”, microparticle swelling was intentionally hindered by the surrounding hydrogel to better mimic the conditions upon administration to a patient. To estimate the microparticles' diameter, optical microscopy pictures were taken from the top at pre-determined time points after inclusion into the agarose gel. It has to be pointed out that a preferential expansion in the plane parallel to the gels' surfaces was observed. This is due to the way the microparticles were placed into the agarose gels: First, only half of the agarose gel was cast into a well. Then, a microparticle was placed in its center, followed by casting the second half of the agarose gel on top. The cohesion between the 2 agarose layers was not as high as the cohesion within one of these layers. Consequently, the microparticles encountered less mechanical resistance when expanding “in between” the two gel layers. Since the pictures were taken from the top, pronounced swelling tends to be overestimated, and values >150 % diameter increase should be viewed with great caution in the agarose set-up.

Fig. 6.

Drug release from and swelling of ibuprofen-loaded PLGA microparticles upon exposure to: (A) well agitated phosphate buffer pH 7.4, or (B) agarose gel. The microparticles were prepared with the “Beaker method” (marked in blue) or “Microfluidics-assisted method” (marked in green). Single microparticles were studied, their initial diameters are indicated on the right-hand side of the diagrams. The same symbols are used to illustrate drug release and swelling of the same microparticle. The drug loading was 15 %. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Interestingly, ibuprofen release in well agitated bulk fluid from microparticles prepared with the “microfluidics technique” was more uniform and faster than from microparticles prepared with the “classical beaker method” at this drug loading (diagram at the top in Fig. 6A). In contrast, the swelling kinetics of the microparticles prepared with the “microfluidics technique” showed a much higher variability compared to the “beaker method” (diagram at the bottom of Fig. 6A).

On the contrary, upon inclusion into agarose gels, drug release from both types of microparticles exhibited substantial variability (Fig. 6B). Overall, the release rate from microparticles prepared with the “microfluidics technique” was again higher than from microparticles prepared with the “beaker method”. As expected, microparticle swelling was slower and less pronounced in agarose gels compared to well agitated bulk fluid.

Fig. 7 shows the observed ibuprofen release kinetics from and swelling kinetics of PLGA microparticles with a higher initial drug loading: 22 %. Again, drug release was monitored upon exposure to: (A) well agitated phosphate buffer pH 7.4, or (B) inclusion into an agarose gel. In contrast to the lower drug loading, ibuprofen release was overall faster from microparticles prepared with the “beaker method” compared to the “microfluidics-assisted technique”. Interestingly, two types of release behaviors could be distinguished, irrespective of the microparticle preparation technique and release set-up: (i) microparticles, which released the drug right from the beginning at a high rate, and (ii) microparticles, showing only limited ibuprofen release during the first couple of days, followed by a steep increase in the release rate (leading to complete drug exhaust). This onset of drug release coincided well with the beginning of significant microparticle swelling. Fig. 8 shows ibuprofen release from and swelling of PLGA microparticles loaded with even higher drug amounts: (A) 36 % and (B) 48 %, respectively. In these cases, only drug release upon exposure to well agitated bulk fluid was measured. As can be seen, the variability in drug release and swelling is considerable, irrespective of the type of preparation method and initial ibuprofen loading.

Fig. 7.

Drug release from and swelling of ibuprofen-loaded PLGA microparticles upon exposure to: (A) well agitated phosphate buffer pH 7.4, or (B) agarose gel. The microparticles were prepared with the “Beaker method” (marked in blue) or “Microfluidics-assisted method” (marked in green). Single microparticles were studied, their initial diameters are indicated on the right-hand side of the diagrams. The same symbols are used to illustrate drug release and swelling of the same microparticle. The drug loading was 22 %. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 8.

Drug release from and swelling of single PLGA microparticles loaded with: (A) 36 %, or (B) 48 % ibuprofen upon exposure to well agitated phosphate buffer pH 7.4. The microparticles were prepared with the “Beaker method” (marked in blue) or “Microfluidics-assisted method” (marked in green). The results obtained with microparticles containing 48 % drug prepared by the “beaker method” are reproduced from (Lefol et al., 2023), with permission. The initial microparticle diameters are indicated on the right-hand side of the diagrams. The same symbols are used to illustrate drug release and swelling of the same microparticle. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.3. Drug release mechanisms

In order to better understand the observed drug release kinetics, the behavior of each single microparticle was analyzed. The diagrams in Fig. 9 show the drug release and swelling kinetics of specific microparticles upon exposure to well agitated phosphate buffer pH 7.4 (A and B), or inclusion in agarose gel (C and D). Microparticles prepared with the “beaker method” are shown in A and C, microparticles prepared with the “microfluidics technique” in B and D. The blue and green curves show drug release (in “bulk fluid” or “agarose gel”) and refer to the left y-axes, whereas the orange curves illustrate microparticle swelling and refer to the right y-axes. Asterisks indicate microparticle fragmentation. The drug loading was 15 % in all cases. Optical microscopy pictures of the microparticles after different exposure times to well agitated phosphate buffer pH 7.4 or inclusion into agarose gels of microparticles prepared with the “beaker method” or “microfluidics-assisted technique” are shown in Fig. 10.

Fig. 9.

Ibuprofen release from and swelling kinetics of single microparticles: (A) prepared with the “Beaker method” upon exposure to well agitated bulk fluid, (B) prepared with the “Microfluidics method” upon exposure to well agitated bulk fluid, (C) prepared with the “Beaker method” upon exposure to agarose gel, and (D) prepared with the “Microfluidics method” upon exposure to agarose gel. The drug loading was 15 % in all cases. Each diagram shows the behavior of a specific single microparticle. The initial microparticle diameter is indicated in each diagram. The blue curves illustrate drug release in well agitated bulk fluid, the green curves in agarose gel. The orange curves show the swelling kinetics of the microparticles. Red flashes indicate time points at which the inner pore network gets direct surface access. Black ovals are used to mark the onset of significant system swelling, coinciding with an acceleration of drug release. An asterisk indicates microparticle fragmentation. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 10.

Optical microscopy pictures of ibuprofen-loaded microparticles: (A) prepared with the “Beaker method” upon exposure to well agitated bulk fluid, (B) prepared with the “Microfluidics method” upon exposure to well agitated bulk fluid, (C) prepared with the “Beaker method” upon exposure to agarose gel, and (D) prepared with the “Microfluidics method” upon exposure to agarose gel. The exposure times are indicated at the top, the initial microparticle diameters on the left-hand side. The drug loading was 15 %.

Importantly, every microparticle exhibited “its own”, specific drug release and swelling profile, irrespective of the type of preparation method and experimental conditions. Interestingly, in many cases ibuprofen release started before substantial microparticle swelling set on. This might be explained as follows: Initially, the microparticles had a non-porous, smooth surface, while continuous pore networks dominated their inner structure (e.g., Fig. 3, Fig. 4, Video 1 &2). Once these pore networks get into direct contact with the release medium, water can easily penetrate into the microparticles and ibuprofen is rapidly released through water-filled channels. The time point of the first contact “release medium - pore network” depends on the specific microparticle structure and experimental conditions: Each microparticle has “its own” inner structure. Some microparticles can be expected to have “weak points”, at which only very thin PLGA layers separate the pore network from the release medium. These microparticles will likely start releasing the drug early. The red flashes in Fig. 9 indicate such events. Importantly, not only the microparticles' specific inner structure, but also the mechanical stress experienced during the release measurement likely plays an important role (Lefol et al., 2023): The convective flow in well agitated bulk fluid can probably more easily cause damage of a microparticle's surface than the “more protecting” conditions in an agarose gel (in the absence of convective liquid flow). For example, comparing the diagrams on the left vs. right hand side at the top of Fig. 6, it can be seen that drug release generally starts earlier upon exposure to agitated bulk fluid than upon inclusion into agarose gel, irrespective of the type of microparticle preparation technique.

In contrast, other microparticles exhibited a lag-time of several days with only limited drug release. These cases are highlighted by black ovals in Fig. 9. Interestingly, the onsets of drug release after the lag-time generally coincide with the onsets of significant microparticle swelling. It has previously been suggested that important PLGA swelling sets on as soon as a critical conditions are reached (Tamani et al., 2020; Tamani et al., 2021; Bode et al., 2018; Bode et al., 2019a; Bode et al., 2019b). At that stage: (i) The system is sufficiently hydrophilic to host large amounts of water. (ii) The polymeric network is sufficiently labile (due to reduced polymer chain entanglement) and cannot offer the required mechanical resistance to prevent the penetration of large amounts of water. (iii) The concentration of water-soluble degradation products generates sufficient osmotic pressure to drive large amounts of water into the microparticles. Once this significant system swelling sets on, the mobility of the drug substantially increases and drug release is accelerated. This type of events is highlighted by black ovals in Fig. 9. The optical microscopy pictures in Fig. 10 clearly illustrate the drastic changes for ibuprofen transport within the microparticles over time. At late time points, the particles have been transformed into highly swollen polymer gels. If drug is still present in the microparticles when significant system swelling sets on, its release is accelerated.

The consequences of the difference in mechanical stress the microparticles are exposed to during the experiments, can also be seen in Fig. 10: Microparticle fragmentation was only observed in the well agitated bulk set-up, but not in the agarose set-up. Fragmentation leads to shorter diffusion pathways and, thus, accelerated release of the drug, which is still inside the system.

Importantly, these hypothesized drug release mechanisms seem to be also valid for the higher drug loadings (22, 36, and 47 %), irrespective of the type of microparticle preparation technique and experimental set-up. Looking at Fig. 7A, it can be seen that certain microparticles release right from the beginning upon exposure to well agitated phosphate buffer pH 7.4, because their pore networks likely get direct surface access at early time points. In contrast, other microparticles show a clear lag-time with only limited drug release, followed by a steep increase in the release rate, which coincides with the beginning of significant microparticle swelling. This is also true for microparticles exposed to agarose gel (Fig. 7B), irrespective of the type of microparticle preparation technique. At 36 and 48 % ibuprofen loading (Fig. 8), the same drug release mechanisms seem to dominate: Certain microparticles release their drug load “early”, because their inner pore network gets into direct contact with the release medium. The other particles “have to wait” until significant system swelling sets on.

Overall, the underlying drug release mechanisms seem to be similar in the investigated microparticles, irrespective of the type of preparation technique (using a microfluidics device or a beaker).

4. Conclusions

A “microfluids-assisted technique” can facilitate the preparation of controlled release PLGA microparticles with a less broad particle size distribution. However, the resulting drug release kinetics are not necessarily predominantly determined by the systems' size. In this study, also the outer & inner structure of the microparticles played a key role. Highly variable drug release kinetics from single microparticles were observed, irrespective of the type of preparation method (using a microfluidics device or a “classical” beaker), and the underlying drug release mechanisms seem to be the same. Furthermore, the experimental conditions (e.g., well agitated bulk fluid vs. agarose gel) can significantly affect drug release. Thus, great care should be taken when developing PLGA microparticles for controlled drug delivery: Multiple factors, including the microparticles' size, outer & inner morphology as well as the experimental conditions used for the in vitro release studies can strongly affect the observed drug release kinetics. Importantly, even in the case of similar microparticle compositions, sizes and morphologies, highly variable release kinetics might be observed from individual microparticles: Each microparticle is unique and releases the drug in “its own” way.

The following are the supplementary data related to this article.

Supplementary Video 1

Inner structure of ibuprofen-loaded PLGA-based microparticles prepared with the “classical beaker method” (before exposure to the release medium): Series of μCT images showing virtual cross-sections. The drug-loading was 15 %.

Supplementary Video 2

Inner structure of ibuprofen-loaded PLGA-based microparticles prepared with the “microfluidics-assisted technique” (before exposure to the release medium): Series of μCT images showing virtual cross-sections. The drug-loading was 15 %.

CRediT authorship contribution statement

L.A. Lefol: Writing – original draft, Visualization, Validation, Methodology, Investigation. A. Sodano: Investigation. P. Bawuah: Visualization, Validation, Methodology, Investigation, Formal analysis. J.A. Zeitler: Writing – review & editing, Visualization, Validation, Supervision, Resources, Project administration, Methodology, Funding acquisition, Conceptualization. J. Verin: Validation, Methodology, Investigation. F. Danede: Visualization, Validation, Methodology, Investigation. J.F. Willart: Writing – review & editing, visualization, Validation, Resources, Methodology, Investigation, Conceptualization. J. Siepmann: Writing – review & editing, Visualization, Supervision, Resources, Project administration, Methodology, Funding acquisition, Conceptualization. F. Siepmann: Writing – review & editing, Visualization, Supervision, Resources, Project administration, Methodology, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

L.A. Lefol, P. Bawuah, J.A. Zeitler, J. Verin, J. Siepmann, F. Siepmann report financial support was provided by Interreg 2 Seas. L.A. Lefol, P. Bawuah, J.A. Zeitler, J. Verin, J. Siepmann, F. Siepmann report financial support was provided by European Regional Development Fund. J. Siepmann, F. Siepmann report financial support was provided by Interreg FWVL. Two co-authors of this article (J Siepmann and F Siepmann) are the guest editors of the special issue this article is part of. Furthermore, the Editor-in-Chief of the journal is one of the co-authors of this article (J Siepmann). The manuscript has been subject to all of the journal's usual procedures, including peer review, which has been handled independently of J Siepmann and F Siepmann. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.