Keywords bead formation - electrospun membranes - membrane characterization - physicochemical
properties - PLGA scaffolds.
Introduction
Poly(lactic-co-glycolic acid) (PLGA) is a widely used biodegradable polymer in biomedical
applications due to its excellent biocompatibility, tunable degradation rates, and
ability to support cell adhesion and proliferation.[1 ]
[2 ] The electrospinning process is a popular technique for fabricating PLGA membranes,
allowing for the creation of nanofibrous structures with high surface area-to-volume
ratios and interconnected porous networks. These properties are ideal for applications
in tissue engineering, drug delivery, and wound healing, where controlled release,
cell adhesion, and scaffold degradation are critical.[3 ]
[4 ] PLGA scaffolds offer significant advantages over traditional systems, including
biphasic calcium phosphate and bioceramic scaffolds, which have shown potential in
enhancing osteogenesis and angiogenesis but may lack the same flexibility in tunable
degradation and surface modification.[5 ]
[6 ]
The morphology of electrospun fibers, including the presence or absence of beads,
can significantly influence the mechanical, physicochemical, and biological properties
of the resulting membranes.[7 ]
[8 ] Bead formation during electrospinning is often considered undesirable. Beads can
create defects within the fibrous network, potentially reducing mechanical strength,
altering pore size distribution, and impacting cellular interactions and material
properties.[9 ]
[10 ]
[11 ] Consequently, extensive research has focused on optimizing electrospinning parameters
to minimize bead formation and produce uniform, bead-free fibers for enhanced membrane
performance.[12 ]
[13 ]
However, the actual impact of beads on the properties of electrospun PLGA membranes
remains under debate. While some studies suggest that bead-free fibers offer superior
mechanical and biological performance,[7 ]
[11 ]
[14 ] others have found minimal differences in the performance of beaded and bead-free
membranes under certain conditions, such as particular mechanical loads, low-stress
environments, or nonload-bearing biological applications (e.g., tissue scaffolding
or filtration systems).[15 ]
[16 ] Additionally, recent studies have demonstrated that beads can play a functional
role in enhancing membrane performance for specific applications, such as water purification
through membrane distillation.[17 ]
[18 ] These discrepancies highlight the need for a deeper understanding of how beads influence
the overall properties of electrospun membranes, particularly in complex biological
environments.
This study aims to systematically investigate the effect of bead morphology on the
physicochemical and biological properties of electrospun PLGA membranes. It evaluates
the impact of bead formation on hydrophilicity, molecular structure, and surface characteristics,
and compares the ability of beaded, fibrous, and mixed membranes to support cell adhesion,
proliferation, and viability using human dental pulp stem cells (DPSCs). By challenging
the assumption that bead-free fibers are inherently superior, this work highlights
the functional benefits of bead morphology and provides new insights into the optimization
of electrospun PLGA membranes for biomedical applications, emphasizing their potential
for tailored scaffold design.
Materials and Methods
Preparation of Synthetic PLGA membranes
To prepare the PLGA membranes, 1 g of PLGA polymer was dissolved in 10 mL of solvent.
Two types of solvent systems were used: pure dimethylformamide (DMF) and a mixture
of DMF and chloroform in varying ratios (25:75, 50:50, 75:25, DMF). The polymer solutions
were stirred continuously overnight to ensure complete dissolution.
Following dissolution, the solutions were electrospun into membranes using an electrospinning
machine (Simatic HMI, Siemens, Germany). The process was optimized by adjusting several
parameters to achieve the desired membrane morphology. The parameters included the
applied voltage (16, 18, 20, 22, and 24 kV), needle gauge (16, 18, 20, and 22), the
distance between the nozzle and the collecting plate (10, 12, 14, 16, 18, and 20 cm),
and the solution flow rate (0.3, 0.5, 0.8, and 1 mL/h). The drum rotation speed was
maintained at a constant 250 rotations per minute throughout the process. [Table 1 ] summarizes the solvent compositions and machine settings used to fabricate three
types of PLGA membranes: beaded (B), fibrous (F), and mixed (M).
Table 1
Composition and machine settings for synthetic PLGA membranes
PLGA membrane
Solvent
10 mL
Temperature
°C
Needle size
(gauge)
Humidity
Distance
(cm)
Voltage
KV
Flow rate
mL/h
Drum (RPM)
Time (h/cycle)
B
DMF
35
20
60–70%
18
18
0.3
250
8/2
F
DMF + chloroform (5 mL + 5 mL)
35
18
60–70%
16
22
0.5
250
8/2
M
DMF
45
18
60–70%
18
16
0.5
250
8/2
Abbreviations: DMF, dimethylformamide; PLGA, poly(lactic-co-glycolic acid); RPM, rotations
per minute.
After electrospinning, the membranes were allowed to air dry overnight to ensure the
complete evaporation of any residual solvents. Subsequently, the membranes were collected
and stored at room temperature in a well-sealed box. Handling membrane B, the beaded
membrane, posed challenges due to its fragile structure, which made it prone to tearing.
To address this, we kept membrane B on the aluminum foil throughout the experiments
to maintain its integrity.
Membrane Morphology and Physicochemical Characterization
Scanning Electron Microscopy
Scanning electron microscopy (SEM) was utilized to analyze the surface topography
of the three different PLGA membranes. Samples from each membrane type were prepared
by cutting them into square pieces, each measuring 0.5 × 0.5 cm. These samples were
then coated with a thin layer of gold using a Quorum Technologies SC7620 sputter coater
to enhance conductivity. The coated samples were subsequently examined using a SEM
(Tescan VEGA 3 XMU, equipped with Oxford Instruments X-max).
Fourier-Transform Infrared Spectroscopy
The Fourier-transform infrared spectroscopy (FTIR) analysis was conducted using a
Jasco FTIR-6300 (Tokyo, Japan). Samples measuring 1 × 1 cm2 were prepared from the three membranes. To enhance the signal-to-noise ratio, the
spectra were collected with a resolution of 2 cm−1 , and 16 scans per sample were obtained in the range of 4000 to 400 cm−1 at room condition at a constant temperature of 25°C, which corresponds to the CH2
group asymmetric and symmetric axial deformations (ν(C-H)). The spectra were collected
from three replicates per membrane and data analysis was performed using Origin Pro
8.5 software. The attenuated total reflection sampling mode was employed for spectra
analysis, and the obtained spectra were subjected to baseline correction and normalization.
Contact Angle Measurement
The water contact angle measurement was conducted to evaluate the wettability of the
membranes. Samples measuring 1 × 1 cm2 were prepared from the three membranes and placed on a flat stage. Five microliter
droplets of distilled water were introduced onto the surface of the membranes and
the sessile drop technique was performed using Rame-Hart automated goniometer (model
290-U1). The mean contact angle value was determined by taking the average of three
separate measurements at intervals of 5, 10, 15, 30, and 60 minutes. The experiment
was conducted in triplicates for a total number of three samples from each membrane.
In
Vitro Biological Characterization
Cell Culture and Seeding
Human DPSCs (Cat no: 300702) obtained from CLS, Germany, were used in this study.
The cells were cultured in Dulbecco's modified Eagle medium/nutrient mixture F12 supplemented
with 10% fetal bovine serum and 1% penicillin-streptomycin. The cells were maintained
at a temperature of 37°C in a humid atmosphere of 95% O2 and 5% CO2 . The cells on reaching confluence were subcultured with 1× trypsin-ethylenediaminetetraacetic
acid solution (59417 C; Sigma; United States), and the medium was replenished every
2 days.
Before the cell seeding, the three study membranes were prepared in size of 1 × 1 cm2 , sterilized with ultraviolet for 30 minutes, and placed at the bottom of 12-well
plates, maintained in position with the help of sterile stainless steel rings. The
membranes were incubated in 500 µL of complete culture media for 2 hours in a CO2 incubator before cell seeding. The media was then aspirated out and cells were seeded
onto the membranes at a density of 1 × 105 cells per membrane in a maximum volume of 50 µL and further incubated for 2 hours.
An extra 1 mL of the complete culture media was then added and maintained in the incubator
for downstream experiments.
Cell Attachment to Membranes by SEM
For SEM studies, DPSCs were seeded onto 1 × 1 cm2 size B, F, and M membranes at a density of 5 × 104 cells/membrane. The seeded membranes were incubated at 37°C, with 5% CO2 and 50% humidity for 3 days. Following incubation, the membranes were gently rinsed
with phosphate buffer saline (PBS) and fixed in 2.5% glutaraldehyde (Sigma) for 1 hour
at room temperature. After fixation, the membranes were washed with PBS three times
and dehydrated through a graded alcohol series. The cell-seeded membranes were coated
with gold using Quorum Technologies SC7620 to prepare for SEM imaging as mentioned
before.
Immunofluorescence Staining Study
For confocal microscopy, cell-seeded membranes for 48 hours were carefully washed
three times with PBS and fixed using 4% paraformaldehyde for 20 minutes at room temperature.
The membranes were further washed three times in PBS, incubated with FITC-phalloidin
(Abcam, United States) for detecting F-actin cytoskeleton for 30 minutes at 4°C. The
membranes were then counterstained using 4′,6-diamidino-2-phenylindole (DAPI) for
the nucleus (using a mounting medium containing DAPI; Abcam) for 5 minutes. The membranes
were then examined under a confocal microscope (Nikon Eclipse Ti-S, Nikon Instruments
Inc., United States).
Viability Assay of DPSCs on Membranes
For cell viability experiments, DPSCs were cultured at a concentration of 5 × 104 cells/membrane on the three different membranes and incubated for 1, 3, and 7 days.
Cell viability was assessed using the XTT assay (Roche Diagnostics, Mannheim, Germany).
According to the manufacturer's protocol, XTT reagent was added to the cells and incubated
for 4 hours. Subsequently, absorbance was measured at 450 nm using a microplate reader
(BioTek 800 TS). The mean absorbance values were used to calculate the percentage
of cell viability. The experiment was repeated three times.
Statistical Analysis
The statistical analysis software used in the analysis was SPSS version 24.0 and GraphPad
Prism version 8.0. The data collected from the study were expressed as mean ± standard
deviation and analyzed using one-way analysis of variance (ANOVA) for intergroup comparison
and post hoc Bonferroni test for pairwise comparison. A p -value of < 0.05 was considered statistically significant.
Result
Preparation of Synthetic PLGA Membranes
Only membranes that were produced through a successful electrospinning process were
selected for further analysis using SEM. Successful trials were identified based on
several key criteria: a smooth and stable electrospinning process, the consistent
formation of a Taylor cone at the needle tip ([Fig. 1 ]), and the absence of any polymer buildup on the needle or collector. These conditions
led to the production of homogeneous membranes, which were free of defects and had
uniform morphology suitable for subsequent SEM examination.
Fig. 1 (A ) Example of a successful electrospinning process of the synthetic polymer solution
with the formation of the Taylor cone (white arrow) and the plume (yellow arrow).
(B –D ) Examples of unsuccessful processes, identified by the absence of a Taylor cone and/or
the accumulation of material buildup at the needle tip.
Membrane Morphology and Physicochemical Characterization
Scanning Electron Microscope
SEM analysis revealed diverse morphologies among the membranes produced via electrospinning.
Three different membrane types were identified based on their structural characteristics:
one with prominent beaded structures (labeled B), another with predominantly fibrous
textures (labeled F), and a third with a combination of beads and fibers, resulting
in a mixed morphology (labeled M) ([Fig. 2 ]). These variations in morphology underscore the significant impact of the electrospinning
parameters on the structural characteristics of the membranes.
Fig. 2 Morphologies of electrospun synthetic poly(lactic-co-glycolic acid) (PLGA) membranes
by scanning electron microscopy (SEM). (A ) Beads membrane exhibiting beads. (B ) Fiber membrane. (C ) Mixed (beads and fibers) membrane.
Fourier-Transform Infrared Spectroscopy
The FTIR spectra of the three membranes ([Fig. 3 ]) were analyzed to identify their chemical compositions. All three membranes exhibited
characteristic absorption bands, indicative of PLGA's molecular structure. For membrane
B ([Fig. 3A ]), the spectrum displayed key absorption bands at approximately 1,750 cm−1 , corresponding to C = O stretching (carbonyl groups), 1,180 cm−1 for C-O stretching, and 1,090 cm−1 for C-O-C stretching. Additionally, peaks were observed around 2,945 and 1,380 cm−1 , representing CH2 stretching and CH3 bending, respectively. Membrane F ([Fig. 3B ]) exhibited similar absorption bands to membrane B. The major peaks were observed
at approximately 1,750 cm−1 (C = O stretching), 1,180 cm−1 (C-O stretching), and 1,090 cm−1 (C-O-C stretching). The peaks corresponding to the CH2 groups at 2,945 cm−1 were also present. For membrane M ([Fig. 3C ]), the FTIR spectrum followed a similar pattern to both membrane A and membrane B,
displaying key peaks at 1,750 cm−1 (C = O stretching), 1,180 cm−1 (C-O stretching), 1,090 cm−1 (C-O-C stretching), and 2,945 cm−1 (CH2 stretching). No significant changes were observed in the chemical composition across
the three membranes, suggesting that all three share similar structural characteristics.
Fig. 3 Fourier-transform infrared spectroscopy (FTIR) analysis for the three poly(lactic-co-glycolic
acid) (PLGA) membranes. (A ) The bead membrane (membrane B). (B ) The fiber membrane (membrane F). (C ) The mixed membrane (membrane M).
Contact Angle Measurement
The average contact angles of the three membranes are summarized in [Table 2 ]. A significant reduction in contact angle was observed for all membranes after 10 minutes
of incubation (p < 0.05), indicating an initial increase in surface hydrophilicity. After 60 minutes
of incubation ([Fig. 4 ]), the contact angles of membranes B and M showed a further significant decrease
compared to membrane F (p < 0.05). These findings suggest that membranes B and M became more hydrophilic over
time, while membrane F maintained a relatively higher contact angle, indicating it
remained more hydrophobic.
Fig. 4 Contact angle measurements for the three poly(lactic-co-glycolic acid) (PLGA) synthetic
membranes were taken at 30 minutes (left) and 60 minutes (right). The figure shows
the wettability and surface properties of the membranes over time, indicating changes
in hydrophobicity or hydrophilicity. Membranes B and M (A and C ) exhibited higher hydrophobicity compared to membrane F (B ) at 60 minutes.
Table 2
The average values of water contact angle test and its spread from 0 to 60 minutes
Water contact angle values (m ± SD)
Time (min)
Membrane
B
F
M
0
133.5° ± 0.8°
135.7° ± 2.3°
138.3° ± 0.1°
5
133.7° ± 1.6°
132.3° ± 2.4°
135.3° ± 0.4°
10
129.3° ± 0.5°
129.1° ± 2.0°
132.5° ± 0.4°
15
126.4° ± 0.5°
125.0° ± 1.7°
132.5° ± 1.2°
30
114.7° ± 0.6°
109.1° ± 1.6°
116.8° ± 0.1°
60
35.8° ± 0.8°
105.3° ± 4.6°
36.6° ± 1.1°
Abbreviation: SD, standard deviation.
Biological Characterization by In Vitro Experiments
Cell Attachment to Membranes by SEM
When DPSCs were seeded onto the membranes, the cells were able to successfully attach
and exhibited a flattened morphology, with minimal peripheral blebs and ruffles ([Fig. 5A–C ]). Interestingly, SEM images of membrane M revealed that DPSCs not only adhered to
the surface but also tended to penetrate the depths of the membrane structure ([Fig. 5C ]).
Fig. 5 Scanning electron microscopy (SEM) images of dental pulp stem cells (DPSCs) seeded
on poly(lactic-co-glycolic acid) (PLGA) membranes after 3 days. (A ) Bead membrane (membrane B), (B ) fiber membrane (membrane F), and (C ) mixed membrane (membrane M).
Immunofluorescence Staining Study
[Fig. 6 ] shows the immunofluorescence staining of DPSCs seeded on the three synthetic polymer
membranes after 48 hours of incubation. The cells were stained with DAPI (blue) to
visualize the nuclei and with FITC-phalloidin (green) to highlight the F-actin cytoskeleton.
The confocal microscope images indicate successful cell adhesion and spreading across
the surfaces of all three scaffolds. Notably, the actin cytoskeletons in cells grown
on membrane M appeared more developed compared to those on the other membranes, as
demonstrated by a densely populated layer of live cells ([Fig. 6C ]).
Fig. 6 Immunofluorescence staining of dental pulp stem cells (DPSCs) cultured on the synthetic
polymer membranes (A–C ) for 48 hours. Cells' nuclei (blue) and mitochondria (green) are shown and demonstrate
successful cell adhesion and spreading across all scaffold surfaces. Notably, membrane
M (The mixed membrane) exhibited significantly enhanced cell growth and activity,
evidenced by a densely populated layer of live cells. (A ) The bead membrane (membrane B). (B ) The fiber membrane (membrane F). (C ) The mixed membrane (membrane M).
Viability Assay of DPSCs on Membranes
[Fig. 7 ] presents the results of the XTT cell viability assay for DPSCs cultured on synthetic
PLGA membranes for 7 days. A significant increase in cell number was observed over
time from day 1 to day 7 across all groups, including the control. Repeated measures
ANOVA between-group analysis revealed a significant mean cell proliferation difference
between the groups (p < 0.05). However, post hoc multiple comparisons, with a significance level of 0.05
(two-tailed), indicated no significant differences between the membrane groups at
day 7. The control group showed significant differences compared to all membrane groups
at all time intervals (p < 0.05).
Fig. 7 Results of the XTT cell viability assay for dental pulp stem cells (DPSCs) cultured
on the different poly(lactic-co-glycolic acid) (PLGA) membranes over 1, 3, and 7 days.
A significant increase in cell number was observed from day 1 to day 7 across all
groups. The control group showed significant differences compared to all membrane
groups at all time intervals. No significant differences between the membrane groups
at day 7. (A ) The bead membrane (membrane B). (B ) The fiber membrane (membrane F). (C ) The mixed membrane (membrane M).
Discussion
In recent years, biomaterial scaffolds have garnered significant attention in dental
tissue engineering due to their potential to promote tissue regeneration.[19 ]
[20 ] These advancements highlight the importance of ongoing innovation in scaffold design,
particularly in optimizing membrane properties to enhance cellular attachment and
integration with surrounding tissues. This study aimed to investigate the effects
of bead formation on the physicochemical and biological properties of electrospun
PLGA membranes. By fabricating three types of membranes, beaded (B), fibrous (F),
and mixed (M), we were able to directly compare their structural characteristics,
chemical composition, hydrophilicity, and cellular interactions. Based on the literature,
the formation of beads could be related to the instability of the liquid jet and altered
surface tension and hence incomplete evaporation of the solvent solution before it
reached the collector. This led to insufficient polymer entanglement and resulted
in the formation of beads rather than fibers.[21 ] Contrary to the common perception that bead-free fibers are superior in biomedical
applications, our findings suggest that the presence of beads may not significantly
impair the functionality of electrospun PLGA scaffolds.
The initial SEM analysis of the membranes fabricated under different electrospinning
parameters allowed us to select three membranes with distinct morphological characteristics.
SEM was utilized to examine the surface morphology of the membranes and to evaluate
their structural integrity, providing high-resolution images critical for understanding
material properties in tissue engineering applications.[22 ] During our experiments, membrane B (beaded) presented significant challenges in
terms of handling. We were unable to separate membrane B from the aluminum foil used
during electrospinning, as it would tear easily during the removal process. This highlights
a key limitation in its practical usability, as it lacked the structural integrity
required for further manipulation during experimentation. Consequently, membrane B
would not be suitable for applications requiring robust handling or mechanical stability.
In contrast, both membrane F and membrane M maintained their structural integrity
throughout the experiments, with membrane M demonstrating superior handling properties.
This makes membrane M a more viable option for practical applications where durability
is essential.
The FTIR spectra indicated that all three membranes had similar chemical compositions,
displaying characteristic absorption bands associated with PLGA's molecular structure.
The spectra showed key absorption bands around 1,750 cm−1 (C = O stretching), 1,180 cm−1 (C-O stretching), 1,090 cm−1 (C-O-C stretching), 2,945 cm−1 (CH2 stretching), and 1,380 cm−1 (CH3 bending), all of which are typical of PLGA's polymer backbone.[23 ]
[24 ] Contrary to previous studies,[7 ]
[10 ] no significant broad peak around 3,400 cm−1 was observed in any of the membranes, indicating the absence of hydroxyl groups or
residual solvents that could suggest surface modification. This confirms that the
variations in membrane morphology did not affect the chemical composition or introduce
significant surface modifications. Therefore, the differences observed in membrane
performance, including hydrophilicity and cellular interactions, are more likely attributed
to the physical structure (presence of beads vs. continuous fibers) rather than any
substantial changes in chemical composition.
The hydrophilicity of membrane surfaces plays a crucial role in enhancing cell attachment
and viability by reducing nonspecific or hydrophobic protein adsorption.[25 ] In a previous study, the gradual reduction in the PLGA water contact angle over
time was described as contact angle relaxation, a phenomenon indicating changes in
surface energy or reorientation of surface molecules upon exposure to water.[26 ] In our study, the contact angle measurements revealed that all membranes were initially
hydrophobic, with water contact angles above 90 degrees,[27 ] consistent with previous reports on PLGA materials.[28 ]
[29 ] However, membranes B and M, which contained beads, showed a significant decrease
in contact angle over time, becoming more hydrophilic. This finding suggests that
the presence of beads may enhance the exposure of hydrophilic functional groups or
alter the surface roughness, facilitating greater water absorption, as supported by
Li and Xia.[9 ] Similar observations have been made with other applications, where bead-containing
nanofibers, such as those used in membrane distillation, have demonstrated enhanced
surface roughness and water interaction, leading to improved membrane performance.[17 ]
Biological characterization through in vitro experiments demonstrated that all three membranes exhibited good biocompatibility,
as they effectively supported DPSC attachment and proliferation. SEM images, which
are critical for visualizing cell–material interactions,[22 ] confirmed that DPSCs adhered well to all membrane types and exhibited a flattened
morphology with minimal blebbing or ruffling. However, cells on membrane M (mixed
morphology, [Fig. 5C ]) appeared to penetrate deeper into the membrane structure. This finding suggests
that the mixed morphology might offer a more conducive environment for cell infiltration,
potentially enhancing cell–matrix interactions in ways that purely beaded or fibrous
membranes do not, aligning with the findings of Bhattarai et al.[15 ]
The immunofluorescence staining study further revealed that cells on membrane M had
more developed actin cytoskeletons and formed a densely populated layer, indicating
robust cell spreading and interaction with the scaffold. These results suggest that
a mixed bead–fiber morphology could provide a more favorable surface topography for
cell adhesion and cytoskeletal organization, potentially improving the scaffold's
suitability for tissue engineering applications.[30 ]
The XTT assay was employed to assess cell viability and proliferation on the PLGA
membranes. This colorimetric assay is based on the reduction of XTT to a water-soluble
formazan product by metabolically active cells, facilitating the quantification of
viable cells without the need for additional solubilization steps.[31 ] The results showed no significant differences in cell viability between the three
membrane types at day 7, although all membranes supported substantial cell proliferation
over time. Notably, while differences in cell viability were not statistically significant,
membrane M (mixed) demonstrated deeper cell infiltration, as evidenced by SEM. This
suggests that the mixed morphology may offer an improved microenvironment for cell
integration, potentially enhancing scaffold performance in applications requiring
high cell–matrix interactions, particularly in long-term applications.
These findings challenge the assumption that bead-free membranes inherently provide
better cellular environments, highlighting that membranes with different morphologies
can be equally effective under certain conditions. This versatility underscores the
potential of electrospun PLGA membranes in biomedical applications, emphasizing that
the optimal morphology should be tailored to specific application needs rather than
defaulting to bead-free structures.
The results of this study have important implications for the design and fabrication
of electrospun PLGA scaffolds. The absence of significant differences in cell viability
and the varying impacts on cell morphology and hydrophilicity suggest that bead presence
may not negatively impact scaffold performance as much as previously thought. Instead,
bead formation could be strategically used to create scaffolds with tailored properties.
For instance, the enhanced hydrophilicity and potential for cell penetration observed
in beaded and mixed membranes could be beneficial for applications requiring rapid
cell infiltration and integration, such as in wound healing or soft tissue regeneration.[14 ] Future research should explore the long-term effects of bead morphology on scaffold
performance, including degradation behavior, mechanical stability, and in vivo functionality, to better understand the role of beads in electrospun PLGA membranes.
Conclusion
This study demonstrated that bead formation in electrospun PLGA membranes does not
impair their physicochemical or biological properties and may even enhance hydrophilicity
and cell penetration in mixed morphologies. These findings challenge the conventional
preference for bead-free fibers and suggest that bead morphology could be strategically
leveraged to optimize scaffold performance. Future research should focus on in vivo applications and long-term functionality of bead-containing PLGA membranes in biomedical
contexts.