Comparative Analysis of Electron Microscopy Techniques for Hydrogel Microarchitecture Characterization: SEM, Cryo-SEM, ESEM, and TEM

Abstract

Hydrogels have emerged as a versatile class of materials with broad applications in biomedical engineering, drug delivery, and tissue engineering. Understanding their intricate structures and morphologies is crucial for tailoring their properties to meet specific biomedical needs. It has been clearly established that the composition and microarchitecture of the materials play a critical role in essential cellular mechanisms such as mechanosensing, adhesion, and remodeling. This question is essential in tissue engineering, where precisely characterizing the microarchitecture of the materials used to model the cell microenvironment is a critical step to ensure the reproducibility and relevance of reconstructed tissues. In this study, we present a comprehensive comparison of four advanced electron microscopy techniques, namely, scanning electron microscopy, cryo-scanning electron microscopy, environmental scanning electron microscopy, and transmission electron microscopy, to observe the hydrogel microarchitecture, including a comparison of the sample preparation methods for each technique. Each technique’s specific advantages and limitations are discussed in detail, highlighting their unique capabilities in characterizing the hydrogel structures. We illustrate this study with two semisynthetic hydrogels, such as gelatin methacrylate and hyaluronic acid methacrylate. Moreover, we delve into the critical sample preparation steps necessary for each method, emphasizing the need to preserve the hydrogel’s native state while obtaining high-resolution images. This comparative analysis aims to select the most suitable electron microscopy technique for their hydrogel studies, fostering deeper insights into the design and development of advanced biomaterials for tissue engineering applications.

Introduction

In tissue engineering and regenerative medicine, hydrogel materials have emerged as a promising class of materials due to their biocompatibility, tunable mechanical properties, and ability to encapsulate bioactive molecules.¹⁻⁴ These three-dimensional cross-linked polymer networks with high water content offer a versatile platform for building an environment that mimics the extracellular matrix, providing a conducive material for cell adhesion, proliferation, and tissue formation.⁵ Fabricating biologically relevant microenvironments capable of driving cellular organization and fate requires a careful understanding of the internal organization of the hydrogel matrix.⁶⁻⁹ The microarchitecture and porosity of hydrogels play critical roles in cell behavior (including migration, adhesion, proliferation, and mechanosensing). However, observing these properties is challenging because it requires a high-resolution technique (submicrometer scale) with a sample preparation that does not damage the microarchitecture of these highly water-containing materials.

In this context, electron microscopy techniques have proven to be interesting tools, enabling one to visualize the hydrogel microarchitecture.⁷ Current literature in the field of tissue engineering frequently discusses a widely used sample preparation technique for observing the microarchitecture of hydrogels: freeze-drying coupled with scanning electron microscopy (SEM). This method is often presented as the standard for characterizing the pore size distribution in hydrogels. However, this sample preparation technique does not provide an accurate representation of the gel’s structure due to alterations that occur during the freeze-drying process.¹⁰ These alterations include the introduction of artifacts, like water crystal formation, which can distort the true architecture of the hydrogel. Consequently, the final images obtained may not reliably reflect the native state of the hydrogel’s microarchitecture.

The primary aim of this work is to identify the electron microscopy technique that, when combined with the optimal sample preparation method, provides the appropriate resolution and preserves the native structure of the hydrogel to accurately visualize its microarchitecture. For this study, we have chosen to work with two widely used hydrogels of different types: gelatin methacrylate (GelMA), a protein-based hydrogel, and hyaluronic acid methacrylate (HAMA), a polysaccharide-based hydrogel^11,12; four electron microscopy techniques and their associated sample preparation methods were then compared with these two hydrogels for microarchitecture observation: SEM, cryo-scanning electron microscopy (Cryo-SEM), environmental scanning electron microscopy (ESEM), and transmission electron microscopy (TEM).

Using SEM requires prior dehydration of the sample, which can alter the native structure of hydrogels and may affect the accuracy of the microarchitecture observation. ESEM can also be employed and is unique in that it does not require sample preparation, though it may have a lower resolution compared to other techniques. Cryo-SEM utilizes high-pressure cryofixation to vitrify the sample, preserving its structure in a near-native state; however, it is technically demanding and requires specialized equipment. TEM is a really high-resolution technique that necessitates ultrathin sections of samples that have been fixed in epoxy resin. The detailed steps of sample preparation and observation for each technique will be discussed further in the Results section of this article. Understanding these limitations will aid in selecting the most suitable electron microscopy approach for accurately characterizing the hydrogel microarchitecture in various tissue engineering applications. In this paper, we did not include CryoTEM in our comparison, but we acknowledge the potential of CryoTEM in studying the hydrogel network architecture. CryoTEM can provide structural information in a native-like hydrated state without the need for sample dehydration.¹³⁻¹⁵ However, it is important to note that while CryoTEM is a promising tool for observing hydrogels’ structure, it remains a challenging and specialized technique. The preparation of cryolamellae, especially for soft and highly hydrated materials such as hydrogels, can be difficult and is not routinely employed in standard TEM workflows. Additionally, CryoTEM requires specialized equipment and expertise, which limits its widespread use in routine studies.

Materials and Methods

Materials

HAMA, lithium-2,4,6-trimethylbenzoylphosphinate (LAP), 3-methacryloxypropyltrimethoxysilane (MAPTMS), phosphate-buffered saline (PBS) solution gelatin from porcine skin type A (90–110 g bloom), methacrylic anhydride (MAA), sodium carbonate (Na₂CO₃), sodium bicarbonate (NaHCO₃), sodium hydroxide (NaOH), hydrochloric acid (HCl), and deuterium oxide (D₂O) were purchased from Sigma-Aldrich and were used without purification.

GelMA was synthesized following the previously reported procedures with modifications.¹⁶ Briefly, 20 g of gelatin was dissolved in 100 mL of carbonate/bicarbonate buffer solution (0.25 M, pH 9.0) under stirring at 50 °C for 2 h. MAA at 0.1 mL/g was added under vigorous stirring at the rate of 0.1 mL/min and left to react for 3 h at 50 °C. Then, the solution was stopped by adding deionized water (DI) (100 mL). The resulting solution was transferred to a dialysis tube (10 kDa MWCO, Thermo Fischer) and dialyzed at 40 °C for 4 days to remove any impurities. Finally, the solution was frozen in liquid nitrogen and lyophilized for 5 days (LyoQuest −85 Plus, Telstar). The resulting white solid was stored at −20 °C until further use. The degree of substitution (DS) was analyzed by nuclear magnetic resonance (¹H-NMR) spectroscopy (500 MHz, Bruker).

Hydrogel Preparation

GelMA (10% w/v) and HAMA (3% w/v) hydrogels were prepared by dissolving the lyophilized powder in PBS solution under stirring at 37 °C for GelMA and at 4 °C for HAMA for 1 h. Next, the photoinitiator, LAP (0.1 and 1.13% w/w, respectively), was added to the polymer solution under stirring conditions for 1 h. In all final polymer solutions, the concentration of LAP remained constant at 2% w/v. Hydrogel photopolymerization was prepared in a cylinder polypropylene mold (8 mm in diameter and 5 mm in height). First, the mold was filled with 300 μL of hydrogel solution, and the MAPTMS-treated coverslip was placed on top to cover the pool. Following that, the polymer solution was photo-cross-linked by UV irradiation (Form Cure, FH-CU-01) at 405 nm for either 30 s or 2 min. Upon completion of photopolymerization, the hydrogel was carefully removed from the mold and detached from the glass coverslip. Then, hydrogels were submerged in PBS at 4 °C for 24 h to reach equilibrium swelling before the sample preparation for electronic microscopy characterization.

Cell Encapsulation and Culture

Isolation and Amplification of MSCs from the Stromal Vascular Fraction of Adipose Tissue (ASC)

Abdominal dermolipectomies (DLAs) were provided by the plastic surgery department of the University Hospital Center (CHU) in Toulouse, France. The gender, body mass index, and age of the donor were recorded before anonymization. The experimental protocols were approved by the institutional ethics committee of the French Ministry of Research (No.: DC-2015-23-49), and informed consent was obtained from all subjects in accordance with institutional guidelines on the handling and use of human tissues. To obtain the stromal vascular fraction (SVF), the adipose tissue was dissected upon receipt of the DLA in the laboratory, mechanically dissociated, and then digested for 45 min at 37 °C with agitation, using collagenase NB4 at 13.6 U/mL in an α-MEM medium supplemented with ASP (0.1% amphotericin and 1% streptomycin/penicillin). After filtration through a 100 μm filter and centrifugation (500 g, 10 min), the cells were washed in αMEM-ASP and centrifuged again (500 g, 5 min). The cell pellet was resuspended in erythrocyte lysis buffer and incubated for 5 min at room temperature (RT). This preparation was then centrifuged (600 g, 5 min) and resuspended in “Endothelial Growth Medium-2” (EGM2) supplemented with ASP. The final cell solution—corresponding to the SVF—was counted and seeded at 4000 cells/cm² for cell amplification in a two-dimensional (2D) monolayer. The medium was changed every 3 days. The 2D cultures were maintained in EGM2 or α MEM + 2% platelet lysate until 80% confluence. The amplified cells, subsequently referred to as “passage 0 cells (P0),” were then enzymatically dissociated and used for further experiments.

Cellular Suspension Embedding

To prepare the GelMA 12 solution, lyophilized GelMA was dissolved in D-PBS with LAP (0,1% w/w, Sigma, USA). The solution was stored at 4 °C until use. Cellular suspensions were prepared from cells harvested at P0 following the laboratory routine protocol.

For encapsulation of cellular suspensions, the cellular suspension was mixed with prewarmed GelMA/0.1% (w/w) LAP solution (37 °C, 1 h) at a high density (66,000 cells/2 μL of GelMA). After homogenization and vortexing, the mixed preparation was sucked up into a PTFE tube (AWG-22L, OD: 1.01 mm; ID: 0.710 mm, PTFE TUBE SHOP, The Netherlands) connected to a 1 mL luer-lock syringe (ref 309628, BD MEDICAL, USA). The PTFE tube containing the hydrogel with cells was then photo-cross-linked using 405 nm light for 40 s (Form cure, Formlabs, Germany). The encapsulated cellular suspension was extruded from the PTFE tube using a syringe, and the resulting cylindrical structure was cut into smaller cylinders, which were individually transferred into 24-well flat-bottom ULA plates (Corning Incorporated, Life Science, USA). They were maintained in an EGM2 proliferation medium for 7 days, with half of the medium being changed every 2–3 days.

After 7 days, the cylinders were fixed with 4% paraformaldehyde (PFA 4%) at RT and then washed with D-PBS. The samples were stored at 4 °C until sample preparation for TEM observations.

Spheroid Formation

Spheroids were formed from either SVF or P0-SVF cells. To promote cell aggregation, 50,000 cells were seeded in a reduced volume of EGM2 medium (50 μL) in ultralow attachment (ULA) 96-well round-bottom plates (Corning Incorporated Life Sciences, USA) and maintained overnight under stirring (150 rpm). For SVF cells, to further improve cell aggregation, cell seeding was followed by plate centrifugation (600 g for 5 min). The following day, EGM2 (150 μL) was added to each well. Cells were maintained in the proliferation medium until spheroid formation, i.e., 5 days for SVF-spheroids and 1 day for P0-SVF-spheroids.

Individual Spheroid Embedding

Individual spheroid embedding was previously described by Escudero et al.¹⁷ To generate GelMA-embedded spheroids, once formed, spheroids were mixed with prewarmed GelMA/0.1% LAP(w/w) solution (37 °C, 10 min). Spheroids were then individually pipetted in a defined volume of GelMA/0.1% LAP solution (1.5 μL) and dispensed onto an antiadhesive PDMS surface, prepared as described above. GelMA droplets containing one spheroid were then photo-cross-linked via exposure to 405 nm light for 40 s (Form cure, Formlabs, Germany). Embedded spheroids were individually transferred into 24-well flat-bottom ULA plates (Corning Incorporated, Life Sciences, USA) and maintained in an EGM2 proliferation medium for 7 days before differentiation. Half of the medium was changed every 2–3 days.

Adipocyte Cell Differentiation in 3D Cultures

GelMA-embedded spheroid differentiation was initiated after a proliferation phase of 7 days in EGM2 medium. Cells were differentiated for 21 days with appropriate adipogenic cocktails. Half of the medium was changed every 3–4 days. Cells were differentiated using variations of an adipogenic cocktail previously described by Deschasseaux et al.¹⁸ These adipogenic cocktails consist of αMEM-ASP supplemented with fetal bovine serum (2%), insulin (55 μg mL^–1), apotransferin (10 μg mL^–1, Sigma, USA), and bone morphogenetic protein 7 (50 ng mL^–1, MiltenyiBiotec, France) with or without intralipids (0.2% diluted from 20% emulsion, Sigma, USA). When specified, the TGFβ pathway inhibitor SB431542 (MiltenyiBiotec, Germany), also termed SB4, was added to the adipogenic cocktail (5 μm). For treatment with UCP1 inducers, cells were treated 3 days prior to the end of the differentiation process with rosiglitazone (100 nm, Sigma, USA), 3,3″,5-triiodo-L-thyronine (T3, 0.2 nm, Sigma, USA), all-trans retinoic acid protected from light (0.1 μm, Sigma, USA), and 8-(4-chlorophenylthio)-adenosine 3″,5″-cyclic monophosphate (8-CPT-cAMP, 200 μm, Abcam, UK). All-trans retinoic acid treatment was renewed every day until the end of the culture to overcome the culture’s molecular instability.

Sample Preparation for SEM

To allow observation of the hydrogels using SEM, complete removal of water from the samples is necessary. Two methods of sample preparation were used and compared.

Supercritical CO₂

Hydrogels were fixed with 2% glutaraldehyde in Sorensen buffer (0.1 M, pH = 7.4) and then rinsed in distilled water. After dehydration through increasing alcohol baths up to absolute alcohol and critical point drying with a Leica EM CPD 300 critical point dryer, the sample was mounted on microscope stubs, followed by platinum sputtering (10 nm) with a Leica EM MED 020 sputter coater. Specimens were examined on an FEI Quanta 250 FEG scanning electron microscope at an accelerating voltage of 5 kV.

Freeze-Drying

The process involves two steps: (i) freezing the hydrogel samples in a bath of liquid nitrogen for 15 min and (ii) lyophilizing with LyoQuest −85 Plus, Telstar equipment (direct transition from the solid to gas state) the frozen samples for 24 h. The resulting dehydrated samples were cut to characterize the polymer structure on both its core and surface. To do that, the samples were previously metalized with 10 nm gold–palladium sputtering (PECS-GATAN equipment).

SEM is an electron microscopy technique capable of producing high-resolution images of a sample’s surface by utilizing the principles of electron–matter interactions. This microscope uses an electron beam that scans the surface of the analyzed sample. Electron beam scanning the interaction of the electron beam with matter produces backscattered or secondary electrons that are used to build an image of the surface. In our study, we focus on secondary electrons, which provide information about the surface of the sample. We used SEM (S-4800-Hitachi) with an acceleration voltage of 10 kV.

Sample Preparation for Cryo-SEM

Cryo-SEM of GelMA and HAMA was made after high-pressure freezing (HPF) as follows: a fraction of the hydrogel was cut and inserted between two HPF specimen carriers (Supplementary Data, Figure S1) dedicated to cryo-fracture (Leica), and it was loaded immediately into an ICE HPF machine (Leica). The sample was fixated and frozen within 5 ms at 2010 bar, and samples were then transferred to cryovials under liquid nitrogen. The HPF specimen carrier with the freeze solution was then inserted into the preparation chamber Quorum PP3000T. For insertion, specimen shuttle cryo stubs and the HPF specimen carrier adapter were used following the description by Payre et al.¹⁹ In this step of the sample preparation process, the samples were vitrified, resulting in an amorphous solid deprived of any crystalline structure. After the quick transfer under vacuum in the preparation chamber, the samples were fractured at −140 °C, sublimed at −90 °C for 20 min, and then coated by platinum sputtering. They were at last transferred in the cryo-SEM Quanta 250 FEG chamber and kept at −140 °C for observation with an FEI Quanta 250 FEG scanning electron microscope at an accelerating voltage of 5 kV (Figure 1iii).

Figure 1.

Electron-based imaging techniques for hydrogels. Standard SEM relies on sample dehydration and drying, including supercritical CO₂ (A) or freeze-drying (B), followed by metal coating. ESEM (C) does not require a particular sample preparation as it remains hydrated within a humidified chamber. In Cryo-SEM (D), samples are vitrified using a high-pressure freezer. For TEM (E), the imaging samples are fixed with glutaraldehyde, dehydrated without drying, and embedded in an epoxy resin before being sliced into ultrathin sections, usually less than 100 nm in thickness (70 nm) (created in BioRender.com).

Sample Preparation for TEM

Hydrogels were fixed with 2% glutaraldehyde in Sorensen buffer (0.1 M, pH = 7.4) for 1 h and washed with Sorensen phosphate buffer (0.1 M) for 12 h. They were then postfixed with 1% OsO₄ in Sorensen buffer (Sorensen phosphate 0.05 M, glucose 0.25 M, OsO₄ 1%) for 1 h, washed twice with distilled water, and stained with 2% uranyl acetate aqueous solution for 12 h. Following the method described by Kiyama et al.,²⁰ water was gradually replaced with a water–ethanol solution. Specifically, the hydrogels were sequentially washed with H₂O:EtOH mixtures in the ratios of 30:70, 15:85, 5:95, and finally 0:100%. This process was then repeated by using acetone. The samples were ultimately embedded in epoxy resin (Epon-Araldite) by following a stepwise protocol: acetone/epoxy resin mixtures of 2:1, 1:1, and 1:2, followed by pure epoxy resin, without any intermediate drying step. After 48 h of polymerization at 60 °C, ultrathin sections (70 nm) were mounted on 100-mesh collodion-coated copper grids and poststained with 3% uranyl acetate in 50% ethanol (Delta Microscopies, Mauressac, France) and with 8.5% lead citrate before being examined using CCD detectors on the Hitachi electron microscope HT7700 at an accelerating voltage of 80 kV.

Measurement of the Pore Dimension

After SEM and TEM observations, pore dimensions were measured using ImageJ software as described in reference (21). Both the largest and the smallest pore lengths were recorded. Approximately 20 measurements were taken for each image. The pore dimensions for each sample were then analyzed by using boxplot representations.

Results and Discussion

SEM

We begin this comparative study with the most common electronic microscopy technique and sample preparation methods described in the literature: supercritical CO₂ and freeze-drying with SEM.

SEM is an advanced imaging technique used to visualize the surface morphology of a wide range of samples under vacuum, including hydrogels.²²⁻²⁸ This technique uses a focused beam of electrons to scan the surface of a sample. When the electrons in the beam interact with the atoms on the sample’s surface, they cause various reactions, emitting secondary and backscattered electrons. These emitted electrons are collected by detectors, allowing reconstruction of a high-resolution image of the sample’s surface, typically up to a few nanometers. Prior to imaging, hydrogel samples are typically dehydrated and dried, manually cut, and coated with a thin conductive layer, such as gold or platinum, to prevent charging under the electron beam.²³ We used two different techniques to prepare the samples: supercritical CO₂ (Figure 1A), which is the most common technique,^29,30 and the freeze-drying process (Figure 1B), which has been described in some publications.^17,31

Supercritical CO₂ Sample Preparation

Figure 2i displays images of GelMA (10%) with 30 s (Figure 2A,i) and 2 min (Figure 2B,i) of photo-cross-linking and of HAMA with 30 s (Figure 2C,i) and 2 min (Figure 2D,i) of photo-cross-linking after supercritical sample preparation. We observe a porous nanoarchitecture of GelMA and HAMA regardless of the cross-linking time. The size of these pores is nanometric but difficult to characterize because the electron beam used by the scanning electron microscope damages the hydrogel at higher magnifications (Figure 3A). In fact, it was not possible to observe hydrogels over 35,000× in magnification without damaging the sample due to the charging effect.

Figure 2.

GelMA 10% and HAMA 3% images from (i) SEM with supercritical CO₂, (ii) SEM with freeze-drying, (iii) ESEM, (iv) Cryo-SEM, and (v) TEM. GelMA and HAMA were characterized for two times of UV exposure: 30 s (A, C) and 2 min (B, D).

Figure 3.

Pore diameter characterization: (A) from SEM images with freeze-drying sample preparation, (B) from EMEB images, (C) from EMEB images with supercritical CO₂ (D), from Cryo-SEM images, and (E) from TEM images.

Freeze-Drying Sample Preparation

Figure 2ii displays images of GelMA with 30s (Figure 2A,ii) and 2 min (Figure 2B,ii) of photo-cross-linking and HAMA with 30s (Figure 2C,ii) and 2 min (Figure 2D,ii) of photo-cross-linking after freeze-drying sample preparation. The microarchitecture of the hydrogels in the case of freeze-drying preparation is completely different in comparison with the supercritical CO₂ preparation. We observe a micrometric pore size whatever the light-curing conditions and the gel (GelMA or HAMA). Based on these observations, the average pore sizes (Figure 3B) for photopolymerized GelMA 30 s and 2 min were, respectively, measured at around 11.7 and 16.7 μm. For HAMA 30 s and 2 min, the apparent pore sizes were, respectively, measured at 19.5 and 37.6 μm. There is no correlation between pore size and cross-linking time; moreover, for the HAMA 30 s gel, the pores all appear to be oriented in the same direction (Figure 2C,ii).

Numerous publications have reported that freezing hydrogels in liquid nitrogen followed by lyophilization can damage and deform their porosity due to several factors.³² (i) Formation of ice crystals: When hydrogels are immersed in liquid nitrogen for freezing, the water present in the pores freezes rapidly but is not fast enough to prevent the formation of ice crystals. It can lead to an increase in the volume of water as it transitions from the liquid to the solid state (eutectic artifacts(34)), thereby creating mechanical stresses within the material’s pores.³⁴ (ii) Lyophilization: After being frozen, hydrogels undergo lyophilization, a low-pressure dehydration process. During lyophilization, frozen water is directly sublimated from the solid to the gas phase without going through the liquid state. This process can result in dimensional changes and deformations in the pore structure as water is removed from the material.¹⁰

Due to these combined effects, freezing in liquid nitrogen followed by lyophilization can alter the porosity and structure of hydrogels, making it challenging to obtain accurate and representative images of their microstructure. To avoid having the need for dehydrating and drying the samples, we will now study ESEM with hydrated native samples.

ESEM

ESEM allows the observation of samples in their native, hydrated state³⁵⁻³⁹ (Figure 1C). Unlike conventional SEM, which is performed in a vacuum and therefore requires samples to be dehydrated and coated with a conductive layer, ESEM can handle wet, nonconductive samples. The principle of ESEM is based on the introduction of a gaseous environment (90% humidity) in the microscope chamber under low vacuum conditions (700 Pa). The gaseous environment prevents the rapid dehydration of the specimen, maintaining it in a nearly native state during imaging. This capability is achieved through the use of a differential pumping system that maintains a gaseous environment around the sample while keeping the electron column under a vacuum. When the electron beam interacts with the sample, secondary and backscattered electrons are emitted and detected by specific detectors adapted to the low-pressure environment. The secondary electrons that form the signal come from the molecules of water inside the microscope chamber. They are created by secondary electrons extracted by the primary electron beam. The technique provides detailed information about surface topography and morphology while preserving the native state of the sample. This makes ESEM particularly valuable in fields such as biology, materials science, and engineering, where maintaining the natural condition of the sample is crucial.

Figure 2 displays images of GelMA with 30 s (Figure 2A,iii) and 2 min (Figure 2B,iii) of photo-cross-linking and of HAMA with 30 s (Figure 2C,iii) and 2 min 30 s (Figure 2D,iii) of photo-cross-linking obtained with ESEM. There appears to be a microarchitecture, but it is difficult to distinguish it and impossible to quantify the pore size whatever the photopolymerization conditions and the gel (GelMA or HAMA) (Figure 3C). Observing the pores of hydrogels using ESEM can be challenging due to several factors. Hydrogels are highly water-absorbent materials, and their high water content can lead to significant electron scattering and charging effects, making it difficult to obtain clear and detailed images of the pore structures.⁴⁰ One primary issue with this observation technique is the achievable magnification. In fact, it was impossible to exceed 2000× magnification without having the current density degrading the sample. As a result, finer details, such as nanometric pores, may not be well resolved in the ESEM images.

Additionally, while the water vapor in the chamber can be controlled, users often lower it to achieve a better image resolution. However, lowering the water vapor also extracts water from the sample, causing alterations, such as shrinkage of the hydrogels.

However, to observe the internal pore structures in greater detail, other techniques, such as Cryo-SEM and TEM, might be more suitable. These techniques are compared below.

Cryo-SEM

Cryo-SEM allows the observation of vitrified hydrogels (Figure 1D), preserving their 3D microstructure without dehydration.^{19,24,33,41−43} The principle of Cryo-SEM sample preparation is based on rapid freezing of the hydrogel sample in liquid nitrogen at cryogenic temperatures (−196 °C) and at high pressure (2000 bar) (Figure 1D). Once frozen, the sample is fractured to expose its internal structures, sublimated to remove surface ice, and then coated by platinum sputtering. This process results in the vitrification of water, preventing the formation of ice crystals that could damage the hydrogel’s micro- and nanostructure. The vitrified hydrogel sample is then imaged using SEM under cryogenic conditions (−140 °C). In Cryo-SEM, an electron beam scans the surface of the frozen sample, causing secondary electrons to be emitted from the surface. These secondary electrons are then detected to produce high-resolution images. Cryo-SEM is often considered more appropriate than conventional SEM for the observation of hydrogels due to the several advantages it offers^39,43: (i) preservation of the native state: Cryo-SEM allows hydrogel samples to be observed in their fully hydrated and near-native state; (ii) elimination of drying artifacts; and (iii) allows observations of the inside of the sample; Figure 2 displays the images of GelMA with 30 s (Figure 2A,iv) and 2 min (Figure 2B,iv) of photo-cross-linking and of HAMA with 30 s (Figure 2C,iv) and 2 min (Figure 2D,iv) of photo-cross-linking obtained with Cryo-SEM. Observation of GelMA with Cryo-SEM shows a porous microarchitecture whatever the light-curing conditions. The observed pore size averages for photopolymerized GelMA 30 s and 2 min are, respectively, 126.1 and 43.2 nm. For HAMA 30 s, it is 28.7 nm. Pore size could not be determined in the HAMA 2 min condition without sample damage (Figure 3D). The maximum reachable magnification with the Cryo-SEM for imaging of hydrogels was 70,000× to avoid degrading the sample due to the charging effect.

TEM

TEM is well described as an imaging technique used to visualize the internal structure and nanoscale features of biological samples.⁴⁴⁻⁴⁷ In TEM, a focused electron beam is transmitted through ultrathin sections (70 nm) of the hydrogel sample.^48,49 Hydrogel samples for TEM observation are typically fixed with glutaraldehyde, dehydrated, embedded in an epoxy resin, and stained with uranyl acetate and lead citrate before being sliced into ultrathin sections, usually less than 100 nm in thickness (70 nm) (Figure 1E). These thin sections allow the electron beam to pass through the hydrogel, generating a variety of electron-transmitted patterns. Detectors collect these transmitted electrons, and a detailed image of the hydrogel’s internal structure is reconstructed based on the intensity and distribution of the electrons. Because of the small probe size, the beam's narrow energy distribution, and the small thickness of the sample, TEM provides exceptionally high resolution, allowing one to investigate the nanostructure, polymer network organization, and interactions with encapsulated molecules within the hydrogel. Figure 2 displays images of GelMA with 30 s (Figure 2A,v) and 2 min (Figure 2B,v) of photo-cross-linking and of HAMA with 30 s (Figure 2C,v) and 2 min (Figure 2D,v) of UV exposure obtained with TEM. With the images obtained using TEM, we can clearly observe nanopores in both GelMA and HAMA hydrogels, regardless of the cross-linking time. The average pore sizes for photopolymerized GelMA 30 s and 2 min are, respectively, 82.9 and 48.3 nm. For HAMA 30 s and 2 min, they are 41.6 and 15.5 nm, respectively (Figure 3E).

In both cases, increasing the photopolymerization time results in a decrease in the observed pore size. This observation can be legitimately correlated with a higher degree of cross-linking in the hydrogel during the 2 min photopolymerization, leading to an increase in cross-linking nodes. Regarding TEM resolution, we were able to observe the hydrogels between 200× and 150,000× in magnification. This extended window of observation allowed us to see that hydrogels do not have several levels of porosity with different sizes, and their porosity is nanometric (Supplementary Data, Figure S2).

TEM for Cell or Spheroid Hydrogel Encapsulated Observation

TEM also stands out for its exceptional capabilities in visualizing cellular structures at the nanoscale level.⁴⁴ This imaging technique offers a unique advantage to simultaneously visualize cells at high resolution and the hydrogel microarchitecture. To highlight this potential of TEM for investigating cell–hydrogel interactions, cells were encapsulated in the hydrogel and cultured for several days before TEM observations.

During sample preparation for observation by TEM, thicker sections (approximately 1 μm) can be utilized to facilitate sample observation within the resin block. Staining these sections with methylene blue enables macroscopic visualization of samples, allowing for the observation of cells and the gel in bright-field microscopy (Figure 4A,B). These observations make it possible to observe the distribution and morphology of the cells in the gel sample as a whole and to see how the cells have migrated in the hydrogel during cell culture before proceeding to observe subsequent sections under TEM.

Figure 4.

(A, B) Histological images of 1 μm sections of 12% GelMA with encapsulated adipose stem cells (ASCs) after 7 days in culture. These sections were stained with methylene blue. (C) TEM images of the same 12% GelMA sample with encapsulated ASCs after 7 days of culture. (C.1) Overview of several cells surrounded by GelMA and degraded GelMA regions. (C.2) Higher-magnification image of C.1 (yellow frame) illustrating the differences in intensity in the TEM images between the GelMA and degraded GelMA areas. (C.3, C.4) Higher magnification further into the microarchitecture of GelMA (red frame) and degraded GelMA (green frame), respectively. (D) TEM images of the 10% GelMA samples with ASC spheroid cells after 28 days in culture. (D.1) Cells surrounded by collagen fibers (CFs, white arrows). (D.2) Higher-magnification image of D.1 (red frame) on CFs. (D.3) Zoomed-in view of the microarchitecture of the fibers, showing the typical banded pattern of CFs.

Moreover, unlike SEM or Cryo-SEM, which primarily provides topographic information, TEM allows for deeper penetration into samples, thereby revealing the internal details of cells encapsulated within hydrogels. This ability to visualize cellular and hydrogel structures at subnanometer resolution offers a unique perspective for studying cell–hydrogel interactions and for optimizing material design for biomedical applications. Figure 4 shows adipose-derived stem cells (ASCs) or ASC spheroids encapsulated, respectively, in a 12% GelMA mixture after 7 days of culture (A–C) and 10% after 28 days of culture (D). Histological images (A, B) reveal cells and positioning within the GelMA matrix after 7 days, while TEM images (C) show GelMA degradation. After 28 days (D), the synthesis of the endogenous matrix fibers by ASC spheroids is observed, with the TEM image (D.3) showing fibers resembling collagen but not showing the characteristic 68 nm banding collagen pattern. Instead, it presents a banding pattern in the range of a few tens of nanometers. This may suggest an incomplete self-assembly related to the mechanical constraints of the spheroid encapsulation within the GelMA matrix. In fact, Feng and co-workers⁵⁰ showed a 26 nm D-band in TEM due to the effects of the in vitro mechanical microenvironment on collagen self-assembly, very similar to what we observed. This hypothesis is also supported by the subsequent work of Rowe and Röder⁵¹ and Buehler and Wong⁵² et al. on the effects of mechanical properties on collagen assembly. These images also suggest that the remodeling of the hydrogel by the cells is consistent. Consequently, the use of TEM for observing cells within hydrogels provides a significant advantage over surface electron microscopy techniques, paving the way for valuable advancements in the fields of tissue engineering and regenerative medicine.

Since the HAMA hydrogel lacks adhesive motifs required for effective cell attachment and survival, it is not a suitable hydrogel alone for cell adhesion. For this reason, no comparative studies have been carried out with the HAMA hydrogel and cells.

Conclusions

In this comprehensive comparative analysis, we have assessed four electron microscopy techniques associated with five preparation sample techniques—SEM, ESEM, Cryo-EM, and TEM—for the characterization of protein-based and polysaccharide-based hydrogels microarchitectures, particularly focusing on GelMA and HAMA hydrogels, which are pivotal in tissue engineering and 3D bioprinting applications.

Each electron microscopy technique offers distinct advantages and limitations in visualizing hydrogel structures (Table 1).

Table 1. Advantages and Disadvantages of Each Technique for Hydrogel Microarchitecture Characterization.

SEM showcased a porous microarchitecture for GelMA and HAMA hydrogels but suffered from sample preparation-induced alterations.

ESEM proved beneficial for observing surface details of hydrated hydrogels, providing insights into their microstructure under near-native conditions. However, challenges such as high electron scattering and limited imaging depth posed constraints in visualizing finer details and internal structures, preventing the observation of nanoporous features.

High-pressure freeze fixation and Cryo-SEM have emerged as effective methods for preserving the near-native structure and revealing fine details of the hydrogel microarchitecture without dehydration artifacts; however, these techniques are limited by the small sample size (1 mm³) and are both time- and cost-intensive in terms of sample preparation. Moreover, we can observe sample damages at high magnification (>70,000×). The resolution limits of Cryo-SEM make it unsuitable for observing nanoscale interactions, such as those between cells and their surrounding matrix.

TEM provided exceptional resolution for observing the nanometric features of the hydrogels. This technique, when combined with histological methods, offers a comprehensive view that encompasses the macroscopic and nanoscopic scales.

Taking into account the in-depth analyses of various electron microscopy techniques, TEM stands out as the most favorable compromise for studying the microarchitecture of hydrogels at this stage. This method offers exceptional resolution, detailing nanostructures and the polymer network organization within hydrogels.

By comprehensively evaluating these electron microscopy techniques, we aim to guide researchers in selecting suitable methods for imaging the hydrogel microarchitecture. Moreover, understanding the impact of sample preparation methods on the hydrogel structure is crucial for improving imaging techniques and advancing tissue engineering applications. In this study, we selected concentrations of 10 and 12% for GelMA and 3% for HAMA to represent a range commonly used in tissue engineering applications.^{12,17,53−56} However, the hydrogel composition, along with the presence and density of cells, can significantly influence artifacts by locally altering the microarchitecture and introducing variations in the hydrogel’s mechanical and hydration properties. This highlights the importance of systematic future studies to better understand how polymer concentration and cell density affect processing artifacts, which would further complement the findings of this work.

In conclusion, the selection of an electron microscopy technique should consider the trade-offs among resolution, sample preparation-induced alterations, and the ability to visualize different scales of hydrogel microarchitecture. Further advancements in sample preparation methodologies and imaging techniques will aid in unveiling a more comprehensive understanding of hydrogel structures, facilitating the design of advanced biomaterials for tissue engineering and regenerative medicine applications.

Glossary

Abbreviations

GelMA

gelatin methacrylate

HAMA

hyaluronic acid methacrylate

SEM

scanning electron microscopy

TEM

transmission electron microscopy

Cryo-SEM

cryo-scanning electron microscopy

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.4c08096.

• Additional experimental details and microscopic observations;Figure S1, sample preparation process for Cryo-SEM in HPF specimen carriers; and Figure S2, TEM observations at different magnifications for GelMA cross-linked for 30 s and HAMA cross-linked for 30 s (PDF)

The authors declare no competing financial interest.