A subscription to JoVE is required to view this content. Sign in or start your free trial.
Research Article
We propose a protocol for assessing bone regeneration at various hierarchical levels. This multimodal approach allows for structural analysis across different length scales, effectively addressing challenges in sample preparation. It also provides a reliable workflow for evaluating the effectiveness of biomaterials in promoting bone regeneration.
The clinical success of an implant largely depends on its ability to promote bone repair and facilitate osteointegration. While significant advancements have been made in designing new scaffolds, a thorough evaluation of implant success at various bone length scales has not been sufficiently investigated. Bone is a hierarchically organized tissue, with an intricate structure that ranges from the organization of collagen fibrils and their mineralization at the nanometer and micrometer scales to the macroscopic arrangement of compact and cancellous bone. This hierarchical organization is crucial for ensuring bone function and mechanical performance. In this context, a comprehensive assessment of the effects of biomaterials on bone repair at multiple length scales is essential for evaluating the effectiveness and safety of implanted materials. However, preparing samples for integrated multimodal analyses poses a technical challenge. In this work, we present a protocol designed to evaluate bone regeneration across various hierarchical levels and analyze the interface between newly formed bone and the implanted scaffold. This approach provides a robust method for assessing the effectiveness of scaffolds in inducing bone repair. The combination of selected 3D and 2D imaging techniques is crucial for a proper evaluation of the complex mineralized tissue that forms during the bone healing process
Bone is a complex, hierarchically organized tissue1,2. Understanding its structural organization is essential for addressing fundamental questions related to how cells contribute to tissue architecture and how structural changes impact mechanical function3. In this context, bone remodeling after injury remains a central topic of investigation. Fracture healing is a complex multistage process, beginning with an initial inflammatory response, followed by angiogenesis and the differentiation of progenitor cells into chondrocytes and osteoblasts4,5. Much has been debated regarding the role of biomaterials to stimulate osteointegration, particularly about biocompatibility and their ability to prevent fibrous encapsulation6.
A commonly adopted method to assess biomaterial-stimulated osteointegration is optical microscopy, which enables both quantitative and qualitative histological analyses. This approach has been shown to be a reliable and effective way to evaluate the success of bone implants7,8. Nevertheless, several issues related to bone hierarchical organization cannot be explored by using a single imaging approach3, particularly the histological one, which is limited at least in part by the magnification restriction of the method.
Advances in three-dimensional imaging have proven essential for understanding bone hierarchical organization. Techniques such as microcomputed tomography (microCT; either synchrotron radiation or conventional laboratory-based X-ray sources)9,10,11, dual-beam scanning electron microscopy (FIB-SEM) associated with slice-and-view tomography12,13,14 and transmission electron tomography15,16 provided new insights into bone architecture at multiple scales.
Based on different 3D imaging approaches, a model of bone hierarchical organization has been proposed, which initially identified nine distinct levels of bone structural organization1. This scheme has recently been revised and expanded to encompass approximately 12 hierarchical levels, serving as a valuable reference for exploring the relationship between structure and function in bone tissue2,17. Importantly, this proposed hierarchical model provides a foundation for evaluating the success of bone repair following injury. Evaluating whether the regenerated bone tissue reproduces the native hierarchical structure is a key standard in assessing the quality of bone healing, particularly when biomaterials are used within the damaged site.
Although the use of multiple imaging techniques to investigate bone organization has been discussed6,7,18,19,20, little has been explored regarding a workflow that integrates different imaging modalities to assess bone mineral organization across multiple spatial scales. As mentioned, bone is a complex composite tissue, consisting of an organic phase (cells, collagen, and macromolecules) and a mineralized matrix. From a sample preparation perspective, this complexity presents technical challenges regarding solvent permeation through the sample, sectioning of a hard specimen, and the integration of different imaging strategies on the same sample.
The key innovation in this work is the establishment of a reproducible, correlative, and multiscale imaging workflow using three-dimensional imaging and complementary techniques. As illustrated in Figure 1A, the workflow is designed so that the same sample can be sequentially examined using different imaging modalities. Because some of these techniques are destructive or require sectioning, the workflow follows an order that begins with non-destructive methods and progresses to approaches that involve cutting or material removal.
This sample preparation and analysis workflow is broadly applicable to hard or mineralized specimens, including cortical and trabecular bone, pre-implantation scaffolds, and the evaluation of the bone/biomaterials interface after implantation. By combining imaging modalities that span different fields of view and resolution ranges (Figure 1B), the workflow enables the analysis of samples from the centimeter scale down to the micrometer and nanometer scales. While the sequence is optimized to support correlative analysis of a single specimen, each technique can also be applied independently according to the experimental goal. This approach enables a detailed assessment of biomaterials' osteointegration and the evaluation of the regenerated bone growth in the healing area.
The results presented in this protocol were obtained from an experimental procedure conducted in a previous study21. All procedures were carried out in accordance with the ethical guidelines of the Ethical Committee on the Use and Care of Experimental Animals at Fluminense Federal University, Brazil (CEUA/UFF: 779). Briefly, 20 male Wistar rats weighing between 300 and 400 g were randomly assigned to two groups: Sham and cHA-microsphere implantation. All animals underwent surgery in which a 2 mm defect was created in the tibial diaphysis. The Sham group served as a clot control to evaluate bone regeneration in a non-critical defect, whereas the experimental group received CHA37 microspheres implanted into the defect site. After 7 and 21 days, animals were euthanized with an overdose of general anesthesia, and tibial bone blocks from each group and time point (7 and 21 days) were collected. Bone quality and the extent of newly formed bone were assessed using optical microscopy and microCT.
1. Sample fixation and resin embedding
2. Sample visualization using microcomputed X-ray tomography (microCT)
NOTE: microCT is used to assess implant positioning and bone regeneration. It also helps identify potential regions of interest for further analysis. This step should be performed with the sample embedded in resin.
3. Sample preparation for optical microscopy
4. Sample preparation for scanning electron microscopy
5. Alignment, filter processing, and segmentation of FIB-SEM slice-and-view tomography
6. Sample preparation for transmission electron microscopy
The initial analysis of resin-embedded samples is performed using microcomputed tomography (microCT). This non-destructive technique facilitates 3D visualization of the entire defect area, enabling both quantitative and qualitative analyses of the sample. Contrast in microCT imaging derives mostly from the density differences of the materials present in the sample. Figure 1 illustrates the microCT analysis of a non-critical defect in a rat tibia that was implanted with the CHA37 biomaterial. The sample was collected 12 weeks post-implantation and prepared following the resin-embedding protocol previously described. The microCT analysis produced a 3D model that reveals the internal structures of the sample (Figure 3A-B) and allows for the differentiation between bone, biomaterial, resin, and newly formed bone (Figure 3C-D). By analyzing parameters such as the volume, distribution of material fragments and the relative volume occupied by trabecular bone versus biomaterial, we can effectively characterize the interaction between bone and material at the micrometric scale24. As an example of such analysis, the work conducted by Zelaya et al21 explored the trabecular bone architecture in tibial defects 21 days after CHA microsphere implantation. Their results showed that the microspheres disaggregate into smaller fragments ranging from 20 to 400 µm in diameter. Additionally, trabeculae were found to form direct contact with the surface of approximately 70% of these fragments and exhibited a relatively uniform thickness of about 31 µm. In contrast, the average thickness of trabeculae developing in defects without biomaterial reached 58 µm at the same implantation time. Altogether, these findings illustrate how 3D measurements can provide important complementary information about the bone regeneration process.
After microCT imaging, the resin-embedded samples (Figure 4A) are sectioned for optical microscopy analysis. Sectioning can be performed either transversally (Figure 4B, orange rectangle) or longitudinally along the bone's long axis (Figure 4B, blue rectangle). Transversal sectioning allows the production of serial sections across the entire defect volume, while longitudinal sectioning provides a single section of the defect. It is important to note that the success of transversal serial sectioning heavily relies on the operator's skill and the thickness of the diamond disk used. Here, we exemplified the expected results by obtaining a single longitudinal section of the defect area (Figure 4C). The sectioning and polishing steps are critical and must be conducted with careful attention to prevent damaging the samples. In this protocol, we utilized a custom-made aluminum tube assembly, which consists of an inner solid cylinder and an outer hollow cylinder (Figure 4D). The sample is mounted onto the inner solid cylinder using mounting wax. This inner component is then inserted into the outer hollow cylinder (Figure 4E-F), ensuring that the sample faces the sandpaper. This setup allows for controlled thinning and polishing of the sample by applying a uniform pressure determined solely by the weight of the inner cylinder, preventing excessive or variable force during polishing.
Analysis through optical microscopy represents a key step in sample localization and in establishing a correlative bridge across different imaging modalities (Figure 5). After polishing and initial examination on the optical microscope, intrinsic sample features, such as defect regions, areas containing biomaterial, and bone morphology, allow the region identified under the light microscope to be matched to the corresponding location in the microCT dataset (Figure 5A). Following this first-level identification, the sample is further examined using polarized light microscopy (Figure 5B-C), which provides valuable insights into the organization of collagen fibers in newly formed bone. The thickness of the sample plays a crucial role at this stage. Thick sections affect light transmission, limiting the evaluation of collagen orientation. As the same sample observed in optical microscopy is subsequently evaluated by dual-beam scanning electron microscopy (FIB-SEM), the correlation regions of interest still rely largely on sample intrinsic structural characteristics (Figure 5E). However, if additional cutting or trimming is required between optical and FIB-SEM analysis, appropriate strategies should be considered to maintain accurate correlation across techniques. In such cases, structural landmarks and fiducial markers (micro-indentations or laser marks) can greatly facilitate the precise relocation of the region of interest.
FIB-SEM allows for high-resolution evaluation of the interface between the newly formed bone and the biomaterial. In this technique, the sample is first imaged using the scanning electron microscope (SEM) mode, enabling precise localization of areas of interest (Figure 5E). Following this, a volume of the selected region can be further examined using slice-and-view tomography (Figure 5F-G). This technique involves sequential milling and imaging of the sample. The volume analyzed through slice-and-view tomography varies depending on the specific equipment and its intended use.
After acquiring slice-and-view tomography data, the image stack is aligned for evaluation. Additional filters can be applied to the aligned stack to enhance the visualization of features of interest. In this study, we applied a sequence of four image-processing filters available in the software to improve overall image quality (Figure 4). First, a Fast Fourier Transform (FFT) filter is applied to remove artifacts caused by gallium beam sectioning during FIB milling (Figure 6A-B, arrows). The curtain artifacts appear on the final image as vertical lines and severely affect the visualization of the features of interest. The artefacts arise from uneven milling rates related to surface roughness and can be partially mitigated by GIS deposition or prior sample polishing. Because curtaining produces a strong directional frequency component, FFT filtering is the most effective first step. If not corrected, these artifacts propagate through subsequent processing steps, being amplified during the use of filters that improve contrast.
Next, the Shading Correction Wizard module was used to address uneven illumination across the block surface (Figure 6B-C, asterisk), which arises from detector geometry, charging effects, or slight variations in beam current during slice-and-view acquisition. Performing shading correction after FFT filtering ensures that only the true background gradient is corrected, without interference from high-frequency curtaining patterns. Following this, a histogram equalization filter was applied to enhance image contrast (Figure 6D). The use of a histogram equalization filter after correcting the curtain effects and the uneven illumination allows the amplification of sample structural details without unintentionally boosting artifact-related intensity variations. While this step improves visual distinction between structures, it also increases image noise, leading to pixel intensity variability that may impact automated segmentation strategies. To mitigate this issue, we employed a denoising filter (Non-Local Means) to reduce image noise while preserving structural details (Figure 6E). A comparison of the same region of interest (ROI) in both the raw and fully processed images shows how the sequential application of these filters significantly enhances image quality and overall interpretability (Figure 6F). Notably, these filters are also available in other widely used software platforms for electron microscopy image analysis, including FIJI (ImageJ), an open-source software. As the algorithms and operations are equivalent, the expected results (using identical parameters) are consistent across software platforms. The main differences are related to software interface, computational requirements, and licensing constraints.
After enhancing the image quality using various filters, segmentation strategies can be applied to the aligned volume stack. Although segmentation can be performed manually, this approach can be time-consuming (see Figure 7A). When there is enough contrast between the features of interest, threshold-based segmentation can be quite effective. However, this method may also capture unwanted features with similar pixel densities, requiring additional manual refinement (Figure 7B). Threshold-based segmentation tools are available in several software packages, including Avizo, Dragonfly, IMOD, and FIJI. As with the image-processing workflow, the segmentation outcomes obtained with these platforms are generally consistent, with major differences arising from interface design and licensing constraints. Therefore, users should verify software requirements and availability through each platform's official website.
Once the segmentation is complete, the structures of interest are rendered in a 3D model (Figure 7C). This enables the visualization of features such as the osteocyte canalicular network. Furthermore, three-dimensional rendering can be enhanced by software-specific analytical modules, like the Thickness Map, allowing the measurement of canalicular diameter throughout the entire osteocyte canalicular network (Figure 7C-D). Although the 3D analyses in this work were performed using Avizo, it is important to note that equivalent functionality is available in FIJI through the Local Thickness plugin. For additional details about its implementation, consult the plugin documentation (https://imagej.net/imagej-wiki-static/Local_Thickness).
TEM analysis allows for the examination of the bone/biomaterial interface at the nanoscale. This analysis can be combined with selected area electron diffraction (SAED), energy dispersive spectroscopy (EDS), and electron tomography to provide a detailed characterization of the sample. Figure 8 shows an interface area of a sample that was sectioned using a diamond knife. Note that fractures are observed on the biomaterial in the direction of the cutting (Figure 8A, arrows). SAED patterns are crucial for distinguishing between bone and biomaterial regions (Figure 8B-C). By indexing the electron diffraction patterns acquired in TEM, it is possible to identify the crystalline phases present in the sample based on the analysis of interplanar distances. In this case, both the biomaterial (Figure 8B) and the bone (Figure 8C) are composed of hydroxyapatite crystals. However, in the bone areas, the 002 ring appears incomplete, indicating that bone crystals grow according to collagen orientation. Additionally, EDS analysis can provide elemental information about the sample (Figure 8D-G). The bone/biomaterial interface can also be further investigated through electron tomography, which offers a 3D analysis at the nanoscale concerning implant osteointegration.
Figure 1: Integrated sample preparation and imaging techniques used for correlative analysis. (A) Sequential sample preparation steps, beginning with specimen retrieval and proceeding through the preparation required for each imaging technique. (B) Overview of the field of view and spatial resolution associated with each technique. Please click here to view a larger version of this figure.
Figure 2: Schematic representation of the reflection-based thickness measurement procedure. (A) System calibration using a marked coverslip. The microscope focus is first adjusted to the lower mark and then to the upper mark; the displacement observed on the fine-focus scale corresponds to the coverslip thickness. (B) Measurement of sample thickness. After placing the sample on the stage, the focus is adjusted sequentially to the bottom and top surfaces of the sample. The difference in the fine-focus scale readings reflects the actual sample thickness. Please click here to view a larger version of this figure.
Figure 3: Microcomputed X-ray tomography evaluation of a rat tibia defect implanted with CHA37 biomaterial, 12 weeks post-implantation. (A) A representative 3D volume of a longitudinal section of the bone is shown, highlighting the defect and the implanted biomaterial. (B) The same volume as in panel A, rendered with higher-density pixels that correspond to the biomaterial fragments. (C) 2D cross-section of a virtual transversal slice (red line in A), depicting different gray levels in cortical, immature bone, biomaterial, resin, and background. (D) Gray value variation from a line profile (yellow line in C), illustrating the intensity levels of the different materials present in the sample. Please click here to view a larger version of this figure.
Figure 4: Sample sectioning and polishing. (A) Resin-embedded sample. (B) Schematic representation of possible sectioning approaches to the sample. Sectioning can be performed either transversely (indicated by the blue rectangle) or longitudinally (indicated by the orange rectangle). (C) Longitudinal section of a tibial defect. (D-E) Custom-made cylinder system used for sample thinning and polishing. The sample is first attached to the inner solid cylinder, which is then inserted into the hollow outer cylinder with the sample surface facing the sandpaper. (F) Diagram illustrating how the assembly of the components is configured. Please click here to view a larger version of this figure.
Figure 5: Correlation of the defect area using different imaging techniques. (A) Microcomputed X-ray tomography 2D slice imaging of the defect area. Biomaterial (BM) fragments are observed at the center of the defect, whereas trabeculae related to bone repair are seen close to the defect border. (B-D) Optical microscopy analysis of a polished section of the sample. Bright field imaging highlights both biomaterial fragments and bone. (C) Polarized light microscopy differentiates the preexisting bone, shown in blue, from the newly formed bone during healing, depicted in orange. A first-order lambda plate with the higher refractive index parallel to the elongated blue lines in the figure was used. (D) Higher magnification of the area selected in C. An interface of bone growing in close apposition to a preexisting cortical bone, but with type I collagen fibers oriented perpendicular (thin yellow region), is shown on the dashed circle. (E) Scanning Electron Microscopy imaging in Secondary Electron (SE) mode. (F) Trench-milled region for slice-and-view tomography acquisition. The bright area within the trench indicates where platinum was previously deposited to protect the area designated for milling. (G) Face of the trench displaying bone morphology seen on a growing trabecula. BM: biomaterial. Please click here to view a larger version of this figure.
Figure 6: Image filtering corrects Focused Ion Beam-Scanning Electron Microscopy (FIB-SEM) artifacts and enhances feature distinction. (A-F) Raw FIB-SEM images display characteristic artifacts such as curtain effects (A, arrows) and uneven illumination (A-B, asterisk). These artifacts were reduced by sequentially applying an FFT filter (B) followed by the Shading Correction Wizard module (C). Afterward, a contrast enhancement filter (D) and the Non-Local Means denoising filter (E) were applied. As result, the internal features of the sample are more clearly distinguished in the final processed image compared to the raw image (F). Please click here to view a larger version of this figure.
Figure 7: Imaging segmentation and 3D model display. (A) Features of interest can be manually segmented using image analysis software. (B) If the features show sufficient differences in pixel intensity, threshold-based segmentation can be applied. (C) After segmentation, a 3D rendered model is generated. Additional analyses, such as the Thickness Map module available in the software, can be utilized to associate measurements, like canalicular diameter, with the rendered structures. (D) Thickness map results were plotted on a frequency histogram of canalicular diameter. The mean diameter was 166.1 ± 60.66 nm. Please click here to view a larger version of this figure.
Figure 8: Transmission Electron Microscopy analysis of the bone/biomaterial interface. (A) A fragment of the biomaterial (BM) is surrounded by mineralized tissue. Cracks in the biomaterial, caused by ultramicrotomy, are indicated by arrows. (B-C) Selected Area Electron Diffraction (SAED) patterns acquired from regions corresponding to the biomaterial and mineralized tissue, respectively. (D-G) Energy-dispersive X-ray spectroscopy (EDS) analysis of the interface. The region of interest is shown in a Scanning Transmission Electron Microscopy (STEM) image (D), along with corresponding elemental maps for phosphorus (P) and calcium (Ca; E-F). (G) The global EDS spectrum from the imaged area shows characteristic peaks for phosphorus (P) and calcium (Ca). Please click here to view a larger version of this figure.
Different imaging techniques have been employed to investigate bone architecture and assess the effectiveness of biomaterials in inducing osteointegration7,21,22,23,24. In this protocol, we established a workflow that incorporates X-ray microtomography, optical microscopy, and electron microscopy to analyze the mineralized bone tissue formed in a rat tibia defect following biomaterial implantation. Executing this workflow requires substantial commitment and access to specialized instrumentation. While microCT and optical microscopy are relatively accessible and involve moderate acquisition times, high-resolution approaches such as FIB-SEM tomography, lamella preparation, and TEM require more advanced equipment. Moreover, data processing for slice-and-view tomography is labor-intensive and can extend over several months for a single dataset, particularly during 3D reconstruction and segmentation. However, bone regeneration studies typically rely on a large number of samples as biological replication can increase data variability and compromise statistical robustness. To address this challenge, it is recommended that all samples undergo an initial global assessment by microCT. This technique provides a rapid, non-destructive, and reliable evaluation of defect healing and biomaterial performance25,26. Following segmentation, biomaterials and newly formed bone volume can be quantified across all samples, enabling comparison between groups and evaluation of reproducibility. Based on these microCT outcomes, a subset of representative specimens can then be selected for more detailed analyses using the subsequent imaging modalities described in this protocol. This two-step strategy ensures adequate biological replication during the initial screening while allowing comprehensive multiscale characterization of representative samples.
Importantly, the trabecular bone parameters obtained by microCT, such as trabecular thickness and inter-trabecular angles, represent valuable data for establishing correlation models with mechanical test outcomes27. Previous studies have successfully integrated microCT-derived volumetric information with finite element analysis to estimate the mechanical properties of fracture callus tissue, demonstrating its usefulness in fracture-healing research27,28,29,30. Additionally, recent work by Reznikov et al. showed that quantifying inter-trabecular angles provides important insights into the relationship between trabecular microarchitecture and normal locomotive function31. The parameter regarding the inter-trabecular angles was later applied to evaluate bone regeneration in the presence and absence of nanostructured cHA/alginate microspheres. Notably, similar inter-trabecular angle motifs were observed in both the clot and biomaterial-treated groups, indicating that the biomaterial does not interfere with the intrinsic topological organization of trabecular bone during regeneration11.
One of the strengths of this protocol is its adaptability; the methodology can be customized for different types of biomaterials, anatomical locations, or animal models. Although the representative results presented in this protocol are based on the analysis of CHA37 samples, recent reviews have broadly demonstrated the applicability of the techniques employed here to other biomaterial systems, including titanium-based implants and hydroxyapatite ceramic scaffolds6,18,19,32,33. Each technique employed in this protocol has distinct advantages, but there are also limitations associated with each method. Key limitations include issues related to sample preparation, artefact generation, and image resolution, which should be considered when interpreting the data obtained from each technique. A major difficulty lies in evaluating the mineral and organic parts of the sample simultaneously. In this protocol, we focused on the analysis of mineralized tissue. Nevertheless, some organic counterparts of the sample can also be observed. For example, the osteocyte lacuna can be identified in the different approaches used in this work. Furthermore, the post-staining of the sample with uranyl acetate also highlights the organic part of the sample, which includes intracellular organelles, cells that are not embedded in the mineralized tissue, and non-mineralized collagen matrix.
Another critical limitation in sample evaluation is the resolution constraints of each technique. While conventional laboratory-based microCT using an X-ray tube provides a detailed overview of bone regeneration in the defect area, it is not capable of accurately resolving osteocyte lacunae or early woven bone trabecular-like structures due to these resolution limitations. In such cases, synchrotron-based microCT (SR-microCT) is recommended for higher-resolution analyses34.
Optical microscopy offers better resolution for identifying the bone/implant interface. The addition of polarized light microscopy provides further insights into collagen fibril orientation, a feature not easily observed using the other techniques in this protocol. The same area examined by optical microscopy can be subsequently mounted on a stub for evaluation by scanning electron microscopy, allowing a 2D visualization of a pre-defined region of the sample.
Considering analysis of smaller regions of the sample, FIB-SEM slice and view tomography, and transmission electron tomography can provide a detailed evaluation of the interface at the nanoscale. This analysis is important to distinguish trabecular and cement sheath structures35 . Furthermore, the use of FIB-SEM, either for the acquisition of slice-and-view tomography or for lamella preparation for electron microscopy, allows the operator to choose the regions of interest in the sample, giving a more user-directed analysis of regions of interest19.
The visualization of the sample's finer structural details is achieved through transmission electron microscopy (TEM). Sample preparation for TEM represents a bottleneck in the study of the implanted biomaterials. The main limitation relates to the hardness of biomaterials. Titanium alloys and ceramic-based biomaterials, for example, cannot be sectioned with a diamond knife in an ultramicrotome. Softer biomaterials, such as hydrogels or the HA/alginate microspheres used in this study, can be sectioned. Nevertheless, even in these cases, ultramicrotomy frequently produces artifacts, including cracks within the biomaterial36. Figure 8A illustrates a typical sectioning artifact observed after ultramicrotome cutting. In addition, attempting to section samples harder than resin-embedded cells reduces the lifetime of the diamond knife. For these reasons, lamella preparation using FIB-SEM offers an optimized approach for evaluating the bone/biomaterials interface. Compared with ultramicrotomy, FIB-SEM enables the targeted extraction of regions of interest with high spatial precision and minimal mechanical damage. Although FIB-SEM preparation also introduces certain artifacts, such as curtaining effects, redeposition, ion implantation, and amorphization, it remains a valuable method for preparing hard samples for TEM analysis37. In general, FIB-milled sections offer a smaller field of view (approximately 5 µm x 6 µm) compared with ultramicrotomy (several hundred micrometers)38.
The TEM approach combines morphological analysis at the nanoscale with crystallographic information obtained through electron diffraction and high-resolution TEM (HRTEM). Furthermore, elemental information can be collected using energy-dispersive X-ray spectroscopy (EDX) and electron energy-loss spectroscopy (EELS). While EDX can be performed using scanning electron microscopy (SEM), conducting EDX in scanning transmission electron microscopy (STEM-EDX) offers improved resolution39. Additionally, three-dimensional analysis of the bone/biomaterials interface at the nanoscale can be explored by electron tomography18,40. In this context, the work by Grandfield et al. explored the bone/implant interface in titanium alloys by combining elemental EDS analysis with electron tomography. Their study provided important insights into the interface between bone and titanium dioxide surfaces, demonstrating how correlative multimodal approaches can reveal features that are not accessible through conventional two-dimensional imaging41.
The multimodal workflow employed in this study provides complementary insights that would not be achievable with a single imaging technique. As mentioned, microCT represents an important approach for assessing global biomaterial fragmentation and overall defect morphology. Nevertheless, fine-scale analysis of the bone-biomaterial interface can be more effectively explored using FIB-SEM, which enables detailed visualization of microscale features and improves the identification of mesoscopic bone organization at the interface. Another important aspect of the correlative strategy relies on the combination of polarized light microscopy and TEM. Whereas polarized light microscopy reveals the organization of collagen fibers, allowing assessment of the organic component and the identification of preferential orientations associated with different stages of bone maturation, TEM provides nanoscale information on the arrangement, size, and alignment of mineral nanoparticles. Through this combination, it becomes possible to directly link tissue-level structural patterns to underlying mineral organization. This integrated perspective highlights how hierarchical features of the bone/biomaterials interface emerge across spatial scales and demonstrates the unique explanatory power of combining optical, microstructural, and ultrastructural imaging modalities within a unified correlative framework.
In conclusion, the imaging techniques outlined here facilitate a comprehensive evaluation of the hierarchical structures of bone. By integrating these methods, we can assess the capacity of biomaterials to induce the formation of organized mineralized tissue that restores both the structure and function of original bone. As bone tissue engineering advances, a workflow like the one presented in this protocol will serve as a robust method for validating biomaterials.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
This work was supported by grants to the authors from the following Brazilian agencies: Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq/Brazil) and Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ). We thank LABNANO/CBPF and CENABIO/UFRJ for the electron microscopy facilities. Language refinement and grammar revision were performed using ChatGPT (OpenAI, GPT-5.1, 2025). All edits were reviewed and verified by the authors.
| Acetone | Merck | 67-64-1 | |
| Aluminium cylinder | home-made | n/a | |
| Avizo version 2022.1 | ThermoFisher | n/a | software for image filtering, segmentation and 3D visualization |
| Cacodylic Acid, Sodium Salt, trihydrate | Ted Pella | 18851 | powder for Cacodylate Buffer. Prepare a stock solution of 0.2 M |
| Diamond Polishing Compound 3 µm size | Ted Pella | 895-8 | |
| Diamond Polishing Compound 6 µm size | Ted Pella | 895-9 | |
| Diamond Wheel x.006 | South Bay Technology | DWH 3063 | |
| Diamond Wheel x.012 | South Bay Technology | DWH 4122 | |
| Felt polishing cloth | Ted Pella | 816-20 | |
| Flat Embedding Capsules | EMS | 70021 | |
| Glass cover slip | Knittel | n/a | |
| Glass slide | Knittel | n/a | |
| Glutaraldehyde, 25% EM grade | Ted Pella | 18426/18427 | |
| Gold sputter K550X | EMITECH | n/a | |
| ImageJ | National Institutes of Health (NIH) | software for image aligment, filtering, segmentation and 3D visualization | |
| IMOD version 5.1 | University of Colorado | software for image aligment, segmentation and 3D visualization | |
| Lapping e Polishing Machine | South Bay Technology | model 910 | |
| Low-Speed Diamond Wheel Saw | South Bay Technology | model 650 | |
| Low Viscosity Embedding Kit | EMS | 14300 | kit for SPURR resin. Prepare according to the Data sheet specifications |
| Micro-CT Skyscan 1273 | Bruker | n/a | |
| Microscope Jeol 2100F 200kV | Jeol | n/a | |
| Microscope Olympus BX51 | Olympus | n/a | |
| Microscope Tescan Lyra3 | Tescan | n/a | |
| Microscope Zeiss Axioplan | Zeiss | n/a | equipped with a first order lambda plate |
| Paraformaldehyde, EM Grade | Sigma | P6148 | powder for formaldehyde solution. Prepare a stock solution of 16% |
| Plastic slide | Exakt | 41500 | |
| QuickStick 135 Mounting Wax | South Bay Technology | n/a | |
| Sandpapers grit sizes 1200 | n/a | n/a | |
| Sandpapers grit sizes 2500 | n/a | n/a | |
| Sandpapers grit sizes 4100 | n/a | n/a | |
| Super Glue | Loctite | n/a | |
| Water Soluble Coolant | South Bay Technology | n/a | |
| Whatmann grade 1 filter paper | sigma | WHA1001150 |
