Visualization of osteocyte lacunae is relevant for both fundamental and applied bone research because morphological parameters may serve as biomarkers to describe pathologies and monitor disease development [3]. Since all visualization methods rely on physical properties of the material, which depend on its composition, many methods will not allow a direct assessment of the osteocyte but rather an indirect assessment of their lacunae, as an enclosure in the surrounding bone matrix, which is why this manuscript focusses primarily on methods for visualization of lacunae. The visualization of osteocyte lacunae provides several challenges regardless of the method used. Firstly, the target of investigation is small (ellipsoid with approximate diameters of 19 × 9 × 5 μm [10]), thus the available resolution of the analysis method is of utmost importance. Secondly, the matrix surrounding osteocytes and their lacunae is calcified, requiring specialized processing methods with a potential of introducing artefacts [6]. Some techniques allow a direct assessment of the osteocyte and their lacunae, but just in demineralized bone or only in 2D. Consequently, the exact research question influences the method of choice. Thirdly, depending on the availability of samples, methodical aspects like sample destruction during measurement should be considered when choosing a method.

In the current literature several in-depth reviews of general 3D lacunae analysis are available [4, 6, 11, 12], while some authors have also reviewed specific methods [scanning electron microscopy (SEM): [7]; atomic force microscopy (AFM): [13]; synchrotron-based methods: [14, 15]; CLSM: [16]]. In the present manuscript, we primarily focus on methods for lacunae visualization. For a brief comparison, we will discuss their advantages and limitations in respect to the following aspects: ability to visualize in 2D or 3D, size of region or volume of interest (ROI/VOI), necessary/common preparation, current minimal resolution, applicability for canaliculi visualization, direct assessment of osteocyte or indirect via lacunae, method-inherent pros and cons, destructiveness of the method regarding the sample, and common artifacts. The following text will highlight the most relevant (dis)advantages. For an overview with all characteristics see Tables 1 and 2 for 2D and 3D methods, respectively.

Visualization in 2D and Thin 3D Sections

While 3D methods are most useful for the visualization of lacunae, some commonly used 2D methods will be discussed here for the following reasons: (1) 2D methods have been and are still commonly used. (2) 2D methods can provide functional aspects through various staining techniques, e.g. labeling specific osteocyte metabolites or cell constituents. Thus, understanding the advantages and weaknesses of 2D methods compared to 3D methods for quantitative analysis of lacunar morphology is necessary to compare existing and future literature containing 2D and 3D results. Here, we will briefly discuss the pros and cons of conventional light microscopy, confocal laser scanning microscopy, scanning electron microscopy (+ acid etching), quantitative backscattered electron imaging (qBEI), transmission electron microscopy (TEM), and atomic force microscopy.

While for 2D methods, variations in lacunar morphometry along the axis perpendicular to the sample surface remain hidden, there may still be valid reasons for choosing specific 2D or semi-3D methods, depending on the research question. Reasons to choose these methods might be the wish to quantify the composition of bone matrix around lacunae without access to synchrotron imaging [17] or to correlate lacunar size with cell biological aspects. Why 2D-based estimations of lacunar morphology are problematic will be discussed using the example of lacunar volume in the Sect. Systematic Error with 2D-Based Estimates for 3D Objects.

Advantages of 2D Electron Microscopy Techniques (Acid etching-SEM, TEM, qBEI/BEI) and AFM

Electron microscopy offers different modes for lacuna and canaliculi visualization. Using secondary electron imaging after acid etching the surface of a PMMA-embedded bone sample allows to visualize the topography of the lacunocanalicular network in a bird’s eye view [18, 19] with a resolution in the nanometer range [20] and a comparatively simple preparation [19] (Fig. 1A). Alternatively, the organic component can be dissolved, resulting in voids where osteocytes and their canaliculi resided [21].

Further, transmission electron microscopy (= TEM) provides imaging of ultra-thin sections of bone with or without demineralization and optional staining [22, 23] at high-resolutions (0.2 nm) [11, 24]. Micro- and nanostructural features [25] as well as cellular and subcellular structures such as organelles or calcified nanospherites in micropetrosis [26] can be imaged and information about mineralization and crystallization can be retrieved [23, 27]. Depending on the detection mode, TEM can be used to visualize hard tissues (scattered electrons), soft tissues (phase contrast) [11] or more complex aspects, e.g. sample composition (high-angle annular dark field) or mineral crystallization (electron diffraction) [28].

Backscattered electron imaging (BEI) creates images with gray values based on atomic numbers and allows to analyze the mineralized matrix surrounding the lacunae and mineralized lacunae (micropetrosis) if present. The mineral dependent gray values not only provide high contrast between lacunae and the surrounding bone matrix but with additional calibration the mineral content can be quantified (quantitative backscattered electron imaging = qBEI) [29] and thus provide insight into mineralization processes (Fig. 1B).

Atomic force microscopy (AFM) theoretically offers up to atomic scale resolution [30] and can be used to create maps of lacunae recording their topography, mineral crystal size [31], and mechanical properties at the nanoscale [13, 32]. In the context of lacunae investigation, this method is predominantly used to investigate the bone surrounding lacunae and its mineralization [33,34,35] or to investigate objects within the lacunae, e.g. mineralized nanospherites [26], but it can also be employed to test the cytoskeleton of osteocytes [36]. Zhou and Du [30] provide a detailed overview of applications of AFM for bone research.

Fig. 1

Exemplary images of lacunae and canaliculi visualized in 2D using (A) SEM on acid-etched bone, (B) qBEI (C) brightfield microscopy of a Ploton silver stained sample and D) confocal laser-scanning microscopy of a rhodamine stained sample

Limitations of 2D Electron Microscopy Techniques (Acid etching-SEM, TEM, qBEI/BEI) and AFM

All scanning electron microscopy-based techniques may be affected by artifacts stemming from crack development in vacuum, sample charging [37], and distortion effects (spatial distortion, drift distortion and scan line shifts) [38]. Correcting the latter is especially relevant if sample are mechanically tested and the deformation is evaluated using digital image correlation (DIC) [38]. Boyde [7] describes in detail best practices for the preparation of bone samples for SEM imaging.

Despite its straightforward sample preparation, SEM with acid etching has inherent limitations. The method is destructive since bone matrix is chemically removed during acid etching [19]. Consequently, the created image will display a mold of the lacunocanalicular network and the concentration of the acid and treatment duration influence the heterogeneity of the etching result [19]. Therefore, the accuracy of the network representation depends on the quality of the resin infiltration process, which may be affected by resin shrinking [11]. Due to the indirect imaging approach, samples with pathologies affected by osteocyte apoptosis or disruption of canaliculi might still show an intact lacunocanalicular network [19]. Further, the images represent a 2D projection of a 3D topography, which renders quantification of specific parameters difficult.

While TEM can provide very detailed insight into 2D properties of osteocytes and their lacunae, it requires complex sample preparation involving either FIB or ultra-thin sectioning, e.g. cryo-sectioning, as embedded bone samples have to be processed to reach a thickness of less than 100 nm [11]. Additionally, visualization of e.g. cell constituents may require optional staining [7].

Both variations of backscattered electron imaging (qBEI, BEI) can by themselves only produce 2D images. Fortunately, it is possible to combine them with focused ion beam milling or microtome cutting in the microscope chamber, allowing 3D visualizations with quantifiable mineral (calcium) content (cf. Sect. Advantages of Volume Electron Microscopy).

Due to its high resolution, the region of interest that can be measured in a realistic time frame is very limited for AFM. While AFM can be used on fresh, wet samples [13], often the surface of bone samples is processed through grinding or polishing, to ensure that height fluctuations are minimal. This might influence the recorded mechanical parameters [39]. Further, tip convolution can be a common source of artifacts [13] resulting in image distortions.

Advantages of Light, Confocal, Light Sheet, and Super-Resolution Light Microscopy

While conventional light microscopy offers 2D insights into lacunae morphology, confocal laser scanning microscopy (CLSM) and fluorescent light sheet microscopy allow 3D visualization of lacunae in thin samples but are limited by the penetration depth of bone to varying degrees.

All methods share the following advantages: (1) soft tissue contrast allowing the visualization of the osteocyte itself, (2) the possibility of mounting wet samples, (3) the option for functional imaging of subcellular and surrounding structures through histological staining, immunohistochemistry or fluorescent labeling.

Light microscopy with conventional staining methods has been used for bone histomorphometry for decades with plenty of detailed descriptions [1, 40] (Example of Ploton silver staining in Fig. 1C). It is limited by the diffraction limit of light (200 nm) [41]. Here, we focus on the advantages of confocal and light sheet microscopy in combination with fluorescent labeling. Although developed with other light sources [42], currently confocal microscopy most often makes use of lasers with specific wavelengths. These specific wavelengths can create auto-fluorescence in biological samples. As confocal microscopy is based on the detection of a point light source through a pinhole, which eliminates light outside of the focal plane [20, 43], blurring of the fluorescent signal is reduced in confocal light microscopy compared to conventional light microscopy [11].

While bone already possesses autofluorescent qualities [6], by combining specific wavelengths through filters or lasers with fluorescent dyes [44] it is possible to visualize osteocytes directly or indirectly and mark sub-cellular structures or other targets. Some fluorescent dyes stain the interstitial fluid and thereby allow indirect assessment of the osteocyte, common examples are: fluorescein isothiocyanate isomer I [45, 46], rhodamine based chromophore (Fig. 1D) [47], basic fuchsin [48]. Other fluorescent dyes or conjugated fluorescent dyes can directly stain the cell, its parts, or the surrounding matrix, e.g. dyes conjugated with phalloidin label the actin skeleton [49] whereas DAPI labels the nuclear chromatin/DNA [44]. A comprehensive list of fluorescent dyes, their applications, and the corresponding laser wavelengths can be found in Canette and Briandet [44].

In addition to osteocyte or lacuna visualization, CLSM can provide information about bone cell biology or bone growth when using tetracycline labelling [16, 50]. Some researchers have developed transgenic mouse models, which selectively express a membrane targeted-GFP (green fluorescent protein) variant in osteocytes, allowing optimized imaging of osteocytes without additional staining [51]. For more information on high-resolution fluorescence microscopy the reader may refer to Schermelleh, et al. [52].

Light sheet microscopy has become more popular in the last 10 years due to its ability to image large volumes in a short time with a high photon efficiency [4, 53]. Several studies highlight its potential for bone research by providing image volumes ranging from zebrafish craniofacial bones [53] to cleared whole porcine cochleae [54]. While the resolution of light sheet microscopy is worse compared to CLSM, light sheet microscopy allows sample sizes of up to 1 cm³ with clearing [4].

While super-resolution light microscopy methods like structured illumination microscopy (SIM), stochastic optical reconstruction microscopy (STORM), and total internal reflection fluorescence microscopy (TIRF) have become available and can supersede the resolution limit of conventional light microscopy, they have rarely been used to image osteocytes. SIM was used to image the lacunocanalicular network [55] and osteocyte processes in murine bone with a resolution of 90 nm [56].

Limitations of Light, Confocal Laser Scanning and Light Sheet Microscopy

Conventional light microscopy is limited to 2D imaging. For conventional light microscopy and CLSM, the histological sample preparation can introduce errors during cutting or microtome sectioning [20]. While both confocal light and confocal laser scanning microscopy are theoretically only restricted by the optical resolution limit and allow 3D visualization of thin samples, the depth of the volume of interest is limited due to the low photon penetrability of mineralized bone [20]. Depth-dependent artifacts arising from signal distortion and attenuation effects limit the attainable depth of the volume of interest in CLSM [6, 20]. While for mineralized bone samples the penetration depth of CLSM amounts to 10s of microns [57], for demineralized or cleared samples the penetration depth is much higher and can theoretically reach up to 150 μm [8]. Heveran et al. [5] recommend to image at least a depth of 60 μm to reduce sectioning effects of partially cut lacunae and to achieve stable values for 3D measures of osteocyte lacunar geometry. During evaluation, an inhomogeneity in lacunar contrast and bright bone marrow fluorescence can influence osteocyte (lacuna) morphology measurements specifically during thresholding [5].

Applications of light sheet microscopy for bone research have been limited by the need for translucency of the sample conflicting with the high density of bone. This issue can be overcome by tissue clearing methods optimized for bone allowing fluorescent 3D visualization of whole murine bones without sectioning using light sheet microscopy [54,