Identification of Cellular Voids in the Human Otic Capsule

Abstract

The otic capsule consists of dense highly mineralized compact bone. Inner ear osteoprotegerin (OPG) effectively inhibits perilabyrinthine remodeling and otic capsular bone turnover is very low compared to other bone. Consequently, degenerative changes like dead osteocytes and microcracks accumulate around the inner ear. Osteocytes are connected via canaliculi and need a certain connectivity to sustain life. Consequently, stochastic osteocyte apoptosis may disrupt the osteocytic network in unsustainable patterns leading to widespread cell death. When studying bulk-stained undecalcified human temporal bone, large clusters of dead osteocytes have been observed. Such “cellular voids” may disrupt the perilabyrinthine OPG mediated remodeling inhibition possibly leading to local remodeling. In the common ear disease otosclerosis pathological bone remodeling foci are found exclusively in the otic capsule. We believe the pathogenesis of otosclerosis is linked to the unique bony dynamics of perilabyrinthine bone and cellular voids may represent a starting point for otosclerotic remodeling. This study aims to identify and characterize cellular voids of the human otic capsule. This would allow future cellular void quantification and comparison of void and otosclerotic distribution to further elucidate the yet unknown pathogenesis of otosclerosis.

BACKGROUND

Bone is a dynamic tissue with a significant turnover (Clarke 2008; Datta et al. 2008; Feng and McDonald 2011) although the static appearance of osteocytes enclosed in lacunae falsely suggests inactivity. Bone is continuously resorbed and rebuilt in the process of remodeling governed by intrinsic and extrinsic factors to maintain a healthy and functional skeleton (Parfitt 1994; Andersen et al. 2009; Florencio-Silva et al. 2015). Osteocytes are interconnected via dendritic processes through the lacuno-canalicular network. This connectivity is important for signal transduction, metabolism and to respond to strain and tissue damage (Vashishth et al. 2000). Cells that are cut off completely from the inter-cellular network, due to degenerative changes like microcracks, must die as they lose their nutritional supply (Tami et al. 2002; Knothe Tate 2003). Even in an otherwise patent network, if the density of living osteocytes falls below a certain threshold, the connectivity may focally become dysfunctional leading to bone degeneration and consequently bone remodeling (Noble et al. 1997; Vashishth et al. 2000; Florencio-Silva et al. 2015; Tiede-Lewis et al. 2017). The bone surrounding the inner ear, the bony otic capsule, is rarely remodeled (Sørensen et al. 1990, 1992; Sølvsten Sørensen et al. 1992). OPG is most likely responsible for the unique bony environment in the otic capsule (Zehnder et al. 2005, 2006; Nielsen et al. 2015). The absence of remodeling results in accumulation of microfractures (Frisch et al. 2001a, 2008, 2015; Hansen et al. 2020) and dead osteocytes (Bloch and Sørensen 2010; Bloch et al. 2012). When dead osteocytes are not replenished and microdamage is not repaired, islets of osteocytes suffer isolation and death due to disruption of the lacuno-canalicular network as evident in bulk-stained human materials. Computer simulations have demonstrated how such areas devoid of living cells, so-called cellular voids, may occasionally form around the inner ear (Bloch and Sørensen 2016).

Cellular voids in the otic capsule could be a missing link in the pathogenesis of bone remodeling disorders around the inner ear, such as otosclerosis. A cellular void represents a disconnection of the lacuno-canalicular network and, consequently, an area where the anti-resorptive OPG signal is prevented from entering. Cellular voids thereby offer a possible explanation of how the focal pathological remodeling in otosclerosis can occur where remodeling is otherwise so effectively inhibited and rare.

When studying osteocyte viability and microdamage at low magnification in undecalcified bulk stained human temporal bones, voids with no or very little stain are frequently encountered. These areas appear osteocyte depleted and are often sharply demarcated from viable tissue in fluorescent light. Their rounded shape breaks up the “patchy” appearance of the otic capsular bone, but at low magnification the actual density of vital osteocytes on either side of the observed void perimeter is not detectable. Based on known osteocyte viability criteria and stereological principles of quantification, this study presents and validates a method to identify cellular voids in undecalcified human temporal bones. Furthermore, it provides a description of void morphology and cellular composition.

MATERIALS AND METHODS

Subjects and Samples

A total of eleven human temporal bones were used from three females and eight males aged 54 to 90 years (mean: 73.6 years). The temporal bones were chosen randomly among older temporal bones from the temporal bone collection at the Otopathology Laboratory at Rigshospitalet in Copenhagen. All samples were undecalcified, bulk stained in basic fuchsin by immersion in 62% ethanol with 1% basic fuchsin and embedded in methyl methacrylate, as described elsewhere (Frisch et al. 2001b, a; Hansen et al. 2020). Samples were horizontally sectioned at a fixed increment of 0.8 mm on a KDG-95 saw microtome (MeProTech, Netherlands) and hand ground to a thickness of 80–120 um, cover slipped and studied without further staining.

Microscopy Setup

All observations were made on a Zeiss Axio Imager M1 microscope (Carl Zeiss, Oberkochen, Germany) equipped with the following objectives: Zeiss EC Plan-Neofluar 1 × 0.025 NA, Olympus UPlanFl 4 × 0.13 NA, Zeiss Plan-Neofluar 10 × 0.30 NA, Zeiss Plan-Neofluar 10 × 0.50 NA, Zeiss EC Plan-Neofluar 40 × 0.75 NA, Zeiss EC Plan-Neofluar 100 × 1.3 NA oil. An AxioCam HRc (Carl Zeiss, Oberkochen, Germany), a Prior H101 motorized stage (Prior, Cambridge, UK) and an AxioCam HRc (Carl Zeiss, Oberkochen, Germany) camera mounted. Observations were made in transmitted light and fluorescent light from a Zeiss HBO 100 mercury arc lamp (Carl Zeiss, Oberkochen, Germany) with a Zeiss filter set 09 (BP: 450–490, beam splitter: FT 510, emission: LP 515). Stereological work was performed in the VIS 2017.2(7.0) stereological software (Visiopharm, Hørsholm, Denmark).

Sampling

The identification of dead and living osteocytes was performed according to criteria first described by Frost (1960a, b) and later applied to temporal bones (Bloch and Sørensen 2010; Bloch et al. 2012). The presence of basic fuchsin stain in the osteocytes categorized the cell as living, whereas an absence of stain in an osteocyte or a calcified lacuna meant the osteocyte was dead (Fig. 4, right image). To identify areas of low cellular density, each temporal bone slide was examined in blue, fluorescent light (≈450 nm) at low magnification (× 4/0.13). Areas with no basic fuchsin stain in either osteocytes, lacunae or canaliculi were identified and subsequently studied at higher magnification (× 100 oil/1.3). The numerical density of living and dead osteocytes was determined using the optical dissector as described below. The quantification was performed in both the cellular voids and the areas immediately surrounding them, termed peri-voids.

Fig. 4.

A cellular void was located (left image (× 4), blue line) and an oversized perimeter was outlined (left image, red line). Optical dissector probes were randomly superimposed (left image, small squares). Reference photo at top showing location of void in relation to the cochlea. Dead and living osteocytes (right image (× 100, oil), D = dead, L = living) were counted in and around the cellular void using the optical dissector probe. Only the 4 living cells marked with * would be counted in this situation (Scale bars: left 500 μm, right 20 μm)

Stereological Tools

Cellular numerical density, N_V, was estimated using unbiased counting frames, often referred to as the optical dissector (Gundersen 1986). This stereological probe allows an unbiased estimation of number of objects, ∑Q⁻ in a known volume of probes ∑V_samp.

$$N_V\left(o,b,j,e,c,t,s\right)=\frac{\sum Q^{-}}{\sum V_{\mathit{samp}}}=\frac{\sum Q^{-}}{h\bullet \left(\frac{a}{p}\right)\bullet \sum P}$$

where h = height of the dissector probe, (a/p) = area of the dissector frame (0.000,334 mm²) and P = the number of frame corner points hitting the area of interest. The optical dissection refers to sweeping a 2D counting frame a constant z-distance through 3D space thus creating a “brick” of tissue consisting of three acceptance planes and five forbidden surfaces (Howard et al. 1985; Howard and Reed 2010). Only particles within the brick or touching the acceptance line are counted, and particles touching the forbidden lines are not (Gundersen 1986)(Fig. 4, right image). This is to avoid oversampling. Counting was performed using the so-called unbiased brick-counting rule due to the varying shape of osteocytes (Howard et al. 1985). The optical dissector probe is placed in the region of interest at a random starting point and moved with a fixed step-length in the X–Y plane to ensure an unbiased systematic uniform random sampling (SURS). Cells may be miscategorized as missing due to cutting and grinding artefacts or poor stain penetration. In order to avoid this, a z-axis distribution was performed on the samples (Dorph-Petersen et al. 2009), which showed a uniform stain penetration. Based on this, a guard height of five µm and a dissector height of 35 µm were used. All samples were plastic embedded without decalcification, which normally produces minimal tissue shrinkage (Hanstede and Gerrits 1983; Nielsen et al. 1995), and both volume estimates and number density estimates were performed after staining and embedding. Therefore, no adjustment for shrinkage was performed.

Stereological Precision

The precision of the stereological estimation, expressed as the coefficient of error (CE), is a function of the counting variance in each section of tissue, the “noise” effect, and the variance between sections, the systematic uniform random sampling (SURS) variation. These factors depend on tissue profile shape, probe setup, sectioning direction and number of sections. The CE is defined as:

$$CE\left(N_V\right)=\frac{\sqrt{{\mathit{Var}}_{\mathit{Noise}}\sum Q^{-}+{\mathit{Var}}_{\mathit{SURS}}\sum Q^{-}}}{\sum Q^{-}}$$

These calculations have been described in detail elsewhere (Bloch et al. 2012; Hansen et al. 2020).

Graphs and Figures

Graphs were created in SigmaPlot 14.0 (Systat Software, Inc., San Jose, California, USA). Images were acquired in VIS 2017.2(7.0) (Visiopharm, Hørsholm, Denmark).

RESULTS

In the outer parts of the temporal bones, where remodeling is not inhibited by inner ear OPG, most osteocyte depleted areas are narrow and elongated (Fig. 1a). They resemble sections of secondary osteons and rarely exceed 200 µm in diameter. Osteocyte depleted areas can also be found around vessels and nerves as elongated and thin structures (< 250 µm in width) (Fig. 1b). In the otic capsule, close to the inner ear, the bone has a “patchy” appearance (Fig. 1c) and areas with dead osteocytes are often twisted or star-shaped with tongues of viable osteocytes in between the dead. These are not regarded as cellular voids in this study. However, some osteocyte depleted areas are morphologically different from the narrow and starshaped areas. They are larger, round or ellipsoid with fewer living osteocyte spikes infiltrating the perimeter, forming a sharply demarcated border to the surrounding tissue (Fig. 1d). These were marked as true cellular voids. Seventeen voids were identified in 10 of the temporal bones. There were between one and seven voids per temporal bone, and they were all located close to the inner ear spaces. The temporal bone with most voids (seven) was from an 86-year-old male. The numerical density of dead osteocytes, N_V,Dead, and living osteocytes, N_V,Alive, were estimated applying the optical dissector to both the cellular voids and the surrounding peri-voids (Figs. 2 and 3) to document the cellular composition of voids compared to the surrounding tissue.

Fig. 1.

Low power fluorescence microscopy images of a region anterior to the cochlea showing long narrow strips of dead osteocytes (arrows and dotted line), b narrow bands of dead osteocytes (arrows and dotted line) around vessels and nerves, c anterior oval window region showing “patchy” osteocyte death, not yet considered voids (dotted lines), d cellular void (arrow and dotted line) in the posterior aspect of human temporal bone. Notice the morphological difference between the osteocyte depleted areas in a-c compared to the actual cellular void in d which is larger, more rounded, and sharply demarcated. (Bars = 600 μm)

Fig. 2.

Numerical density of dead (black bars) and living (grey bars) osteocytes in 17 cellular voids

Fig. 3.

Numerical density of dead (black bars) and living (grey bars) osteocytes in 13 peri-voids

The count demonstrated a higher numerical density of dead osteocytes, compared to living osteocytes, in the cellular voids. Only one cellular void had a N_V,alive of more than 20,000 cells/mm³. The mean density of living osteocytes in the remaining cellular voids was 6,155 cells/mm³, compared to an N_v,Dead of 63,507 cells/mm³. The opposite relationship between dead and living osteocytes was found in the peri-void areas where most cells were alive with an average N_V,Dead of 21,152 cells/mm³ and an average N_V,Alive of 49,150 cells/mm³. Based on a pilot study, the CE of the numerical density estimation was 0.12.

To test the precision of the identification method, a cellular void was identified and marked as a region of interest. The void measured 572 × 745 µm meaning that the distance from the center to the perimeter was between 286 to 372.5 µm. In an area larger than the cellular void, optical dissector counting frames were superimposed in SURS positions (Fig. 4). Living and dead cells were counted in a blinded fashion so that the frame position was unknown to the observer. The number of living and dead osteocytes in 50 µm concentric zones from the center of the cellular void are shown in Fig. 5. From zero to 250 µm from the center of the cellular void, no living osteocytes were identified.

Fig. 5.

Numerical density of dead and living osteocytes with increasing distance from the center of a cellular void. No living osteocytes were found in the zones 0–250 μm corresponding with the size of the void. Trend lines: N_V,Dead (red), N_V,Alive (green)

DISCUSSION

Systematic studies of cellular viability and bony degeneration around the inner ear has led to the discovery of osteocyte depleted areas in the bony otic capsule, so-called cellular voids. This study set out to develop a method for identifying cellular voids in human temporal bones and to describe their morphology and cellular composition.

Specimens and Sampling

Empty and calcified lacunae (micropetrosis) are histological indications of osteocyte death (Frost 1960a, b; Frost 1960a, b). However, empty lacunae may also represent osteocytes that were lost during tissue preparation, although this is mostly related to thinner sections (≈7 μm) (Jilka et al. 2013). In this study only thick sections of undecalcified basic fuchsin bulk-stained temporal bone samples were used. Due to the surface guard heights of the optical dissector, only cells or empty lacunae fully embedded in the tissue were counted. This eliminated the risk of missing any osteocytes lost during tissue preparation. Other methods of identifying dead osteocytes exist but were not used in this study (Noble et al. 1997; Riahi and Noble 2012; Tiede-Lewis et al. 2017).

Viable osteocytes continuously maintain the patency of their lacunae and canalicular network (Holmbeck et al. 2005; Inoue et al. 2006) and thereby their morphology and bulk stainability. Basic fuchsin traces this porous network and therefore lacunae and canaliculi that are stained represent open canaliculi and living cells (Frost 1960a, b). After death, the osteocyte is gradually dispersed and the lacuna and canaliculi eventually calcify (Frost 1960a, b; Jilka et al. 2013). For this reason, dead osteocytes generally show no basic fuchsin stain. Lacunae may remain accessible and contain stainable cellular debris for a period of time after death, which could lead to an underestimation of dead osteocyte numbers (Frost 1960a, b). However, such lacunae will often show other signs of non-viability, like more rounded lacunae (Noble et al. 1997; Bloch et al. 2012) and missing canaliculi, and generally it is possible to distinguish between living and dead osteocytes. Furthermore, this method has previously been used in temporal bone histology to quantify living and dead osteocytes (Bloch and Sørensen 2010; Bloch et al. 2012). Therefore, the identification of dead osteocytes was based on empty lacunae lacking basic fuchsin staining.

The undecalcified bulk-stained temporal bone specimens were cut on a saw microtome and hand ground to their final thickness. This technique may result in a variation of stain intensity and section thickness. For the identification of cellular voids, the sections cannot be too thin or under stained. Otherwise, the border between void and living bone is blurred. The bulk staining procedure ensures very good penetration and a z-axis distribution confirmed this. The thick sections (80–120 μm) are suitable for unbiased stereological quantifications, which is an essential tool for counting cells in a three-dimensional space.

Stereological estimations require different degrees of randomness in the orientation and sectioning of specimens to be unbiased. Temporal bones are traditionally cut horizontally to facilitate recognition of the complicated anatomical structures of the inner ear. Unbiased stereology has previously been applied to temporal bone histomorphometry when estimating number and surface (Bloch et al. 2012; Bloch and Sørensen 2014; Hansen et al. 2020). In surface estimations, no difference was found between the horizontal and the randomly oriented sections, leading to the conclusion that the rounded structure of the inner ear is globally anisotropic to some degree (Hansen et al. 2020). However, for number estimations any orientation of the specimen may be used.

Fluorescent light has been applied to bone histomorphometry in previous studies (Prentice 1967; Vashishth et al. 2000; Frisch et al. 2015). Using blue fluorescent light has previously facilitated the identification of microfractures (Frisch et al. 2001b; Hansen et al. 2020) in basic fuchsin stained temporal bones, and was also chosen for this study. It makes the red/purple basic fuchsin stain visible against the bright green bony matrix background. In this case, the absence of a basic fuchsin stain was indicative of cellular voids.

Cellular Void Characteristics

At low magnification truly osteocyte depleted areas are readily identifiable in fluorescent light as bright green islands in a darker sea spotted red by basic fuchsin stained osteocytes. Cellular voids contain almost exclusively dead osteocytes, and it is the absence of stained canaliculi and living osteocytes that facilitate the identification of cellular voids. In viable bone, stained canaliculi are responsible for a darker appearance of the green bone matrix at low magnification.. True voids are large, rounded and sharply demarcated.

In the otic capsule, primary osteons, skein bone and globuli ossei persist in a fetal-type environment of bony components that are more tortuous and disordered (Mendoza and Rius 1966; Sørensen 1994). The combination of this primitive bony architecture and stochastic osteocyte death gives the aging otic capsule a “patchy” appearance. Osteocyte depleted areas close to the inner ear are therefore often irregular. They have intertwining septae of living cells and can appear starshaped. It is plausible that cellular voids emerge from these areas when the septae of living osteocytes die because the connectivity falls below a certain threshold. This could give rise to the rounded, sharply demarcated osteocyte depleted areas recorded in this study. In the anterior oval window region fissula ante fenestram (FAF) is regularly found. This small cleft in the temporal bone contains connective tissue. FAF does not stain in basic fuchsin and may resemble a cellular void but is not regarded as such in this study. However, FAF may functionally be a void.

Composition of Cellular Voids

Based on the observations and unbiased quantifications presented in this study a description of cellular voids has been provided. These recommendations apply to basic fuchsin bulk-stained temporal bones observed in blue, fluorescent light (Table 1).

Table 1.

Cellular void characteristics in basic fuchsin bulk-stained human temporal bones

Cellular voids are characterized by:
- a sharply demarcated border to the surrounding bony tissue
- a light background
- presence of calcified lacunae and dead osteocytes
- an absence of stained living osteocytes and canaliculi
- a round or ellipsoid shape—longer tongues as a periphery of hollow structure (larger vessel or nerve) or star shaped dead areas, are not counted as cellular voids

Seventeen cellular voids were identified in this study all displaying the characteristics described in Table 1.The density of dead osteocytes was much higher in the voids compared to peri voids. One void had a N_V,Alive of more than 20,000 cells/mm³ but the average N_V,Alive of the remaining 16 voids was 6,155 cells/m³. The average N_v,Alive in the peri-voids was 49,150 cells/mm³, which is eight times higher than in cellular voids. The total cell counts vary from specimen to specimen, probably due the different ages of the temporal bones.

In order to test the accuracy of the cellular void identification method, a void was marked and measured at low magnification, and numerical densities of living and dead osteocytes were quantified at high magnification in a blinded fashion across an oversized area containing the void. The high magnification count showed that the shift between living and dead areas correlated with the border traced at low magnification. This demonstrated the suitability of low magnification scanning of human bulk-stained temporal bones in fluorescent light for accurate identification and outlining of cellular voids.

Although the present sample size is too small to demonstrate a positive correlation between age and void size and number, it is likely that cellular voids will accumulate in the otic capsule with time, as previously documented for dead osteocytes and microcracks.

CONCLUSION

This paper aimed to describe and characterize cellular voids and validate the microscopical identification of them in basic fuchsin bulk-stained human temporal bones. Using well known methods of temporal bone histology, histomorphometry, unbiased stereology and osteocyte viability criteria, it was possible to distinguish the morphology and composition of cellular voids from normal capsular bone. The accuracy and feasibility of the light microscopy identification method described was tested. This method is a new tool dedicated to measure and map the amount and distribution of osteocytic cellular voids in human perilabyrinthine bone and to study the possible role of voids in the pathogenesis of otosclerosis.

Declarations

Conflict of Interest

The authors declare that they have no conflict of interest.