A Stereological Study of Mouse Ovary Tissues for 3D Bioprinting Application

Abstract

Introduction

The use of 3D-bioprinted ovaries has been proven to be a promising technique for preserving fertility. Stereology is an accurate method to obtain quantitative 3D information and the stereological data is the basis for 3D bioprinting ovaries.

Methods

In this study, six female mice were used to acquire the ovarian tissues. One of the two paraffin-embedded ovaries of each mouse was cut into 5 µm sections, and the other was cut into 15 µm sections and then subjected to haematoxylin and eosin staining and anti-follicle stimulating hormone receptor antibody immunohistochemistry. The volume and volume fractions of ovaries were measured by the Cavalieri method. Then, the numerical densities and total numbers of ovarian granulosa cells (OGCs) and primordial, preantral and antral follicles in serial sections were estimated using design-based stereology.

Results

The ovarian volume was 2.50 ± 0.32 mm³. The volume fractions of the cortex, medulla, follicles and OGCs were 86.80% ± 2.82, 13.20% ± 2.82%, 5.60% ± 0.25% and 81.19% ± 2.57%, respectively. The numerical densities of OGCs, the primordial, preantral and antral follicles were 2.11 (± 0.28) × 10⁶/mm³, 719.57 ± 18.04/mm³, 71.84 ± 3.93/mm³ and 17.29 ± 3.54/mm³, respectively. The total number of OGCs and follicles per paraffin-embedded ovary were 5.26 (± 0.09) × 10⁶ and 2013.66 ± 8.16.

Conclusions

The study had obtained the stereological data of the mice ovaries, which contribute to a deeper understanding of the structure of the ovaries. Meanwhile, the data will supply information for 3D bioprinting ovaries.

Introduction

Increasing numbers of childbearing age women are now surviving malignant tumors, and fertility concern is paramount for them.21 While cryopreservation and transplantation of ovarian tissue is considered an increasingly procedure to restore fertility, it is not an absolutely safe method for some types of cancer cells with the risk of reimplanting together with the frozen-thawed tissues.3 With the development of biomedical technologies, three-dimensional (3D)-bioprinting has been applied for tissue engineering and has been used to achieve important milestones.5,9,15 In recent years, 3D-bioprinting of natural or synthetic polymers to rebuild an artificial ovary has been proven to be a promising technique for preserving fertility.4,11,19 However, the cells in these studies were not printed together with the hydrogel, and there is no report about 3D-bioprinted ovary employing ovarian decellularized extracellular matrix (dECM) bioink encapsulating ovarian cells.

We have now successfully developed a dECM bioink encapsulating ovarian cells that can be used for 3D-bioprinting (Data not yet published). However, 3D-bioprinting involves more than just a simple stack of cells and bioink, and it also requires the construction of a printing platform and the design of a printing scheme. Therefore, a full understanding of the structure of tissue, the distribution of major cells and the numerical density of cells is of great significance for 3D reconstruction functional ovary.

Although two-dimensional (2D) quantitative data can be obtained by morphometric approaches, the use of design-based stereological techniques is preferable, as stereology is an accurate method to obtain quantitative 3D information about microstructure based on the comprehensive observation of 2D slices.7 It uses uniform and random sampling of organs, slices, field of vision and spatial direction rather than randomly selection of samples, and each part of the study area has the same probability of being selected. In this study, we aimed to acquire the normal mouse ovaries’ stereological data to gain the accurate distribution and the numerical density of ovarian granulosa cells (OGCs) and follicles, which will supply information for 3D-bioprinting ovaries.

Materials and Methods

Animals

All procedures that involved animals were approved by the local Institutional Animal Ethics Committee (approval letter no. 2019-P056). Wild-type female C57BL/6 J mice (n = 6) (SPF (Beijing) Biotechnology Co., Ltd.) aged 10 weeks (22 g) were kept on a 12 h light and 12 h dark cycle and provided free access to mouse food and water.

Tissue Collection and Processing

Daily vaginal smears were taken to ascertain oestrus stage. Mice were anaesthetized, and all ovaries were collected in dioestrus. The ovaries were fixed in Bouin’s fluid, dehydrated using graded alcohol solutions, and embedded in paraffin wax. One of the two paraffin-embedded ovaries of each mouse was cut into 5 µm sections, and the other was cut into 15 µm sections and then stained using haematoxylin and eosin (HE) for the following stereology analysis. Meanwhile, the 5 µm sections were subjected to follicle stimulating hormone receptor (FSHR, a specific marker of OGCs) immunohistochemistry staining (1:200; Servicebio; cat. no. GB11275-1) to measure the diameters of mouse OGCs, primordial, preantral and antral follicles.

Estimation of the Ovarian Volume and the Volume Fractions of the Ovarian Cortex, Medulla, Follicles and OGCs

The volume of the ovary was determined by the Cavalieri method (Fig. 1): After staining with HE, every 8th 5 µm section was chosen from a random starting point (sampling fraction (f) = 1/8), and 16 sections per ovary were selected and examined using a microscope (BX63; Olympus) at 10× magnification. The ovarian volume was estimated by the point counting method (the intersection of the crosshairs is the counting point) as follows:

$${\text{V}}_{\text{ovary}}=\Sigma P_{\text{total}}\times a\left(p\right)\times d\times \left(1/f\right),$$

where “ΣP_total” is the total number of points superimposed on the sections; “a(p)” is the area per point, “d” is the distance between the sampled sections, and “(f)” is the sampling fraction. The volume of each ovary fraction was estimated as follows:

$$V_{\text{vmedulla}}=\Sigma P_{\text{medulla}}/\Sigma P_{\text{total}},$$

where “ΣP_total” is the total number of counted points and “ΣP_medulla”, “ΣP_cortex” and “ΣP_follicle” indicate the total number of points superimposed on the medulla, cortex and follicles, respectively. “ΣP_follicle” is subtracted from “ΣP_cortex” to obtain “ΣPOGC”. The volume of the ovarian cortex, medulla, follicles and OGCs were estimated by multiplying “V_ovary” by the corresponding volume fraction.

Figure 1.

The Cavalieri method was used to calculate the volumes of the mouse ovary, cortex and medulla. Sixteen 5 µm-thick sections per ovary were randomly selected (sampling fraction (f) = 1/8), and using the point counting method and a microscope with a 10 × objective. The intersection of the crosshairs is the counting point (arrow) and the number of points superimposed on the images was counted. “H” is the height of the ovary; “d” is the distance between the sampled sections; “A” is the sampled section; and “a” is the side length of the point.

Estimation of the Numerical Density of OGCs: Optical Disector

The disector consists of two parallel sections (the reference section and the look-up section). An unbiased counting frame was placed in the reference section, and the particles were counted if they appeared in the reference section but not in the look-up section. The number of particles measured in the reference space per unit volume of one tissue was the numerical density (N_v). Then, the total particle number was calculated by multiplying N_v by the total volume of the reference section.20 In a systematic random manner, 10 15 µm-thick sections per ovary were chosen. OGCs were counted using a 100 × oil immersion objective (N_A = 0.5) on an Olympus BX63 microscope and a microphone (DP73; Olympus) connected to a computer. The microscope measured the movement of the stage in the z-axis. By moving the microscope stage at identical distances, microscopic fields were selected. To avoid surface cutting irregularities that may have biased the count, the 3 µm area at the top of each section was used as a guard area, and the next 5 µm of the 15 µm-thick section was optically sectioned. Subsequently, two counting frames (365.26 µm² per frame) were superimposed over the reference section, and OGCs were counted if they appeared in the unbiased counting frame applied to the reference section, were not intersected by exclusion lines and did not appear in the look-up section (Fig. 2). The N_v of OGCs was calculated using the following formulas:

$$N_{\text{v}}=\Sigma Q/\Sigma V\left(\text{dis}\right),\Sigma V\left(\text{dis}\right)=\Sigma P\times a^2\times h,$$

where “ΣQ” is the whole number of OGCs selected in the disector height, “ΣV(dis)” is the total volume of the disector, “ΣP” denotes the total number of counted frames in all microscopic fields, “a²” indicates the area of each frame and “h” is the height of the disector.

Figure 2.

Use of an optical disector to estimate the numerical density of ovarian granulosa cells per ovary. The 3 µm area at the top of the sections was used as a guard area, and the next 5 µm of the 15 µm section thickness was optically sectioned. The frame itself contains inclusion (green) and exclusion (red) boundaries. Granulosa cells were counted if they appeared within the unbiased counting frame, were not intersected by exclusion boundaries and did not appear in the look-up section. For example, particles No. 1–6 in the reference were included as they satisfy the counting criteria, and No. 3 was included, as it touches only the inclusion boundary. Cells No. 7–8 were not included, as they touch the exclusion line. Bar = 5 µm.

Estimation of the Numerical Density of Primordial, Preantral and Antral Follicles: Physical Disector

According to the shape, size, growth rate and histological characteristics of follicles, the growth process can be divided into four stages: the primordial, preantral, antral and preovulatory stages. According to the demand of 3D printing, we performed a quantitative study of the follicles in the first three stages. After staining with HE, every 7th 5 µm-thick section was chosen from a random starting point (Fig. 3) and examined using a 40 × oil immersion objective (N_A = 0.5) on an Olympus BX63 microscope. Sample tissue sections were traced to define tissue boundaries, and then separate unbiased counting frames (measuring 6400 µm² for primordial follicles, 16,900 µm² for preantral follicles and 22,500 µm² for antral follicles) superimposed on the reference sections were used to count primordial, preantral and antral follicles. CellSens Dimension software (v1.18; Olympus) was used to match fields between the reference section and the consecutive look-up section. Oocyte nuclei were counted in the same manner as OGCs. Approximately 100 oocytes per ovary were analysed, and the N_v was obtained using the above formulas.

Figure 3.

Sampling of ovarian tissue to estimate the numerical densities of follicles per ovary using a physical disector. A pair of sections was analysed with a physical disector. Every 8th and 9th section were chosen, and both sections were aligned in a single focal plane. Follicles were counted if they appeared in the reference section but not in the look-up section.

Presentation of the Data

All data are continuous variables and are presented as the mean ± SEM.

Results

The Diameters of Mouse OGCs, Primordial, Preantral and Antral Follicles

As shown in Figs. 4a and 4b, the diameter of mouse OGCs was 4.98 ± 1.69 µm. The mean diameters of the follicles of the three stages were 29.70 ± 12.21, 52.92 ± 21.12 and 56.32 ± 17.05 µm (Figs. 4c to 4g).

Figure 4.

The diameters of ovarian granulosa cells and oocyte nuclei. a Ovarian granulosa cells. Bar = 5 µm. The average diameter was 4.98 ± 1.69 µm (b). Primordial follicles were defined as oocytes surrounded by a layer of squamous granulosa cells (c). Bar = 5 µm. Preantral follicles include primary follicles and secondary follicles. The former were surrounded by a single layer of cuboidal granulosa cells (d). Bar = 5 µm. The latter possessed more than one layer of cuboidal granulosa cells and no visible antrum (e). Bar = 50 µm. Antral follicles possessed a clearly defined antral space (f). Bar = 100 µm. The mean diameters of follicles of the three stages were 29.70 ± 12.21 µm, 52.92 ± 21.12 µm and 56.32 ± 17.05 µm (g).

The Ovarian Volume and Corresponding Volume Fraction

The ovarian volume of C57BL/6 J mice at 10 weeks of age was 2.50 ± 0.32 mm³. The volume fractions of the cortex, medulla, follicles and OGCs were 86.80% ± 2.82, 13.20% ± 2.82%, 5.60% ± 0.25% and 81.19% ± 2.57%, respectively.

The Numerical Density of OGCs

The N_v of OGCs in C57BL/6 J mice at 10 weeks of age was 2.11 (± 0.28) × 10⁶/mm³. The total number of OGCs per paraffin-embedded ovary was 5.26 (± 0.09) × 10⁶ (Table 1).

Table 1.

Statistical table of the number of nuclei.

	Ovarian granulosa cell	Primordial follicle	Preantral follicle	Antral follicle
Q	5581	87	23	5
P	1451	3780	1145	661
Nv (n/mm3)	(2.11 ± 0.28) × 106	719.57 ± 18.04	71.84 ± 3.93	17.29 ± 3.54
N(n)	(5.26 ± 0.09) × 106	1791.74 ± 5.77	178.88 ± 1.26	43.04 ± 1.13

“Q” is the total number of cells selected in the disector height. “P” denotes the total number of counted frames in all microscopic fields. The number of particles measured in the reference space in unit volume of ovary is the numerical density (N_v). Then, the total cell number (N) was calculated by multiplying N_v by the total volume of the reference section

The Numerical Densities of Primordial, Preantral and Antral Follicles

The N_vs of the primordial, preantral and antral follicles in 10-week-old C57BL/6 J mice were 719.57 ± 18.04/mm³, 71.84 ± 3.93/mm³ and 17.29 ± 3.54/mm³, respectively. The total number of follicles of the three stages per ovary was 1791.74 ± 5.77, 178.88 ± 1.26, and 43.04 ± 1.13, and the total number of follicles per paraffin-embedded ovary was 2013.66 ± 8.16 (Table 1).

Discussion

The stereological data is the basis for ovarian reconstruction by 3D printing. In this study, the volume and corresponding volume fractions of mouse ovaries were determined by using Cavalieri’s principle. Additionally, we used an optical disector and physical disector to determine the N_vs of OGCs and follicles.

Stereology has become an essential tool for various fields that require quantitative 3D microstructural information.8,10,16 Generally, the stereological apparatus used to count the number of particles in 3D space is called the disector. The disector can provide essentially unbiased 3D data because it gives information on the existence of particles but not on the size, shape or direction of the particles.22 The different focal planes of thick histological sections are called optical sections, and conventional sections are called physical sections. Disectors that can assess physical and optical sections and are called physical disectors and optical disectors, respectively. Large structures that can be observed by light microscopy (such as glomeruli and islets) and tiny structures that can only be visualized by electron microscopy (such as mitochondria and synapses) should be analysed with a physical disector; an optical disector is suitable for the counting of small particles, such as cell nuclei (less than twenty microns in diameter)23 and synaptic granules.17 In our study, the minimum diameter of mouse oocytes was more than 10 µm, and therefore, a physical disector was used. Additionally, OGCs (which had a maximum diameter of less than 10 µm) were analysed using an optical disector. The main parameter of stereology we chose was N_v (the number of particles in the reference space per unit volume of ovarian tissue), as it allowed us to calculate the number of cells that would be encapsulated in the 3D-bioprinted structure regardless of its shape. In addition, the volume of any structure, regardless of its shape, can be calculated by Cavalieri’s principle,6 and the total number of cells in the tissue can be determined by multiplying N_v by the volume of tissue.23 We also estimated the volume fractions of the cortex, medulla, follicles and OGCs by the dot counting method, and the corresponding printed volume was calculated by multiplying the volume of the artificial ovary by the corresponding volume fraction, which is conducive to the design of the 3D ovary-printing scheme.

On the other hand, rebuild an artificial ovary should imitate the natural ovary and requires appropriate primary ovarian cells for survival and development of the artificial ovary. Follicles are important functional structures of ovaries, and are composed of an oocyte and many small ovarian cells surrounding it. Ovarian granulosa cells (OGCs) are important primary ovarian cells that play an oestrogen-mediated regulatory role, support the development and maturation of follicles, and maintain hormonal balance in the ovarian niche.12–14 In addition, abnormal apoptosis of OGCs is an important mechanism underlying premature ovarian insufficiency (POI).2 The type and number of follicles in a delivery scaffold should be optimized. Grafted preantral follicles have a higher survival and growth rate than grafted primordial follicles.1 A delivery scaffold must include enough follicles to produce mature oocytes after transplantation but not too many follicles that the small size of the scaffold cannot be maintained. A small scaffold may be conducive to the prevention of ischaemic injury and neovascularization after transplantation.18 In view of this, we estimated the normal mouse ovaries’ number of follicles and OGCs, which can provide a basis for the ovarian reconstruction.

Furthermore, for different types of bioink and cells, 3D bioprinting construction schemes are different, and the purposes of biological printing are different. By changing the cell plating density, applying different types of cells and hydrogels, designing different cell space locations and utilizing other experimental conditions, we can perform in-depth studies on the communication between cells and the interaction between cells and the external environment. Researchers have found that impaired OGC function is one of the direct causes of POI,2 clinical pharmacologic hormone replacement therapy (pHRT) is the main treatment strategy for POI, but the mode of pHRT is controversial. 3D-bioprinted ovarian constructs that recapitulate native OGC-oocyte interactions, the generation of which requires accurate measurements of the volume fraction and N_v of OGCs, may be an alternative approach. Besides, 3D-bioprinted ovaries could provide further insights into the impact of mouse follicle stage on graft outcome in an artificial ovary environment, which requires an accurate number of follicles of different levels. That is, we determined the ovarian volume and the volume fractions, number and N_v of primordial, preantral, and antral follicles and OGCs in mouse ovaries by stereology to aid in the development of 3D-bioprinted ovaries.

However, some shortcomings in the study are inevitable. First, challenges to the application of stereology remain, although most of standard stereology techniques for measuring volume, volume fraction, surface area, curvature densities and so on are simple point counts upon correct sampling and do not require any expensive instruments. Stereology involves many steps, such as correct experimental design, slicing, sampling and software calculation. Deficiencies in any step will lead to different results between similar studies. Second, we used a physical disector to estimate the numerical density of follicles owing to the limitation of experimental conditions. Perhaps in some cases, the combination of physical disector and optical disector can be used to obtain more accurate results, but this speculation requires further exploration. Third, the study did not estimate the theca cells and blood vessels. Angiogenesis has always been a challenging problem for artificial ovaries, which will be the next step.

Conclusion

The study had obtained the stereological data of the mice ovaries, the data will supply information for 3D bioprinting ovaries.