Spatio-temporal remodelling of the composition and architecture of the human ovarian cortical extracellular matrix during in vitro culture

Abstract

Study question: How does in vitro culture alter the human ovarian cortical extracellular matrix (ECM) network structure?

Summary answer: The ECM composition and architecture vary in the different layers of the ovarian cortex and are remodelled during in vitro culture.

What is known already: The ovarian ECM is the scaffold within which follicles and stromal cells are organized. Its composition and structural properties constantly evolve to accommodate follicle development and expansion. Tissue preparation for culture of primordial follicles within the native ECM involves mechanical loosening; this induces undefined modifications in the ECM network and alters cell-cell contact, leading to spontaneous follicle activation.

Study design, size, duration: Fresh ovarian cortical biopsies were obtained from six women aged 28-38 years (mean ± SD: 32.7 ± 4.1 years) at elective caesarean section. Biopsies were cut into fragments of ∼4 × 1 × 1 mm and cultured for 0, 2, 4, or 6 days (D).

Participants/materials, setting, methods: Primordial follicle activation, stromal cell density, and ECM-related protein (collagen, elastin, fibronectin, laminin) positive area in the entire cortex were quantified at each time point using histological and immunohistological analysis. Collagen and elastin content, collagen fibre characteristics, and follicle distribution within the tissue were further quantified within each layer of the human ovarian cortex, namely the outer cortex, the mid-cortex, and the cortex-medulla junction regions.

Main results and the role of chance: Primordial follicle activation occurred concomitantly with a loosening of the ovarian cortex during culture, characterized by an early decrease in stromal cell density from 3.6 ± 0.2 × 106 at day 0 (D0) to 2.8 ± 0.1 × 106 cells/mm3 at D2 (P = 0.033) and a dynamic remodelling of the ECM. Notably, collagen content gradually fell from 55.5 ± 1.7% positive area at D0 to 42.3 ± 1.1% at D6 (P = 0.001), while elastin increased from 1.1 ± 0.2% at D0 to 1.9 ± 0.1% at D6 (P = 0.001). Fibronectin and laminin content remained stable. Moreover, collagen and elastin distribution were uneven throughout the cortex and during culture. Analysis at the sub-region level showed that collagen deposition was maximal in the outer cortex and the lowest in the mid-cortex (69.4 ± 1.2% versus 53.8 ± 0.8% positive area, respectively, P < 0.0001), and cortical collagen staining overall decreased from D0 to D2 (65.2 ± 2.4% versus 60.6 ± 1.8%, P = 0.033) then stabilized. Elastin showed the converse distribution, being most concentrated at the cortex-medulla junction (3.7 ± 0.6% versus 0.9 ± 0.2% in the outer cortex, P < 0.0001), and cortical elastin peaked at D6 compared to D0 (3.1 ± 0.5% versus 1.3 ± 0.2%, P < 0.0001). This was corroborated by a specific signature of the collagen fibre type across the cortex, indicating a distinct phenotype of the ovarian cortical ECM depending on region and culture period that might be responsible for the spatio-temporal and developmental pattern of follicular distribution observed within the cortex.

Large scale data: N/A.

Limitations, reasons for caution: Ovarian cortical biopsies were obtained from women undergoing caesarean sections. As such, the data obtained may not accurately reflect the ECM distribution and structure of non-pregnant women.

Wider implications of the findings: Clarifying the composition and architecture signature of the human ovarian cortical ECM provides a foundation for further exploration of ovarian microenvironments. It is also critical for understanding the ECM-follicle interactions regulating follicle quiescence and awakening, leading to improvements in both in vitro activation and in vitro growth techniques.

Study funding/competing interest(s): Medical Research Council grant MR/R003246/1 and Wellcome Trust Collaborative Award in Science: 215625/Z/19/Z. The authors have no conflicts to declare.

Trial registration number: N/A.

Keywords: extracellular matrix; fertility preservation; mechanobiology; ovary; primordial follicle activation; tissue stiffness.

Figure 1.

Primordial follicle activation and early growth during in vitro culture. (A) Photomicrographs of human primordial (a), transitory (b), primary (c), and secondary (d) follicles after haematoxylin and eosin staining. Scale bar = 50 µm. (B) Distribution of the follicles according to their developmental stage at 0, 2, 4, and 6 days of culture (n = 6; 5703 follicles counted). Data are expressed as mean ± SEM, ANOVA followed by Tukey test. Different superscripts indicate P < 0.05.

Figure 2.

Distribution of the stromal cells within the human ovarian cortex during culture. (A) Representative images of the ovarian stroma at 0 (a), 2 (b), 4 (c), and 6 (d) days of culture after haematoxylin and eosin staining. Scale bar = 100 µm. (B) Box plots of stromal cell density (n = 6). ANOVA followed by Tukey test, *P < 0.05.

Figure 3.

Dynamics of collagen, elastin, fibronectin, and laminin proteins during in vitro culture. (A) Representative images of whole human ovarian cortex (2.5× magnification, a–d) and at higher magnification (10× magnification, a’–d’) of PicroSirius Red stained collagen (a, a’) and immunostained elastin (b, b’), fibronectin (c, c’), and laminin (d, d’). Collagen fibres stain in red, elastin in green, and DAPI in blue, fibronectin and laminin in brown. Scale bar = 1 mm (a–d) and 500 µm (a’–d’). (B–E) Quantification of the area of extracellular matrix (ECM)-related proteins positive staining per day of culture (n = 6). Data presented as box plots, ANOVA followed by Tukey test. *P < 0.05, **P < 0.01, ***P < 0.001. (F) Summary of the remodelling of all studied ECM component proteins in the human ovarian cortex during in vitro culture. For each protein, lines of best fit with 95% CI as shaded regions are represented.

Figure 4.

Cortical sub-regions- and culture-induced collagen and elastin remodelling. (A–C) collagen, (D–F) elastin. (A, D) Representative images of entire human cortical sections stained with PicroSirius Red (A) or elastin (D). Collagen fibres stain in red, elastin in green, and DAPI in blue. Scale bar = 1 mm. (B, E) Selected regions of interest within the outer cortex, the mid-cortex, and the cortex–medulla junction sub-regions. Scale bar = 100 µm. (C, F) Quantification of the area of PicroSirius Red- (C) and elastin- (F) positive staining depending on the cortical sub-region and the culture period (n = 6). Data are expressed as mean ± SEM, univariate general linear models. Different superscripts indicate P < 0.05 within each of the cortical sub-region; ****P < 0.0001.

Figure 5.

Collagen deposition in the human ovarian cortical stroma visualized under polarized light. (A, B) Representative PicroSirius Red (PSR) staining of an ovarian cortical section (A) and of examined cortical sub-regions (B)—outer cortex (a), mid-cortex (b), and cortex–medulla junction (c)—viewed under polarized light. Scale bars = 500 and 50 µm, respectively. (C) PSR-stained samples from (B) were analysed using a custom FIJI macro to threshold the three main colours seen under polarized light in: red, yellow, and green. (D) Quantification of the relative percentage of each collagen colour type (n = 6). Collagen fibre thickness varies according to its location within the cortex but is not significantly remodelled over time within each sub-region. Data are expressed as mean ± SEM, univariate general linear models. Comparison between cortical sub-regions: **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 6.

Characteristics of the human ovarian cortical collagen architecture. (A) Representative image of a region of interest stained with PicroSirius Red and viewed under polarized light. Field size = 137 × 137 µm². (B) Graphical output from curvelet transform-fibre extraction (CT-FIRE) showing automatic extraction of collagen fibres. (C) Graphical outputs from CurveAlign showing fibre overlaid image that is converted into a direction heat map. (D) Quantification of collagen fibres density, width, length, straightness, angle, and alignment (n = 6 patients; >17 000 fibres analysed per region and day of culture). Collagen fibre metrics vary according to its location within the cortex but are not significantly remodelled over time within each sub-region. Data are expressed as mean ± SEM, univariate general linear models. Comparison between cortical sub-regions: **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 7.

Follicle spatial distribution within the human cortex during in vitro culture. Follicles were classified according to both their developmental stage and location within the cortex at 0 and 6 days of culture (n = 6; 2805 follicles counted). Data are expressed as mean ± SEM, multivariate general linear models. ****P < 0.0001.