Follicle structure and vitality

A schematic illustration of the in vitro culture (IVC) method used in this study is depicted in Fig. 1a. Light microscopy images of the isolated follicles are presented in Fig. 1b, demonstrating that healthy follicles are composed of oocytes and somatic cells. Individual isolated follicles stained with H&E are shown in Fig. 1c, confirming the structural integrity of the follicles. Figure 1d shows that the isolated follicles survived after mechanical isolation. Conclusively, our results revealed that mechanical isolation does not damage follicle structure and viability, and follicles isolated in this manner can be used for IVC and transplantation.

Fig. 1

Structure and viability of the follicles. (a) Schematic illustration of in vitro culture. (b) Light microscopy images of isolated follicles. (c) Isolated follicle stained with hematoxylin and eosin. (d) Follicular viability assessment by staining with calcein-AM and ethidium homodimer-I

Follicle growth and steroid production in vitro

Morphologically normal follicles were cultured after isolation. Light microscopy showed that the follicles had intact spherical structures in the ALG and CTP groups during IVC (Fig. 2a). A fluid-filled antrum cavity was observed on day 12 in both the ALG and CTP groups, consistent with the in vivo morphology. Conversely, the link between oocytes and granulosa cells (GCs) was broken in the 2D groups because the GCs adhered to the bottom of the culture dish, resulting in eventual follicular damage (Fig. 2a). The follicle diametres in the CTP groups on days 6 and 12 were significantly larger than those in the ALG groups (P < 0.05) (Fig. 2b). Moreover, the survival rate of follicles in the CTP groups was significantly higher than that in the ALG groups on day 12 of IVC (P < 0.05) (Fig. 2c), suggesting that CTP hydrogels could support follicles for longer duration than ALG hydrogels in vitro.

Fig. 2

Growth and hormone secretion. (a) Morphology of follicles cultured from days 0–12. (b) Growth curve over a 12-day culture period. (c) Follicular survival rate in vitro measured on days 6 and 12. (d-f) Androstenedione, oestradiol, and progesterone levels secreted into the culture medium on days 6 and 12. The data are represented as the mean ± standard deviation. * indicates significance at P < 0.05 compared with 2D, and # indicates significance at P < 0.05 compared with alginate hydrogels (ALG) determined using one-way ANOVA, followed by Student–Newman–Keuls post-hoc analysis

Ovarian hormones, including androstenedione (A), oestradiol (E), and progesterone (P), are secreted by theca cells and GCs. The levels of A in the ALG groups were significantly lower than that in the CTP groups on day 6 of IVC but higher than that in the CTP groups on day 12 of IVC (P < 0.05) (Fig. 2d). Levels of E in the CTP groups were significantly higher than those in the ALG groups on days 6 and 12 (P < 0.05) (Fig. 2e). Levels of P in the ALG groups were significantly lower than those in the CTP groups on day 6 of IVC (P < 0.05) but higher than those in the CTP groups on day 12 of IVC (P > 0.05) (Fig. 2f). These results imply that follicular growth was supported in the CTP hydrogels, and more oestrogen was produced in the CTP groups than in the ALG groups.

Characterisation of differential gene expression using RT-PCR and oocyte meiotic competence

At the end of the culture period, the expression levels of folliculogenesis-related genes were assessed. PCNA, FSH-R, and AMH are considered key markers of GCs. Figure 3a-c shows that the expression levels of PCNA, FSH-R, and AMH were higher in the CTP groups than those in the other two groups, indicating that the CTP groups contained more GCs (P < 0.05). BMP15 is an oocyte-secreted factor crucial for regulating follicular growth [27, 33]. Our findings showed that mRNA expression of BMP15 was higher in the CTP groups than in the ALG groups (Fig. 3d). Cyp19a1, HSD3B1, CYP17a1, and CYP11a1 are important enzymes that convert androgen to oestrogen; the higher the activity of these enzymes, the more oestrogen is produced [34]. Figure 3e-h demonstrate that the expression levels of Cyp19a1, HSD3B1, CYP17a1, and CYP11a1 were higher in the CTP groups than those in the other two groups, consistent with the results shown in Fig. 2e. STAR is known to be responsible for transporting cholesterol and providing substrates for oestrogen production [35]. Figure 3i shows that the expression of STAR mRNA in the CTP groups was higher than that in the other two groups (P < 0.05). These results reveal that the CTP hydrogels are more suitable for follicular growth, as indicated by increased GC proliferation, hormone secretion, and steroidal aromatase activity.

Fig. 3

Real-time polymerase chain reaction and oocyte meiotic competence. (a-i) Analysis of PCNA, FSH-R, AMH, BMP15, CYP19a1, HSD3B1, CYP17a1, CYP11a1, and STAR expression levels on day 12 of culture. (j-l) Representative images of oocyte stages after in vitro follicle maturation in CTP hydrogels. GV germinal vesicle in a mature oocyte (j) GVBD germinal vesicle breakdown (k), The first polar body (arrow) was extruded in metaphase II (M II) (l). GAPDH was used as the internal control. The data are represented as the mean ± standard deviation. * indicates significance at P < 0.05 compared with 2D, and # indicates significance at P < 0.05 compared with alginate hydrogels (ALG) determined using one-way ANOVA, followed by Student–Newman–Keuls post-hoc analysis

Subsequently, the quality of oocytes obtained from follicles cultured in hydrogels was measured based on their ability to resume meiosis. Follicles encapsulated in CTP hydrogels could resume meiosis and produce meiotically competent oocytes, as shown in Fig. 3j-k. Moreover, the percentage of M II stage oocytes obtained from follicles cultured in CTP hydrogels (78.9%) was significantly higher than that cultured in ALG hydrogels (56.7%), as shown in Table 2.

Table 2 Oocyte meiotic competence Evaluation of the grafts post-transplantation

After 1 week of grafting, primary, secondary, and antral follicles were observed in ALG and CTP groups (Fig. 4e). Follicular recovery rate on day seven in CTP hydrogels was 28% (42 normal follicles out of 150 grafted follicles), compared to 17.2% (25 normal follicles out of 145 grafted follicles) in ALG hydrogels, showing a significant difference between the two groups (χ2 = 4.862, P = 0.027) (Fig. 4i). After 10 weeks of grafting, antral follicles were observed only within the CTP groups and not in the ALG groups (Fig. 4j). Moreover, large amounts of undegraded ALG hydrogels were observed, as indicated by the light pink stain; however, only a small amount of undegraded CTP hydrogels were observed (Fig. 4j).

Fig. 4

Evaluation of the grafts post-transplantation. (a) Blood vessels and endothelial cells stained with CD34; black arrows indicate capillaries. (b) Ki67-positive cells were stained brown (black arrows). (c) CD45-positive cells observed in the grafts. (d) Terminal deoxynucleotidyl deoxyuridine triphosphate nickend labeling (TUNEL) assay illustrating apoptotic cells. Red fluorescence: TUNEL-positive nuclei; blue fluorescence: DAPI. (e) Representative images of haematoxylin and eosin (H&E) staining of the grafts at 1 week after transplantation. Primary follicles (white arrows), secondary follicles (black arrows), antral follicles (asterisk), and CTP hydrogels (black rectangles) were observed. Average optical density values of CD34 (f), Ki-67 (g), CD45 (h), and follicle recovery rate (i) were quantified. (j) Representative images of H&E staining of the grafts at 10 weeks after transplantation. Antral follicles (asterisk) and ALG hydrogels (white rectangle) were observed. $ indicates significance at P < 0.05 compared with the CTP-F groups; & indicates significance at P < 0.05 compared with the ALG groups, determined using one-way ANOVA, followed by Student–Newman–Keuls post-hoc analysis

In this study, CD34 and Ki-67 levels were used as indicators of revascularisation and cell proliferation, respectively. Figure 4a shows that the average optical density (AOD) of CD34 in the CTP groups was higher than that in the ALG groups after 1 week of transplantation (P < 0.05). Moreover, the Ki-67 staining results showed significantly increased cell proliferation in mice transplanted with CTP grafts compared to those with ALG grafts (Fig. 4b). CD45 immunostaining was used to evaluate the healing process after injury induced by scraping the peritoneal surface of the mice. No fibrous capsules were observed around the grafts, and CD45-positive cells were identified in the grafts (Fig. 4c); the three groups showed no significant differences. Results of the TUNEL assay showed that a small number of apoptotic cells appeared in all three groups, with fewer apoptotic cells in the CTP groups than those in the ALG groups (Fig. 4d). Conclusively, our results demonstrated that the CTP matrix is biodegradable and can promote angiogenesis and cell proliferation.

Hormone levels and their effects on body composition after AO transplantation

After isolation, the follicles were encapsulated in CTP or ALG hydrogels to construct AOs and then transplanted into OVX mice to investigate whether AO could mimic the functions of natural ovaries. Figure 5a shows a schematic illustration of the experimental design. Serum levels of both follicle-stimulating hormone (FSH) and luteinizing hormone (LH) were elevated in OVX mice owing to the loss of the negative feedback control loop (Fig. 5c-d). CTP-F mice exhibited lower levels of serum E2 and higher levels of FSH and LH than the sham mice (Fig. 5b-d). These results indicated the successful construction of the OVX model. Transplantation with ALG and CTP grafts significantly increased the serum E2 levels and decreased the levels of serum FSH and LH compared with those in OVX and CTP-F mice (P < 0.05). The physiological levels of FSH, LH, and E2 were remarkably maintained until 8 weeks in the CTP groups, and the function of the CTP grafts was sustained for 2 weeks longer than that of the ALG grafts (Fig. 5b-d).

Fig. 5

Hormone levels in the artificial ovaries and their effects on body composition. (a) Schematic illustration of the animal experiments. (b-d) Serum E2, FSH, and LH levels were measured twice a week for 10 weeks. (e) Body weight was measured at 6 and 10 weeks after transplantation. (f) Visceral fat was measured at 6 and 10 weeks after transplantation. (g) Rectal temperature was measured twice a week for 10 weeks. * indicates significance at P < 0.05 compared with the sham groups; # indicates significance at P < 0.05 compared with the OVX groups; $ indicates significance at P < 0.05 compared with the CTP-F groups; & indicates significance at P < 0.05 compared with the ALG groups, determined using one-way ANOVA, followed by Student–Newman–Keuls post-hoc analysis

Oestrogen loss leads to body fat accumulation [36] and body weight [37] gain. Figure 5e-f depicts this phenomenon by comparing the sham and OVX groups. Six weeks after transplantation, mice in the OVX groups transplanted with the ALG and CTP grafts exhibited body weights comparable to those in the sham groups (Fig. 5e). Although the mean body weights of mice in the ALG and CTP groups were higher than those in the sham groups at 10 weeks after transplantation, the mean body weights of mice in the CTP groups were lower than those in the ALG groups (P < 0.05) (Fig. 5e). Moreover, the amount of visceral fat in mice in the CTP groups was lower than those in the ALG groups at 10 weeks after transplantation (P < 0.05) (Fig. 5f). Rectal temperature was significantly higher in OVX mice than in sham mice, suggesting that the OVX mice experienced hot flushes (P < 0.05) (Fig. 5g). Mice in the OVX groups transplanted with ALG and CTP grafts exhibited rectal temperatures comparable to those in the sham groups at 6 weeks after transplantation. However, the rectal temperatures of mice in the ALG groups were similar to those in the OVX groups at week 8, and the mice in the CTP and OVX groups showed similar rectal temperatures at week 10 (Fig. 5g). These results indicated that AOs constructed using ALG and CTP hydrogels could function normally in OVX mice and that ALG and CTP grafts completely prevented the increase in body weight and rectal temperature in OVX mice at 6 weeks after transplantation. The therapeutic effects of CTP grafts were notably superior to those of ALG grafts in OVX mice at 10 weeks after transplantation.

Transplantation of AOs restored OVX-induced uterine atrophy

The uterus begins to atrophy after losing oestrogen-induced protection. We observed that the uteri of OVX mice became thinner and longer than those of sham mice (Fig. 6a-b). However, the OVX mice transplanted with ALG and CTP grafts demonstrated general uterine morphologies similar to those of the sham mice at 6 and 10 weeks after transplantation. Figure 6e shows that the uterine indices of OVX and CTP-F mice were significantly lower than those of the sham mice (P < 0.05); however, the uterine indices of OVX mice transplanted with both ALG and CTP grafts did not differ from those of the sham mice at 6 weeks after transplantation (P > 0.05). Conversely, 10 weeks after transplantation, the CTP grafts resulted in uterine indices comparable to the values in the sham groups (P > 0.05), and the ALG grafts yielded comparable indices to those in the OVX groups (P > 0.05). Subsequently, we evaluated the histological changes in the uterus after transplantation. We observed that the endometrium became thinner, and the number of glands decreased in OVX and CTP-F mice compared with the sham mice (Fig. 6c-d). In contrast, the endometrial thickness increased and more glands appeared in the ALG and CTP groups, similar to the levels observed in the sham groups at 6 weeks after transplantation (Fig. 6f-g). At 10 weeks after transplantation, endometrial thickness and gland density of mice in the ALG groups were significantly different from those in the sham groups; however, these values were similar between mice in the CTP and sham groups (Fig. 6f-g). Our results demonstrated that CTP grafts improved uterine atrophy associated with oestrogen deficiency better than ALG grafts at 10 weeks after transplantation.

Fig. 6

Effects of artificial ovaries transplantation on the uteri of ovariectomised (OVX) mice. (a) Representative images of the uterus at 6 weeks after transplantation. (b) Representative images of the uteri at 10 weeks after transplantation. (c) Images of the uterine sections stained with haematoxylin and eosin (H&E) at 6 weeks after transplantation (En, endometrium). (d) Images of the uteri stained with H&E at 10 weeks after transplantation (En, endometrium). (e) Uterine index after transplantation. (f) The thickness of the endometrium and (g) the number of glands in the uterus were calculated in each group after transplantation. The data are represented as the mean ± standard deviation. * indicates significance at P < 0.05 compared with the sham groups; # indicates significance at P < 0.05 compared with the OVX groups; $ indicates significance at P < 0.05 compared with the CTP-F groups; & indicates significance at P < 0.05 compared with the ALG groups, determined using one-way ANOVA, followed by Student–Newman–Keuls post-hoc analysis

AO transplantation restored OVX-induced vaginal atrophy

The vagina begins to atrophy and lose mass after OVX [38]. The oestrus cycle in mice is generally categorised into four stages (proestrus, oestrus, metestrus, and dioestrus) [39]. The oestrus cycle of mice in the OVX and CTP-F groups was arrested at the metestrus/dioestrus stage; however, mice in the ALG and CTP groups exhibited dynamic and regular cycling even after transplantation (Fig. 7a). Moreover, 90% and 85% of the mice in the CTP and ALG groups, respectively, showed dynamic cycling at 6 weeks after transplantation (Fig. 7d). The number of mice with oestrus cycles decreased gradually over time, and no mice in the ALG groups experienced oestrus cycles; however, approximately 25% of the mice in the CTP groups experienced oestrus cycles at week 10 (Fig. 7d). Figure 7e shows that the vaginal indices of mice in the ALG groups were not different from those in the OVX groups; nevertheless, the vaginal indices were significantly different between the CTP and OVX groups at 10 weeks after transplantation (P < 0.05). Finally, the thicknesses of the vaginal epithelium and epithelial cellular layer were quantitatively analysed using H&E staining (Fig. 7b-c). OVX mice exhibited thinner vaginal epithelia and fewer vaginal epithelial cellular layers than sham mice (Fig. 7f-g). Conversely, compared with the vaginal epithelial or epithelial cellular layer thickness of mice in the sham groups, those in the ALG and CTP groups showed no significant differences at 6 weeks after transplantation (Fig. 7f-g). At 10 weeks after transplantation, the thicknesses of both vaginal epithelia and epithelial cellular layers of mice in the CTP groups were similar to those in the sham groups; however, these parameters significantly differed between mice in the ALG and sham groups (Fig. 7f-g).

Fig. 7

Effects of artificial ovaries transplantation on the vagina in ovariectomised (OVX) mice. (a) Images a1-a4, b1-b4, and c1-c4 depict representative cytology images of the OVX mice, ALG mice, and CTP mice, respectively. (b) Images of the vagina stained with haematoxylin and eosin (H&E) at 6 weeks after transplantation. (c) Images of the vagina stained with H&E at 10 weeks after transplantation. (d) Proportion of mice with oestrus cycles after transplantation. (e) Vaginal indices after transplantation. (f) The thickness of epithelial cells and (g) number of epithelial layers in the vagina were calculated for each group after transplantation. The data are represented as the mean ± standard deviation. * indicates significance at P < 0.05 compared with the sham groups; # indicates significance at P < 0.05 compared with the OVX groups; $ indicates significance at P < 0.05 compared with the CTP-F groups; & indicates significance at P < 0.05 compared with ALG, determined using one-way ANOVA, followed by Student–Newman–Keuls post-hoc analysis

AO transplantation restored OVX-induced bone loss

Oestrogen deficiency can lead to osteoporosis and fractures [40]. To examine the effects of AOs on the loss of bone mass in OVX mice, femurs were collected for micro-CT imaging and H&E staining (Fig. 8a-d). H&E staining revealed that the structure of the trabeculae was disordered, number of trabeculae decreased, and number of fat vacuoles increased in the bone marrows of mice in the OVX and CTP-F groups compared to those in the sham groups (Fig. 8a). Micro-CT imaging confirmed that bone mineral density (BMD), trabecular number (Tb.N), trabecular thickness (Tb.Th) and percent bone volume (BV/TV%) were significantly reduced, the trabecular separation (Tb.Sp) and structural model index (SMI) markedly increased in the OVX mice compared to sham mice (Fig. 8e). The ALG and CTP grafts notably alleviated bone loss in the distal femur of OVX mice; however, none of the treated groups exhibited bone loss restoration similar to the sham groups at 6 and 10 weeks after transplantation (Fig. 8e). At 10 weeks after transplantation, the parameters BV/TV%, Tb.N, Tb.Sp and Tb.Th in the ALG mice were not significantly different from those in the OVX mice; however, these parameters differed significantly between the ALG and the sham mice (P < 0.05). However, the BMD, Tb.N, Tb.Th and BV/TV% in the CTP mice were significantly different from those in the OVX and ALG mice, as shown in Fig. 8e. Conclusively, transplantation with AOs partially protects against OVX-induced osteoporosis, and the therapeutic effects of CTP grafts were found to be superior to those of ALG grafts at 10 weeks after transplantation.

Fig. 8

Artificial ovaries transplantation restored ovariectomised (OVX)-induced bone loss after transplantation. (a) Representative images of haematoxylin and eosin (H&E) staining and (b) micro-computed tomography (micro-CT) imaging analysis at 6 weeks after transplantation. (c) Representative images of H&E staining and (d) micro-CT imaging analysis at 10 weeks after transplantation. (e) Quantitative analyses of BV/TV%, Tb.Sp, Tb.N, Tb.Th, BMD, and SMI. The data are represented as the mean ± standard deviation. * indicates significance at P < 0.05 compared with the sham groups; # indicates significance at P < 0.05 compared with the OVX groups; $ indicates significance at P < 0.05 compared with the CTP-F groups; & indicates significance at P < 0.05 compared with ALG groups, determined using one-way ANOVA, followed by Student–Newman–Keuls post-hoc analysis

Degradability and biocompatibility of scaffolds in vivo

The biocompatibility of CTP hydrogels in vivo is crucial for its application in the fabrication of AOs. Figure 9a shows that the ALG hydrogels were non-degradable, consistent with previous data [12]. Conversely, the CTP hydrogels were degradable, consistent with the 3D-hydrogel criteria (Fig. 9a). Furthermore, we observed no evident signs of tissue damage or inflammatory lesions in any of the major organs (heart, liver, spleen, lung, and kidney) (Fig. 9b). Biochemical analyses of hepatic and kidney function revealed no significant differences between mice in the CTP and sham groups (P > 0.05) (Fig. 9c-f). Thus, the biocompatibility of the CTP hydrogels is satisfactory for its future biomedical applications.

Fig. 9

Degradability and biocompatibility of scaffolds in vivo. (a) Representative images of the grafts. (b) Representative images of haematoxylin and eosin staining of the heart, liver, spleen, lung, and kidney tissue sections from the CTP and ALG groups for biocompatibility validation in vivo at 10 weeks after transplantation. (c-f) Biochemical analyses of hepatic (ALT and AST) and kidney function (BUN and CREA) after implantation with hydrogels for 10 weeks