Rapamycin prevents cyclophosphamide-induced ovarian follicular loss and potentially inhibits tumour proliferation in a breast cancer xenograft mouse model

Abstract

Introduction

The primordial follicle pool is approximately assembled at birth and represents female reproductive capacity. Certain chemotherapies induce abnormal activation and depletion of primordial follicles within this fixed pool and irreversibly diminish the ovarian reserve, resulting in premature ovarian insufficiency (POI). Chemotherapy-induced infertility in the adolescent and young adult (AYA) generation is of particular concern. Breast cancer is the most common cancer among AYAs (Nakata et al., 2022) with an increasing incidence (Barr et al., 2016). Disease-free and overall survival outcomes in the AYA breast cancer setting are worse, with more aggressive disease and higher rates of mastectomy than in patients with adult breast cancer (Sauder et al., 2020; Youlden et al., 2020; Cathcart-Rake et al., 2021). Since higher rates of triple-negative breast cancer and human epidermal growth factor type 2 (HER2)-positive subtypes were detected in patients with AYA breast cancer, 55.5% of patients have received chemotherapy, including neoadjuvant chemotherapy (Kataoka et al., 2014). Cyclophosphamide, the most gonadotoxic anticancer drug, is frequently administered to treat breast cancer. Therefore, protecting fertility against chemotherapy-induced ovarian dysfunction is critical, and the prevention of cyclophosphamide-induced ovarian dysfunction in patients with breast cancer warrants further study.

There are three fertility preservation strategies in female patients with cancer: frozen embryo/egg preservation, ovarian tissue cryopreservation and transplantation (Diaz-Garcia et al., 2018; Dolmans et al., 2021), and ovarian protection with drugs (Roness et al., 2016). Drug-assisted ovarian protection is a valuable option as it can be used in combination with either of the other two methods and has no noticeable disadvantages; however, no drug with a high level of supportive evidence has been reported. Currently, the GnRH agonists have accumulated the most evidence as ovarian protective agents. However, the European Society for Medical Oncology (ESMO) (Lambertini et al., 2020) and American Society of Clinical Oncology (ASCO) (Oktay et al., 2018) guidelines emphasize that GnRH agonists are a therapeutic option but not a substitute for embryo/oocyte freezing and are recommended for use in situations where these are not available. The ESHRE Guideline Group on Female Fertility Preservation (2020) similarly states that GnRH agonists, although a treatment option, have limited protective effects on ovarian reserve and future fertility. Although several randomized controlled trials have shown the efficacy of GnRH agonists (Moore et al., 2015), others have failed to demonstrate efficacy (Munster et al., 2012), whereas others have not shown efficacy but argue limited effectiveness (Leonard et al., 2017). Hence, ovarian protective agents other than GnRH analogues are eagerly anticipated in clinical practice.

A breakthrough discovery was made regarding chemotherapy-induced ovarian dysfunction: cyclophosphamide induces ovarian damage in part by activation of the mammalian target of the rapamycin (mTOR) pathway, leading to primordial follicle activation and follicular ‘burnout’ (Adhikari and Liu, 2010). Based on this accumulated evidence, we previously reported the ovarian protective effects of mTOR inhibitors (Tanaka et al., 2018). Others have supported our supposition on the ovarian protective effect of mTOR inhibitors (Goldman et al., 2017; Zhou et al., 2017; Xie et al., 2020; Chen et al., 2022a; Kashi et al., 2023). Meanwhile, mTOR inhibitors have several other effects (Li et al., 2014); in terms of safety, mTOR inhibitors have proven relatively safe long-term oral agents in human solid organ transplantation and tuberous sclerosis (Somers and Paul, 2015; Ventura-Aguiar et al., 2016). Notably, mTOR inhibitors also reportedly elicit antitumour effects in breast cancer (Zeng et al., 2010; Baselga et al., 2012).

The greatest challenge facing existing research on protective agents against chemotherapy-induced ovarian hypofunction is the relative lack of reports in cancer-bearing mice. The most common approach for studying ovarian-protective agents is to administer chemotherapeutic and ovarian-protective agents to non-cancerous mice. However, chemotherapy dosage regimens have not been studied based on antitumour efficacy but on the basis of chemotherapy-induced ovarian dysfunction; that is, unlike real-world clinical practice, chemotherapy was administered as a single dose or cyclical administration of low doses. Furthermore, since the subject mice are not carcinoma bearing, it is unclear whether cyclophosphamide administration sufficiently reproduces antitumour effects. Moreover, the lack of observed effects elicited by ovarian protective agents on tumours has proven problematic.

Accordingly, the primary objective of the present study was to develop a breast cancer mouse model that is cyclically administered cyclophosphamide (a potent standalone antitumour agent) and simultaneously provided an ovarian protective agent, rapamycin (an mTOR inhibitor known for its antitumour properties). Subsequent analyses assessed the combined effects of these treatments on ovarian protection and breast cancer antitumour activity.

Materials and methods

In vivo xenograft breast cancer model

All procedures were approved by the Institutional Animal Care and Use Committee (approval number: 2022-6-16). Female BALB/c nude mice at the age of 4 weeks were purchased from SLC (Japan), housed under standard conditions with a 12-h light/12-h dark cycle, and fed standard pelleted rodent chow. The sample size was determined based on a previous in vivo study with rapamycin and cyclophosphamide co-treatment in breast cancer (Zeng et al., 2010). The mice were acclimatized for 1 week. MDA-MB-231 human mammary adenocarcinoma cells, expressing low-level HER2, also express low levels of oestrogen and progesterone receptors and are highly invasive. MDA-MB-231 cells were purchased from the American Type Culture Collection (ATCC, USA). Cells were cultured in Dulbecco’s modified eagle medium (DMEM) supplemented with 10% foetal calf serum (FCS), l-glutamine, and penicillin/streptomycin solution (all Gibco Invitrogen, USA) in a humidified 5% CO₂, 95% ambient air atmosphere at 37°C. An in vivo xenograft breast cancer model was established by injecting 5 × 10⁶ MDA-MB-231 cells (in 100 μl of DMEM) into the mouse flank subcutaneously.

Short-term treatment protocol for evaluating ovarian follicle apoptosis

In brief, 5 × 10⁶ MDA-MB-231 cells were injected into the mouse flank subcutaneously. To replicate AYA generation in human clinical practice, the treatment protocol was initiated when the mice were 8 weeks old (3 weeks after tumour cell injection). Eight-week-old mice are considered to approximate 20- to 30-year-old humans (Dou et al., 2017). Three weeks after tumour injection, mice were randomized into three treatment groups: vehicle group (control), cyclophosphamide-only group (Cy), and cyclophosphamide + rapamycin combination group (Cy + Rap) (n = 5–6/group). Cyclophosphamide (FUJIFILM, Japan) was dissolved in phosphate-buffered saline (PBS); rapamycin (FUJIFILM, Japan) supplied in dimethyl sulfoxide (DMSO) at −20°C was diluted in PBS before use. Additionally, 5 mg/kg of rapamycin or vehicle (DMSO in PBS) was intraperitoneally administered daily until sacrifice. After 7 days of rapamycin administration, a single 120 mg/kg dose of cyclophosphamide or vehicle control was administered intraperitoneally. Mice were sacrificed 24 h after cyclophosphamide administration. The ovaries were harvested and immediately submerged in 4% (v/v) paraformaldehyde (PFA) and analysed using the Terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) assay (Fig. 1).

Figure 1.

Experimental schema: tumour cells were transplanted into 5-week-old mice, and 3 weeks after transplantation (at 8 weeks of age), the treatment protocol was initiated. Rapamycin (green) at a dosage of 5 mg/kg, or its vehicle, was administered daily from treatment Week 1, Day 1, through Week 5, Day 1. Cyclophosphamide (red) at a dosage of 120 mg/kg, or its vehicle, was administered weekly from treatment Week 2, Day 1, through Week 4, Day 1. To analyse ovarian follicle apoptosis, mice were sacrificed on treatment Week 2, Day 2, i.e. 24 h after a single cyclophosphamide administration. To perform follicle counting and assess the mTOR pathway in the ovary, as well as the antitumour effects, mice were sacrificed on treatment Week 5, Day 1, i.e. 7 days after the last cyclophosphamide administration.

Open in new tabDownload slide

Long-term treatment protocols for evaluating follicle and tumour development

In brief, 5 × 10⁶ MDA-MB-231 cells were injected into the mouse flank subcutaneously (n = 20 mice). Three weeks after tumour injection, 19 mice with palpable tumours were randomly divided into three groups (n = 6 or 7/group): control, Cy, and Cy + Rap groups. Rapamycin (5 mg/kg) or vehicle was administered intraperitoneally (daily) for 29 days. Additionally, 120 mg/kg cyclophosphamide or vehicle was administered intraperitoneally (weekly) on Day 1 of Weeks 2–4. Mice were euthanized on Day 1 of Week 5. Considering that steady-state mTOR inhibitor concentrations are not reached for ∼7 days (Kirchner et al., 2004), we administered mTOR inhibitors for 7 days before the first exposure to chemotherapy.

In cancer xenograft models, treatment protocols are typically initiated when the tumour volume reaches a predetermined value after tumour transplantation. Accordingly, the age of mice at treatment initiations can vary considerably among studies. However, in studies on ovarian dysfunction, age-related ovarian dysfunction must be excluded; therefore, the number of weeks following tumour cell transplantation and before initiating treatment was fixed in the present study. Tumour growth was evaluated using tumour volume index rather than absolute tumour volume after confirming that no significant differences existed in the absolute tumour volume between the groups at treatment initiation. Tumour volume index indicates the relative volume, compared to the volume 3 weeks after tumour cell injection (at treatment initiation) set to 100%. The body weights of the mice were recorded daily.

At the time of sacrifice, the ovaries and tumours were harvested and immediately immersed in 4% PFA or liquid nitrogen (Fig. 1). The follicle count assay was performed on ovaries immersed in PFA via haematoxylin and eosin (HE) staining; the mTOR pathway assay was performed based on phospho-S6 kinase (pS6K) immunohistochemistry. Meanwhile, the proliferation of cells in tumours immersed in PFA was analysed based on Ki-67 immunohistochemistry. The ovaries and tumours immersed in liquid nitrogen were analysed for mTOR activation by western blotting.

Ovarian follicular apoptosis assay (TUNEL immunostaining)

The commercial TUNEL assay In Situ Cell Death Detection Kit-POD (Roche, Switzerland) was employed to evaluate follicle cell apoptosis. After dewaxing and rehydrating, the sections were pretreated with proteinase K and immersed in 3% H₂O₂. Next, 50 µl of TUNEL working solution was added to each sample and mixed with label solution and enzyme solution; negative controls were treated with label solution. The sections were incubated at 37°C for 60 min in the dark and analysed by diaminobenzidine (DAB) staining. If more than 50% positive cells and/or positive oocytes were present, the primordial follicle was defined as being in an apoptotic state. As a positive control, a paraffin-embedded ovarian section treated with DNase I (Takara Bio, Japan)was appropriately stained.

Ovarian follicle classification and count

The ovaries were fixed in 4% PFA for 24 h, dehydrated, placed horizontally along the long-axis, and paraffin-embedded such that the cross-sectional area was as large as possible when cut. The whole ovaries were sliced every 25 µm into 5-µm sections and follicles were enumerated for the entire ovary. The 25-μm serial sections were rehydrated and HE-stained for morphological observation and differential follicle counts. Blind follicle classification and counting were performed on all ovarian sections; the totals for all sections were summed. Differential follicle counts were performed, and the follicle stage was classified according to accepted published definitions (Fig. 2) (Myers et al., 2004). A follicle was defined as primordial if it was surrounded by a single layer of squamous granulosa cells. A primary follicle was an oocyte surrounded by a single layer of cuboidal granulosa cells. Secondary follicles were defined as two or more layers of cuboidal granulosa cells without an antrum. Antral follicles are defined as two or more layers of cuboidal granulosa cells within the antral cavity. Only follicles containing an oocyte were counted to prevent double follicle counts. Zona pellucida remnants represented atretic follicles. The ratio of primary to primordial follicles was calculated by dividing the total number of primary follicles in each treatment group by the total number of primordial follicles in each treatment group. The minor axis index of the ovary was defined as the number of sections in which ovarian tissue was observed in consecutive sections.

Figure 2.

Follicle count method. (A) The ovaries were placed horizontally in the long-axis direction and paraffin-embedded so that the cross-sectional area would be as large as possible when cut. Ovaries were cut every 25 µm from end to end into 5- µm thickness sections and were rehydrated and stained with haematoxylin and eosin (HE) for morphological observation and differential follicle counts. Follicle classification and counting were performed blindly on all sections of the ovaries, and the results were then summed up across all sections. (B) Primordial follicle oocytes were surrounded by a single layer of squamous granulosa cells (arrows). Bar = 10 μm. (C) A primary follicle was defined as an oocyte surrounded by a single layer of cuboidal granulosa cells (arrow heads). Bar = 20 μm. (D) A secondary follicle was defined by two or more layers of cuboidal granulosa cells with no antrum. Bar = 50 μm. (E) Antral follicle was defined by two or more layers of cuboidal granulosa cells with antral cavity (star). Bar = 100 μm. (F) Zona pellucida remnants representing end-stage atretic follicles (circle); Bar = 100 μm.

Open in new tabDownload slide

Ovarian follicular mTOR pathway activation assay (pS6K immunostaining)

pS6K is a downstream serine/threonine kinase of mTOR (Goldman et al., 2017). Localization of mTOR pathway activation within the ovary was assessed by pS6K immunohistochemistry. Using the fragment next to the largest diameter observed in the HE-stained ovarian specimen, sections were cut and stored as representative sections for immunostaining. Paraffin-embedded ovarian sections were warmed and serially de-paraffinized in xylene and ethanol, incubated in EDTA buffer (pH 9.0), heated in a microwave for 10 min, and immersed in 3% H₂O₂ for 10 min at room temperature to block endogenous peroxidase activity. Nonspecific binding was blocked with an antigen-unmasking solution for 20 min at room temperature. The sections were incubated in a 1:50 dilution of the primary antibody phospho-rpS6 (9205; Cell Signalling Technology, USA) overnight at 4°C. For immunohistochemistry, sections were incubated with appropriate secondary antibodies for 2 h at room temperature, followed by DAB staining. The proportion of pS6K-positive cells among the primordial follicles was quantified (n = 4–5/group). As a positive control, a paraffin-embedded kidney section was appropriately stained.

Tumour cell proliferation assay (Ki-67 immunostaining)

Tumour cell proliferation was visualized via anti-Ki-67 immunohistochemistry (1:400) as previously described in an in vivo breast cancer study (Namba et al., 2006; Zeng et al., 2010; Macaskill et al., 2011). Three representative fields of view for tumours excised from each group (n = 2–3/group) were analysed under 40× magnification. Ki-67-positive nuclei were quantitated as percentages of the total nuclei. The average of three fields of view served as the cell proliferation index for the tumours. As a positive control, a paraffin-embedded spleen section was appropriately stained.

Ovary and tumour mTOR pathway activation assay (western blotting)

Mouse ovaries were pulverized, subsequently extracted by radio-immunoprecipitation assay buffer (Nacalai Tesque Inc., Japan), and centrifuged at 10 000 × g at 4°C for 10 min. The supernatant was collected and stored at −80°C until western blotting was performed. The protein content was determined using a bicinchoninic acid assay (BCA) protein assay kit (Thermo Fisher Scientific Inc., USA); 15 μg of the protein from each sample was loaded onto 4–20% polyacrylamide gel for sodium dodecyl sulphate polyacrylamide gel electrophoresis. Proteins of interest were separated by electrophoresis and transferred to polyvinylidene fluoride membranes. The membranes were blocked in 3% bovine serum albumin solution for 1 h and probed with specific primary antibodies overnight at 4°C. Primary antibodies against mTOR (# 2983), phospho-mTOR (Ser2448 # 2971), S6K (# 2708), phospho-S6K (Thr389 # 9205), 4E-BP1 (rabbit mAb 9644), phospho-4E-BP1 (rabbit mAb 9451), and β-actin (# 4970) were all rabbit monoclonal antibodies purchased from Cell Signalling Technology (USA). Horseradish peroxidase-conjugated goat anti-rabbit IgG (7074, CST) was used to detect proteins, and a Chemi-Lumi One Super kit (Nacalai Tesque Inc., Japan) was employed to visualize the bands. β-Actin expression was measured as a loading control.

Statistical analysis

The results were analysed using the JMP version 17 software (JMP, USA). A one-way ANOVA was used where appropriate. Data are presented as mean ± SEM. Differences among groups were tested using ANOVA. If an overall significant difference was found, a post hoc Tukey’s test was performed for multiple comparisons. Statistical significance was set at P < 0.05.

Results

Rapamycin and cyclophosphamide treatment do not induce obvious adverse effects

Treatment with 5 mg/kg/day rapamycin for 29 days with or without three cycles of 120 mg/kg/week cyclophosphamide did not elicit obvious abnormalities at necropsy. No significant weight losses were observed among the mouse groups (Fig. 3).

Figure 3.

Final body weights of the mice. There were no statistically significant differences in body weight in any group, despite repeated cyclophosphamide (Cy) and rapamycin (Rap) administration over a long period. Data represent the mean ± SEM values.

Open in new tabDownload slide

Antitumour effects of cyclophosphamide and rapamycin in mouse xenograft breast cancer model

HE-stained tumours showed no necrosis in the control group, whereas extensive necrosis was observed in the Cy and Cy + Rap groups (Fig. 4). During the first week of treatment, the vehicle control and Cy groups, both administered the placebo, exhibited similar tumour volume indices (mean: 176% and 161%, respectively). In contrast, the Cy + Rap group, administered daily rapamycin, had a tumour volume index of 125%. After the second week of treatment (4 weeks after cell transplantation), the tumour volume index for the Cy group (weekly cyclophosphamide monotherapy) and Cy + Rap group (cyclophosphamide weekly + rapamycin daily) stabilized at a mean of 160–197% and 130%, respectively, until the end of treatment. In contrast, tumours in the control group grew steadily, with a mean tumour volume index of 578% 4 weeks after treatment initiation. Statistically significant differences in the final tumour volume index were observed between the vehicle control and both treatment groups (Fig. 5).

Figure 4.

Haematoxylin and eosin (HE) staining of tumours. Typical HE staining of the tumours in each group is shown. Bar = 100 μm. Cy, cyclophosphamide; Rap, rapamycin.

Open in new tabDownload slide

Figure 5.

Antitumour effects of cyclophosphamide and rapamycin in mouse xenograft breast cancer model. Subcutaneous transplantation tumour model in nude BALB/c mice was established using the human breast cancer cell line MDA-MB-231. The mice were divided into control, Cy, and Cy + Rap groups with 6 or 7 mice per group. Treatment protocol schema and the growth curve of normalized tumour volume are shown: *P < 0.001 compared with the control. All data represent the mean ± SEM values. Cy, cyclophosphamide; Rap, rapamycin.

Open in new tabDownload slide

Western blotting revealed that phospho mTOR (pmTOR) and pS6K were expressed in the control group; high mTOR expression in human clinical triple-negative breast cancer was reproduced in our in vivo xenograft model. Phosphorylation of mTOR and S6K was inhibited more significantly by rapamycin treatment than in the control or Cy groups (Fig. 6). Moreover, the Cy + Rap group exhibited markedly fewer Ki-67-positively stained cells than the Cy and control groups, indicating decreased tumour cell proliferation (P < 0.01; Fig. 7).

Figure 6.

Rapamycin inhibits mTOR pathway in mouse xenograft breast cancer model. Representative western blots from tumour xenograft tissue are shown. Phosphorylation of mTOR and S6K was inhibited by rapamycin treatment (Cy + Rap) than in the control or Cy groups. Cy, cyclophosphamide;; pmTOR, phospho mTOR; pS6K, phospho S6K; Rap, rapamycin; S6K, S6 kinase.

Open in new tabDownload slide

Figure 7.

Inhibitory effect of rapamycin and cyclophosphamide on tumour cell proliferative activity. The tumour proliferation inhibitory effect of rapamycin was evaluated using Ki-67 immunostaining. Brown indicates Ki-67 stain-positive cells, while blue indicates Ki-67 stain-negative cells. The Cy + Rap group had a significantly lower Ki-67 positive cell rate than the Cy and control groups. P < 0.01, bar = 100 μm. The data represent the mean ± SEM values. Cy, cyclophosphamide; Rap, rapamycin.

Open in new tabDownload slide

Short-term treatment with cyclophosphamide does not induce apoptosis in primordial follicles but can induce apoptosis in growing follicles and this effect is inhibited by rapamycin

Based on the TUNEL staining of ovaries resected 24 h after short-term cyclophosphamide administration, no positively stained primordial follicles were detected in the control, Cy, or Cy + Rap groups. Conversely, numerous granulosa cells within growing follicles were TUNEL positive in the Cy group but relatively negative in the control and CY + Rap groups (Fig. 8).

Figure 8.

Terminal transferase dUTP nick end labelling (TUNEL) staining of ovaries resected 24 h after cyclophosphamide administration. Brown indicates TUNEL stain-positive cells, while blue indicates TUNEL stain-negative cells. No TUNEL stain-positive primordial follicles can be observed in the control, Cy, and Cy + Rap groups. All primordial follicles in the control, Cy, and Cy + Rap groups were negative for TUNEL staining (indicated by red circles). Conversely, numerous granulosa cells of growing follicles were TUNEL positive in the Cy group and relatively negative in the control and Cy + Rap groups. Scale bar = 100 μm. Cy, cyclophosphamide; Rap, rapamycin.

Open in new tabDownload slide

Protective effect of rapamycin as an mTOR inhibitor on the in vivo model of long-term cyclophosphamide-induced ovarian follicle loss

Low-magnification histological analysis found no obvious morphological differences in the stroma of the control group and the Cy and Cy + Rap groups (Fig. 9). We also investigated the effects of the mTOR inhibitor (rapamycin) and cyclophosphamide on the number of follicles in each group. The ovarian tissue of the Cy group had significantly fewer primordial follicles (P < 0.001) than the control group. Meanwhile, the ovarian tissue of the Cy + Rap group had significantly more (∼2.5 times) primordial follicles (P < 0.001) than the Cy group. That is, the Cy + Rap group had 62% of the primordial follicles observed in the control group, whereas the Cy group had 26%. This indicates that the addition of rapamycin can inhibit cyclophosphamide-induced primordial follicle loss.

Figure 9.

Haematoxylin and eosin (HE) staining histological examination of ovaries. Low-magnification HE staining revealed that the Cy and Cy + Rap groups exhibited no destruction of the ovarian structure, such as stromal fibrosis despite cyclic cyclophosphamide administration. the Cy + Rap group showed a reduced growing follicle count (e.g. secondary and antral follicles); scale bar = 100 μm. Cy, cyclophosphamide; Rap, rapamycin.

Open in new tabDownload slide

There was also a statistically significant difference in the number of primordial follicles between the control and Cy + Rap groups (P < 0.001). The Cy + Rap group had fewer growing (secondary and antral) follicles than the control groups. Meanwhile, the atretic follicle count did not differ significantly among all groups. The Cy + Rap group also had significantly smaller ovaries than the control group (P < 0.05), indicating that there were fewer secondary and antral follicles, which contribute extensively to the ovary size, in the Cy + Rap group compared with the control group.

The ratio of primary to primordial follicles in the Cy group was twice that of the control group. Meanwhile, the addition of rapamycin (the Cy + Rap group) reduced this ratio to a level that did not differ significantly from the control group (Fig. 10). This indicates that rapamycin can prevent cyclophosphamide from stimulating excessive differentiation of primordial follicles into primary follicles.

Figure 10.

Follicle counts. (A) Primordial follicle count: the Cy group had significantly fewer primordial follicles than the control group. The Cy + Rap group had significantly more primordial follicles than the Cy group. Cy, cyclophosphamide; Rap, rapamycin. (B) Primary follicle count: the Cy and Cy + Rap groups had significantly fewer primary follicles than the control group. (C) Secondary follicle count: the Cy + Rap group had significantly fewer secondary follicles than the control and Cy groups. (D) Antral follicle count: the Cy + Rap group had significantly fewer antral follicles than the control group. (E) Atretic follicle count: No statistically significant differences were observed. (F) Minor axis index of ovary: the minor axis index of the ovary was defined as the number of sections in which ovarian tissue was observed in consecutive sections. The Cy + Rap group had significantly lower minor axis index of ovary than the control group. (G) Ratio of primary to primordial follicles: ovaries of Cy-treated mice had over twice the ratio of primary to primordial follicles compared with all other treatment groups. Ovaries of mice cotreated with Cy + Rap had ratios of primary to primordial follicles matching untreated controls. *P < 0.05, **P < 0.001. All data represent the mean ± SEM values.

Open in new tabDownload slide

Western blotting showed that mTOR, S6K, and 4E-BP1 (located downstream of the mTOR pathway) were activated by cyclophosphamide and inhibited by rapamycin (Fig. 11). Thus, rapamycin is supplied to the ovary and elicits an inhibitory effect against mTOR. Moreover, the pS6K-positive primordial follicle rate in the Cy group was 2.7 times higher than that in the control group. This effect was suppressed in the Cy + Rap group to a level that did not differ significantly from that of the control group (Fig. 12). Hence, cyclophosphamide locally activates the mTOR pathway, particularly in the primordial follicle rather than in the ovarian medulla. Meanwhile, rapamycin inhibits cyclophosphamide-induced primordial follicle activation via mTOR inhibition.

Figure 11.

Cyclophosphamide upregulates and rapamycin downregulates downstream targets of the mTOR pathway in the ovary. Whole ovaries from mice treated with placebo or cyclophosphamide (Cy) or cyclophosphamide + rapamycin (Cy + Rap) were lysed for immunoblot. Representative results from western blotting are shown. In the Cy group, mTOR and downstream targets of the mTOR pathway (S6K and 4E-BP1) were more phosphorylated than control. In the Cy + Rap group, mTOR and downstream targets of the mTOR pathway were less phosphorylated than in the Cy group. S6K, S6 kinase.

Open in new tabDownload slide

Figure 12.

Representative results from phosphorylated (p)S6K staining. The control and Cy + Rap groups had many primordial follicles, most were pS6K negative. In contrast, there were fewer primordial follicles in the Cy group, but relatively more primordial follicles were pS6K positive. P < 0.001. n = 4–5/group. Bar = 100 μm. The data represent the mean ± SEM values. Cy, cyclophosphamide; Rap, rapamycin; S6K, S6 kinase.

Open in new tabDownload slide

Discussion

In this study, we realized an in vivo murine xenograft breast cancer model. Breast cancer is classified into four subtypes according to the positivity or negativity of hormone receptors (oestrogen and progesterone receptors) and HER2. In this study, triple-negative breast cancer (hormone receptor negative, HER2 negative) was selected as the subtype because, although triple-negative breast cancer is a minor subtype among all breast cancer cohorts, it is more common in younger breast cancer (Colleoni et al., 2002; Futamura and Yoshida, 2022). Moreover, despite the fact that patients with triple-negative breast cancer have a high recurrence rate compared with non-triple-negative breast cancer (Costa et al., 2018), specific endocrine therapies and targeted therapies are ineffective. Therefore, cytotoxic chemotherapy, primarily cyclophosphamide-based multi-combination chemotherapy, has become the main strategy for treating triple-negative breast cancer (Yin et al., 2020). Additionally, in oestrogen receptor-positive breast cancer, if ovarian function is preserved by ovarian protection agents, blood oestradiol levels may increase and cause a tumour-enhancing effect on the transplanted breast cancer. However, in triple-negative breast cancer, ovarian protection with the preservation of ovarian function and elevated blood oestradiol levels are unlikely to cause tumour progression.

We also used immunodeficient mice to reproduce ovarian dysfunction in a carcinoma-bearing state. BALB/c nude mice were selected as they are only T-cell deficient and near-normal immunodeficient, compared with new strains of immunodeficient or severe combined immunodeficiency mice. Notably, cellular immunodeficiency (i.e. T-cell immunodeficiency) is associated with human breast cancer (Hadden, 2003). Given that the prognosis of breast cancer does not differ considerably between preoperative and postoperative chemotherapy (Early Breast Cancer Trialists’ Collaborative Group, 2018), the incidence of preoperative chemotherapy, particularly involving cyclophosphamide, is steadily increasing. Hence, our study, conducted in a xenograft mouse model, mirrors this evolving clinical scenario.

In the present study, doses and treatment cycles were selected based on those used in human clinical settings and mouse experimental models. In humans, a minimum of 9 months is required from the primordial follicle to ovulation (Zhou et al., 2017). Cyclophosphamide, typically prescribed for neoadjuvant and adjuvant chemotherapy, is administered in a tri-weekly cyclic schedule, often spanning 4–6 cycles, all within a minimum of 9-month timeframe. In this regard, we considered a single dose inappropriate for replicating the human clinical model and opted for cyclic dosing. Cyclophosphamide was administered weekly, considering that the time between primordial follicle formation and ovulation was 3 weeks in mice (Zhou et al., 2017).

Follicle activation occurs via phosphoinositide 3-kinase (PI3K)/AKT/mTOR pathway activation, resulting in the accumulation of numerous growing follicles, many of which undergo apoptosis through the PI3K/AKT/mTOR pathway (Kalich-Philosoph et al., 2013; Qin et al., 2022). The activation of AKT/mTOR by cyclophosphamide and the FOXO3a signalling network leads to FOXO3a export from the nucleus into the cytoplasm, promoting cellular proliferation (Liu et al., 2007; Jang et al., 2016). Moreover, γH2AX becomes elevated when the AKT/mTOR pathway is activated, whereas high intracellular levels of AKT/mTOR reportedly increase DNA damage, resulting in apoptosis (Bellusci et al., 2019). The present study demonstrates that rapamycin exerts a protective effect against cyclophosphamide-induced primordial follicle loss. The mechanism of cyclophosphamide-induced follicle loss was found to be due to excessive primordial follicle development (so-called ‘burn out’), whereas rapamycin inhibits this effect, as evidenced by the ratio of primary to primordial follicles. Western blotting and immunostaining results further revealed that cyclophosphamide activated the mTOR pathway locally in primordial follicles, whereas rapamycin inhibited this effect. However, TUNEL staining 24 h after cyclophosphamide administration showed that primordial follicles do not directly undergo apoptosis, but rather granulosa cells of the growing follicles; this is inhibited by rapamycin (Fig. 13). Meanwhile, the smaller ovarian size in the Cy + Rap group may primarily reflect the fewer growing follicles, which comprise a larger proportion of the ovarian volume. Similarly, a previous study on the preventive effect of rapamycin on age-related follicle loss found more primordial follicles and fewer growing follicles, resulting in smaller ovaries in the rapamycin group compared with the control group (Dou et al., 2017). In the Cy + Rap group, the number of growing follicles was low despite the prevention of apoptosis in growing follicles. This finding may be due to the rapamycin-mediated prevention of primordial follicle activation outweighing the rapamycin-mediated prevention of apoptosis in the growing follicle. Notably, activation of the mTOR pathway continued 7 days after cyclophosphamide administration. In mice, 3 weeks pass between the primordial follicle development and ovulation. In our study, the cyclophosphamide-induced primordial follicle activation persisted for more than one-third of the ovulation cycle, suggesting that mTOR inhibitors should be taken orally for an extended period after cyclophosphamide administration. Moreover, the residual toxicity from cyclophosphamide 1 week after the weekly regimen suggests an accumulation of cyclophosphamide ovarian toxicity.

Figure 13.

Effects of cyclophosphamide and rapamycin on breast cancer and ovary. (A) Normal physiology; the primordial follicle is a finite resource and is activated into a primary follicle by balanced regulation. The growing follicle has natural survival and does not undergo apoptosis. (B) When cyclophosphamide is administered to breast cancer patients, the primordial follicle is over-activated through the mTOR pathway, resulting in the apoptosis of many growing follicles and loss of the primary follicle. In breast cancer, activation of mTOR induces tumour progression through cell proliferation. Cyclophosphamide, as an alkylating anticancer drug, exerts a strong antitumour effect on breast cancer through the mechanism of DNA damage. (C) Cyclophosphamide + rapamycin cotreatment prevents loss of primordial follicles via rapamycin-induced mTOR pathway blocking. Rapamycin has the potential to inhibit cell proliferation by blocking mTOR activation in breast cancer.

Open in new tabDownload slide

Whether cyclophosphamide directly induces primordial follicle apoptosis is debateable. Although some authors claim the direct loss of primordial follicles following cyclophosphamide administration (Ganesan and Keating, 2016; Luan et al., 2019; Nguyen et al., 2019b), most investigators are of the opinion that over-activation of the primordial follicle (‘Burn out’) is the primary mechanism (Kalich-Philosoph et al., 2013; Goldman et al., 2017; Kano et al., 2017; Lande et al., 2017; Roness et al., 2019; Sonigo et al., 2019; Shai et al., 2021). Meanwhile, Kashi et al. (2023) demonstrated that activation of the primordial follicle is largely responsible, with a small amount of direct DNA damage. The discrepancy in results likely stems from the time at which activation and apoptosis of the follicle occur. DNA damage and apoptosis by cyclophosphamide occur in 12–24 h, whereas activation of the primordial follicle is observed 24 h to several days later (Kalich-Philosoph et al., 2013; Goldman et al., 2017; Zhou et al., 2017; Chen et al., 2022a; Kashi et al., 2023). There is also the question of which indexes (e.g. γH2AX, TUNEL, cPARP, or caspases) are most suitable for evaluating the direct damage and apoptosis caused by cyclophosphamide in primordial follicles. In the present study, we evaluated apoptosis by TUNEL staining 24 h after cyclophosphamide administration and found that apoptosis did not occur in primordial follicles. We also evaluated the mTOR pathway and primordial follicle activation 7 days after cyclophosphamide administration and observed that primordial follicle activation occurs via the mTOR pathway. Hence, the results of this study suggest that the loss of primordial follicles by cyclophosphamide is primarily due to activation of primordial follicles via the mTOR pathway.

mTOR inhibitors exert potential antitumour effects against certain subtypes of breast cancer. The level of evidence in human clinical practice is particularly high for hormone-receptor-positive breast cancer. A phase III randomized controlled trial study reported the efficacy and safety of mTOR inhibition with hormone-receptor-positive breast cancer (Piccart et al., 2014). Meanwhile, probable antitumour effects of mTOR inhibitors have also been reported in triple-negative breast cancer. In triple-negative breast cancer, 72% of the tumours harbour phosphorylated mTOR, which is associated with shorter overall- and recurrence-free survival rates for patients with early-stage triple-negative breast cancer (Ueng et al., 2012). Phosphorylation reactions stimulated by the PI3K/AKT/mTOR are responsible for cancer cell growth, proliferation, angiogenesis, tumour metastasis, and invasion (Arcaro and Guerreiro, 2007). Indeed, considerable in vivo and in vitro evidence have shown the antitumour efficacy of mTOR inhibitors as single agents or as sensitizing drugs with standard care drugs, including cyclophosphamide, and numerous studies, including our own, have reported that a main antitumour mechanism of mTOR inhibitors is their inhibitory effect on cell proliferation, as evidenced by employed Ki-67 immunostaining (Akcakanat et al., 2009; Montero et al., 2014; Zeng et al., 2010; Zhang et al., 2012, 2014; Eloy et al., 2016; Wan et al., 2016; Dankó et al., 2021; Tan et al., 2023). In this study, there was no statistically significant difference in tumour volume between the cyclophosphamide group and cyclophosphamide plus mTOR inhibitor group. The efficacy of anticancer drugs varies depending on the genetic profile of the cell line. MDA-MB-231, the cell line employed in this study, is sensitive to rapamycin in vivo (Zhang et al., 2012; Wander et al., 2013) but less sensitive than other triple-negative breast cancer cell lines (Liu et al., 2011; Yunokawa et al., 2012), suggesting that other triple-negative breast cancer cell lines may exhibit a stronger tumour volume suppression effect following rapamycin treatment. Hence, rapamycin, at the minimum, does not weaken the anticancer effect of cyclophosphamide against triple-negative breast cancer and may augment an anti-tumour cell proliferation effect (Fig. 13). However, the lack of statistically significant differences in tumour volume prevented the confirmation that rapamycin elicits an antitumour effect on breast cancer. Prior to the clinical application of mTOR inhibitors for antitumour effects, phase III trials assessing their potential against triple-negative breast cancer are warranted.

Mucositis and stomatitis are the most commonly reported adverse effects of mTOR inhibitors (Nguyen et al., 2019a). In this study, no obvious stomatitis was observed during autopsy, and no weight loss due to mTOR inhibitors was observed, suggesting that feeding difficulties due to stomatitis were negligible.

A limitation of this study is that although rapamycin elicits follicle-protective effects, it cannot fully prevent the follicle loss induced by cyclophosphamide. Nevertheless, the advantages of rapamycin as a follicle protective agent can be exploited in combination with hormone therapy, recombinant anti-Müllerian hormone (Kashi et al., 2023), or a GnRH agonist. Combination therapy with stem cells is also expected to be effective (Takahashi et al., 2021). In combination with other mechanism-based agents, mTOR inhibitors may exhibit more complete follicle protective against high doses of cyclophosphamide.

Additionally, we were unable to examine follicle functionality and fertility in this study. Although two studies have evaluated litter size after cyclophosphamide and mTOR inhibitor co-treatment as ovarian protectants (Goldman et al., 2017; Kashi et al., 2023), both studies evaluated litter size 8–16 weeks after termination of anticancer therapy. If the present experimental model is prolonged and matching is performed for 8 weeks after cyclophosphamide and rapamycin completion, a humane endpoint will likely be reached since the mice will be followed for several months without administration of antitumour drugs in the cancer-bearing state. Thus, new protocols are needed to evaluate follicle functionality and fertility and antitumour effects in a single experimental model. For example, using the breast cancer xenograft mouse model, antitumour effects can be measured after administering rapamycin and cyclophosphamide to mimic neoadjuvant chemotherapy. Tumour resection should be performed at the end of Cy and rapamycin treatment, and with the litter size evaluated after a matching protocol.

In conclusion, this study demonstrated that mTOR inhibitors exert an ovario-protective effect during cyclical cyclophosphamide regimens, which have high antitumour effects as single agents in breast cancer. In terms of antitumour efficacy, although there was no statistically significant difference in tumour volume, our results suggest that mTOR inhibitors could potentially inhibit tumour proliferation.

Data availability

The data in this article are available in the article, and when necessary by request from [email protected].

Acknowledgments

The authors thank Takefumi Yamamoto (Central Research Laboratory of Shiga University of Medical Science) for their excellent technical support. The authors thank Suzuko Moritani (Division of Diagnostic Pathology, Shiga University of Medical Science) for pathological assistance.

Authors’ roles

T.M. conceptualized the study. Y.T. designed the study, performed the experiments, collected and analysed the data, and wrote the article. A.N. assisted in performing the histological immunostaining and western blot assays and in discussing the results. F.Y. prepared the carcinoma-bearing mice. Y.Y. performed statistical analyses. A.I. assisted with all aspects of the experiment. H.Y. and T.A. assisted in the discussion of the results. T.A. and S.T. supervised the study. A.Take., T.H., and A.Taka. edited the manuscript.

Funding

This work was not financially supported.

Conflict of interest

The authors declare that they have no conflict of interest.

References