CHEK2 signaling is the key regulator of oocyte survival after chemotherapy

Genetic inactivation of CHEK2 prevents oocyte elimination caused by chemotherapy drugs in ex vivo organ culture

Previous studies showed that genetic inactivation of CHEK2 prevents POI in mice after exposure to radiation (19,26). Studies using the CHEK inhibitor BML-277 showed a protective effect against the chemotherapy drugs CDDP, DOX, and CTX derivative 4-HC (22, 24, 25). However, it remains unclear whether this protective effect is due to specific inhibition of CHEK2, CHEK1, or both kinases due to limited selectivity. To test the direct role of CHEK2 in the elimination of PMFs damaged by genotoxic chemotherapy drugs and determine whether CHEK2 inhibition would be effective and sufficient in reducing PMF loss, we tested whether genetic ablation of CHEK2 function prevents primordial oocyte elimination after treatments. We chose chemotherapy drugs known to induce DNA damage via different mechanisms and characterized by different severities of ovarian adverse effects: DNA alkylating agents CDDP and MAFO, and DNA topoisomerase II poisons DOX and ETO (27). We used MAFO, a preactivated CTX derivative that can be used in ex vivo culture to imitate the in vivo activity of CTX, which needs to be metabolized by the liver to activate its alkylating properties. However, both drugs have been shown to generate similar metabolites (28). To better control for effective drug doses and duration of exposure for comparative analysis, we utilized an ex vivo organ culture system (Fig. 1A), where we cultured whole ovaries to prevent potential PMF activation due to fragmentation of ovarian tissue (29). We used ovaries from 1-week-old females, the stage at which mouse ovaries predominantly contain fully formed PMFs and a small population of primary/secondary follicles. This approach also facilitates investigation of direct toxicity in primordial oocytes without the potential confounding effects from damage to large growing follicles seen in postpubertal ovaries. Ovaries were cultured for 5 days after completion of 48 hours of drug treatment, or a total of 1 week, to ensure observation of long-term primordial oocyte survival rather than delayed apoptosis. To determine doses that deplete primordial oocytes in mice, we exposed ovarian explants from wild-type females to increasing doses of MAFO (0.1 to 1 μg/ml), CDDP (0.1 to 0.5 μg/ml), DOX (0.025 to 0.1 μg/ml), and ETO (0.05 to 0.5 μg/ml) (fig. S1). Doses used in this ex vivo study are lower than the reported plasma concentrations in patients at the highest single recommended dose [CDDP: 4.3 μg/ml, CTX: 33.4 μg/ml, DOX: 3.66 μg/ml, and ETO: 19.66 μg/ml (30, 31)], yet they are sufficient to significantly deplete primordial oocytes in mice. PMFs lack direct contact with blood vessels; drugs reach the oocytes through diffusion, which may explain the toxicity at lower doses in organ culture. The number of primordial oocytes was reduced compared to vehicle controls after exposure to all tested drugs in a dose-dependent manner (fig. S1). Significantly reduced survival of primordial oocytes was observed in wild-type ovaries for MAFO at 1 μg/ml (7.3% ± 9.0%; P < 0.0001), CDDP at 0.5 μg/ml (0.04% ± 0.08%; P < 0.0001), and DOX at 0.1 μg/ml (2.4% ± 0.6%; P < 0.0001), and to lesser extent for ETO at 0.5 μg/ml (39.0% ± 13.7%; P = 0.0021) (Fig. 1, B and D). In agreement with published data, ETO showed lowest toxicity in juvenile ovaries, suggesting mechanistic differences in activity compared to the other topoisomerase II poison tested (DOX). When Chek2^−/− ovaries were treated with the same doses of drugs, no significant reduction in primordial oocyte numbers was observed (Fig. 1, C and D). We calculated oocyte survival as a percentage of ovarian reserve in untreated wild-type and Chek2^−/− females. When compared to wild-type controls, survival of primordial oocytes lacking CHEK2 was improved after MAFO (7.3% versus 107.3% ± 23.7%; P = 0.0002), CDDP (0.04% versus 83.2% ± 42.32%; P < 0.0001), DOX (2.4% versus 92.9% ± 24%; P < 0.0001), and ETO (39% versus 120.7% ± 20.3%; P = 0.0006).

Because there have been reports that CDDP and CTX induce hyperactivation of PMFs and their transition to growing follicles (32–34), we counted the larger growing oocytes (diameter > 30 μm), which are typically found in primary and secondary follicles, in 2-week-old ovaries treated with each of the four drugs. We observed no significant increase in growing oocyte numbers: To the contrary, in some cases (CDDP and DOX), we observed a reduction in the number of growing oocytes after drug treatments (Fig. 1E and fig. S1). This indicates that follicle hyperactivation is not the major mechanism of PMF loss in prepubertal ovaries or in ex vivo explants and is in agreement with other studies (25, 35, 36). In summary, survival of primordial oocytes in Chek2^−/− ex vivo–treated ovaries demonstrates that CHEK2 is directly responsible for coordinating elimination of PMFs after treatment with four different chemotherapy drugs and that CHEK2 inhibition would likely be sufficient to prevent loss of PMF reserve in cancer patients treated with these drugs.

CHEK2 deficiency prevents PMF loss after in vivo treatment with an acute dose of alkylating chemotherapy drugs

To confirm that abrogation of CHEK2 activity is sufficient to mitigate chemotherapy-induced PMF loss in prepubertal mice treated in vivo with highly ovotoxic alkylating agents CTX and CDDP (37), 1-week-old control (Chek2^+/−) and Chek2^−/− females were treated with vehicle or an acute dose of CTX (150 mg/kg) or CDDP (5 mg/kg). Female weights were recorded before treatment and then weekly after. Ovaries were collected 2 weeks after treatment (at 3 weeks of age) for histological analyses of the PMF reserve. CTX treatment obliterated oocytes in PMFs in control females where follicle remnants were often observed (3.1% ± 2.2%; P < 0.0001) (Fig. 2, A and C). In contrast, abundant PMFs with oocytes were present in ovaries from CTX-treated Chek2^−/− females (91% ± 40.8% versus 3.1%; P < 0.0001) (Fig. 2, B and C). CTX treatment caused additional side effects in pups including hair loss and growth retardation. Improved weight gain was evident in treated Chek2^−/− pups compared to controls (60% ± 14.3% versus 10% ± 7.2%) (Fig. 2E), suggesting that inhibition of CHEK2 activity could potentially alleviate other adverse side effects caused by CTX, although hair loss was not prevented. CDDP treatment significantly decreased PMF numbers in ovaries from control females (38.7% ± 26.5%; P = 0.0002), but to a lesser extent than CTX (Fig. 2, A and D). Compared to treated controls, Chek2^−/− females showed abundant PMFs (126% ± 59.6% versus 38.7%; P = 0.0009) as well as growing follicles (Fig. 2, B and D). Overall, CTX- and CDDP-treated Chek2^−/− females retained >90% of their ovarian reserve and showed healthy primary and secondary follicles in the ovary at 3 weeks of age and thus were expected to be fertile. Here, we focused on PMF survival rather than fertility outcomes, but other studies indicate that oocytes that survive genotoxic treatments can produce healthy offspring (23, 24). Together, these results indicate that CHEK2 is predominantly if not exclusively responsible for triggering primordial oocyte elimination and that targeting CHEK2 activity in vivo will be sufficient to protect PMFs against CDDP- and CTX-induced damage.

CDDP and MAFO treatments cause abundant DNA DSBs in oocytes, which exhibit markers for homologous recombination but not nonhomologous end joining repair

CHEK2 has been shown to coordinate elimination of oocytes in response to unrepaired, programmed meiotic and radiation-induced DSBs (19). To examine whether the four tested drugs cause DNA DSBs in oocytes, which would activate CHEK2 signaling, we immunostained vehicle- and drug-treated wild-type ovaries for general DNA DSB marker γ-H2AX. After 24 hours of treatment, abundant γ-H2AX foci were present in all primordial oocytes in ovaries treated with MAFO and CDDP. However, in ovaries treated with DOX or ETO, elevated levels of γ-H2AX foci were rarely observed in oocytes (Fig. 3, A and D). DNA damage markers were present in granulosa and other ovarian cell types after all four treatments, indicating that drug doses used in this study induce DNA damage (Fig. 3A). Because DSBs can be repaired by high-fidelity homologous recombination (HR) or more error-prone nonhomologous end joining (NHEJ), we immunostained treated ovaries for HR marker RAD51 and NHEJ marker 53BP1.

RAD51 foci were present in oocytes treated with MAFO or CDDP, indicating that these drugs activate the HR repair pathway (Fig. 3, B and E). RAD51 foci were rarely observed in DOX- and ETO-treated oocytes, consistent with the γ-H2AX staining results, while RAD51 foci were found in somatic cells (Fig. 3, B and E). Oocytes in postnatal ovaries are arrested at dictyate stage of meiotic prophase I, comparable to mitotic G2 phase. In this phase, HR is the predominant repair pathway due to the presence of sister chromatids, but NHEJ repair is also thought to be active. A previous report shows that DNA-PKcs, a component of NHEJ repair, is detected in 10% of primordial oocytes after radiation-induced DNA damage (38). Another study reported DNA-PKcs activation in 30 to 60% of oocytes after CTX treatment (39). 53BP1 directs DSB repair toward NHEJ and localizes to DSB sites forming discrete foci that colocalize with γ-H2AX (40). Remarkably, there was no evidence for 53BP1 foci in oocytes treated with any of our tested chemotherapy drugs (Fig. 3C). However, they were readily detected in somatic cells in treated ovaries (Fig. 3C). This indicates that DNA damage in oocytes is predominantly, if not exclusively, repaired by high-fidelity HR, which has a lower risk of generating deleterious mutations in surviving oocytes, as was shown for radiation (38). Lack of DNA damage markers in primordial oocytes within 24 hours of DOX and ETO treatments suggests that DSBs may not be the primary type of damage caused by these drugs that leads to CHEK2 activation and oocyte apoptosis. It is possible that DSB formation after DOX treatment is delayed or caused indirectly by oxidative stress as γ-H2AX–positive primordial oocytes were previously reported in human and mouse ovaries after DOX treatment (41). Additionally, differences in accumulation of DNA damage in oocytes could be due to interspecies differences in drug pharmacokinetics and metabolism.

To determine whether CHEK2 is activated in oocytes by chemotherapy drugs, we immunostained drug-treated ovaries for CHEK2 phosphorylated at threonine-68 (42). Of note, we found that multiple commercially available antibodies either do not detect CHEK2 in mouse paraffin-fixed ovaries by standard methods or showed unspecific staining also present in CHEK2-deficient samples. However, we identified one specific antibody against pCHEK2 T68 that shows expected nuclear signal in wild-type but not in Chek2^−/− ovaries treated with radiation. Activated pCHEK2 was readily detected in primordial oocytes 3 hours after radiation and was still present in oocytes 24 hours after radiation (Fig. 4). MAFO, CDDP, and DOX treatment showed delayed CHEK2 activation compared to radiation; pCHEK2 was not detected in primordial oocytes treated for 3 to 6 hours. Primordial oocytes positive for pCHEK2 were present in ovaries treated for 24 hours, but not as abundant as those damaged by radiation. CHEK2 activation was rarely detected in ETO-treated oocytes. This delayed and asynchronous pattern of CHEK2 activation most likely reflects a more complex mechanism of drug toxicity for alkylating agents compared to instant physical damage caused by radiation. Why CHEK2 activation, measured as ATM-dependent T68 phosphorylation (43), is not readily detected in all primordial oocytes after treatment with MAFO and CDDP—although they are positive for DNA damage markers (Fig. 3, A and B)—remains unclear and will need additional studies or better reagents. Nevertheless, our genetic data clearly demonstrate that CHEK2 is involved in elimination of primordial oocytes after chemotherapy treatments.

Together, we show that CDDP and MAFO induce lethal DSBs in prophase I–arrested primordial oocytes and in dividing ovarian somatic cells. Repair of DSBs in oocytes is initiated by high-fidelity HR but cannot be completed due to activation of proapoptotic responses coordinated by CHEK2. Moreover, our results confirm previous reports that topoisomerase II poisons such as DOX and ETO do not induce abundant DSBs in oocytes at least within 24 hours of treatment (44), but rather in ovarian somatic cells that are actively proliferating. Nevertheless, primordial oocyte elimination is still observed in CHEK2-proficient ovaries and increased survival in CHEK2-deficient ovaries, indicating that DOX and ETO affect primordial oocyte survival by a different mechanism, potentially related to oxidative stress. Additional studies are needed to identify how cellular damage caused by these drugs leads to CHEK2-dependent oocyte elimination. In summary, our results indicate that inhibition of CHEK2 activity in chemotherapy-treated ovary prevents elimination of primordial oocytes with and without DNA DSBs by providing more time for repair of cellular damage and allowing oocyte survival.

Damage induced by CDDP and MAFO does not induce hyperphosphorylation of TAp63 and triggers activation of p53-dependent apoptosis

In response to radiation-induced damage, CHEK2 phosphorylates two proapoptotic factors: TAp63 specifically in oocytes and p53 in all cell types (19, 22, 45). Unlike p53, which requires phosphorylation to avoid degradation (45), TAp63 is constitutively expressed in primordial oocytes and maintained in an inactive form until phosphorylated by CHEK2 (19, 22). Phosphorylated TAp63 is detected in irradiated ovaries as mobility shift on Western blots (Fig. 5A and fig. S2C) (22, 46). Reports suggest that inactivation of TAp63 is sufficient to protect primordial oocytes from CDDP-induced damage (23, 24), but may not be sufficient for CTX as two studies report contradictory results (23, 47). Because MAFO and CDDP induce DSBs and trigger CHEK2-dependent apoptosis, phosphorylation and activation of TAp63 are expected. However, TAp63 shift was not detected up to 24 hours of MAFO or CDDP treatment, suggesting that majority of primordial oocytes failed to activate TAp63 in this timeframe (Fig. 5A). These results suggest that damage induced by MAFO (2.4 μM) and CDDP (1.6 μM) does not lead to rapid phosphorylation of TAp63 as observed after radiation where instant physical damage induces DDR. Activation of CHEK2 by CDDP and MAFO was also delayed compared to radiation; therefore, it is possible that DNA damage caused by alkylating agents triggers activation of CHEK2 and subsequently TAp63 with different dynamics (Fig. 4). This is supported by reports showing TAp63 phosphorylation after ex vivo treatments with a higher dose of CDDP (10 μM) (22, 24) or after CTX treatment in vivo (39, 47).

To determine whether TAp63 is involved in MAFO- and CDDP-induced primordial oocyte elimination downstream of CHEK2, we utilized a novel mouse model with a mutation in the Trp63 gene (fig. S2A). Here, serine 621 was replaced with alanine (S621>A, S582>A in human) at the TAp63 C terminus, which was previously shown to be phosphorylated by CHEK2 (19, 22). This mutation prevents activation by CHEK2 but does not affect TAp63 expression (fig. S2, C and D) and results in increased resistance of oocytes to low dose of radiation (0.5 Gy) (fig. S2D). TAp63A/A mutant females are fertile with normal ovarian histology (fig. S2B). We exposed TAp63A/A mutant ovarian explants ex vivo to MAFO and CDDP at the same doses and regimen as wild-type and Chek2^−/− ovaries (Fig. 5, B and E, and Fig. 1). After treatment with MAFO and CDDP, primordial oocyte numbers were significantly reduced in TAp63A/A ovaries (Fig. 5, B and D). Primordial oocyte survival after MAFO treatment was lower in TAp63A/A ovaries at ~21.9% ± 12.6% (P < 0.0001) versus ~107.3% in Chek2^−/− (P < 0.0001), and after CDDP treatment at ~12.7% ± 13% (P < 0.0001) versus ~83.2% (P < 0.0001) (Fig. 5E). The reduced number of surviving primordial oocytes in TAp63A/A mutants suggests that other CHEK2 target/s contribute to primordial oocyte elimination after MAFO- and CDDP-induced damage. We detected phospho-p53(S15) in treated ovaries and specifically in primordial oocytes by immunostaining (Fig. 5C). Phospho-p53 was also detected in primordial oocytes after treatment with CTX in vivo (39). To test whether p53 contributes to primordial oocyte elimination after MAFO- and CDDP-induced damage, we analyzed oocyte survival in double mutant ovaries lacking activity of both TAp63 and p53 (48). Compared to Trp63A/A single mutant, primordial oocyte survival was significantly improved in Trp63A/A Trp53^−/− double mutant ovaries after MAFO (84.4% ± 21% versus 21.9%; P = 0.002) and CDDP treatment (99.4% ± 36.4% versus 12.7%; P = 0.001) (Fig. 5, B and E). This confirms that p53 is activated in oocytes after damage with alkylating agents and that it contributes to primordial oocyte elimination. CDDP induces higher number of DSBs than MAFO in ovarian explants (Fig. 3D); therefore, it is possible that p53 activation in oocytes depends on the amount of cellular damage induced by chemotherapy drugs. Together, genetic analysis reveals the critical role for CHEK2 in coordinating PMF elimination after MAFO- and CDDP-induced damage by activating two proapoptotic factors TAp63 and p53 (Fig. 5F). Moreover, loss of CHEK2 or p53 in ovarian somatic cells is expected to prevent apoptosis in ovarian somatic cells and thus indirectly contribute to primordial oocyte survival.

Transient pharmacological inhibition of CHEK2 improves primordial oocyte survival after radiation, CDDP, and MAFO treatments

We showed that genetic ablation of CHEK2 activity prevents depletion of PMF reserve in mice treated with chemotherapy drugs (this study) and radiation (19). Because of similarities in structure and function between CHEK1 and CHEK2, many available checkpoint kinase inhibitors have limited selectivity and can block the activity of both kinases, albeit with different affinities (49). CHEK1 is an essential kinase that coordinates DDR and cell cycle progression in all dividing cells and shares many downstream targets with CHEK2 (50). CHEK1/2 inhibitors have been shown to potentiate the effects of genotoxic chemotherapy drugs against cancer and some are being tested in clinical trials (51–53). Therefore, inhibitors blocking CHEK2 and CHEK1 could have dual benefits for female patients: improved cancer cell elimination and oocyte protection. To test the principle that CHEK1/2 inhibitors could prevent oocyte depletion after radiation and chemotherapy treatments, we selected four inhibitors shown or predicted to target CHEK2 and to some degree CHEK1, and tested them in ovarian explant culture: AZD7762 (54), CCT241533 (55), LY2606368 (56), and PF477736 (57). The pharmacokinetics of these inhibitors in mouse is unknown. Therefore, we used our ex vivo ovarian explant culture system to better control the timing and dosing of drugs and inhibitors without the confounding influence of transport and metabolism found in vivo. We treated ovarian explants with increasing doses of inhibitors in combination with low-dose ionizing radiation 0.5 Gy known to deplete primordial oocytes as described previously (26). Ovarian explants were precultured with inhibitors for 2 hours before radiation to assure cellular presence of the inhibitor before DNA damage and for additional 24 hours after radiation. Primordial oocyte numbers were counted in immunostained whole ovaries after an additional six days of culture without inhibitor to ensure quantification of oocyte survival and not a delay in apoptosis. Treatment with CCT241533 (0.05, 0.5, and 5 μM), LY2606368 (0.1, 1, and 10 μM), and PF477736 (0.5 and 5 μμ) had no impact on primordial oocyte survival after radiation at any dose tested (Fig. 6), indicating failure to sufficiently block CHEK2 activity in oocytes although similar doses have been used in cell culture with cancer cell lines.

In contrast, AZD7762 treatment improved primordial oocyte survival compared to vehicle-treated irradiated ovaries in a dose-dependent manner, although growth of the ovarian explants was retarded (Fig. 7A). We calculated survival as a percentage of oocyte reserve in nonirradiated ovaries. Survival was estimated at 57.3% ± 36.2% after 10 μM AZD7762 (P = 0.001) and 45.6% ± 20.6% (P = 0.008) after 1 μM AZD7762 versus 10.1% ± 4.4% with vehicle. To test whether AZD7762 could also protect primordial oocytes against chemotherapy drugs, ovaries were cotreated with AZD7762 (0.1, 1, and 10 μM) and MAFO or CDDP (Fig. 7A). More primordial oocytes survived chemotherapy-induced damage after cotreatment with AZD7762 (Fig. 7A); 59.6% ± 32.4% versus 4.5% ± 2.7% compared to MAFO alone (P = 0.0003) and 63.1% ± 42.5% versus 3.9% ± 2.4% (P = 0.0006) in CDDP alone. To validate that AZD7762 indeed prevents primordial oocyte death by inhibition of CHEK2-dependent signaling, we tested activation of TAp63 and p53 by Western blot. As expected, TAp63 was phosphorylated after radiation, which resulted in mobility shift and p53 was readily detected (Fig. 7C). In ovaries treated with AZD7762, TAp63 remained unphosphorylated and p53 was reduced (Fig. 7C), which confirms that AZD7762 efficiently inhibited CHEK2 activity in oocytes and prevented TAp63 activation and oocyte apoptosis. Because of lack of detectable TAp63 mobility shift after MAFO and CDDP treatments (Fig. 5A), activation of TAp63 was not tested for these drugs.

AZD7762 blocked activation of CHEK2 during drug treatment and improved oocyte survival. This suggests that inhibition of the CHEK2 pathway resulted in sufficient DNA repair to evade apoptosis even after inhibitor withdrawal. To determine whether AZD7762 inhibition allowed DSB repair, we analyzed γ-H2AX staining in ovaries 24 hours after treatment with MAFO or CDDP and 5 days after drug and inhibitor withdrawal (7 days total). Although γ-H2AX was still present in surviving primordial oocytes, the number of foci was significantly lower at the end of culture for MAFO 11.3 ± 3.8 versus 15.3 ± 5.4 (P = 0.0001) and CDDP 16.3 ± 6.4 versus 30.24 ± 6.4 (P < 0.0001) (Fig. 8). DSB repair assessment, beyond the time allowed by explant culture, may be needed to determine the full repair capacity of the surviving oocytes. Reduction in γ-H2AX foci suggests that transient inhibition of CHEK2-dependent checkpoint with AZD7762 prevents triggering apoptosis and enables DNA repair.

Despite improved oocyte survival, retardation of ovarian growth was also observed in cotreatment with chemotherapy drugs resulting in visibly smaller explants after 7 days of culture (Fig. 7A). This suggests that at doses required for CHEK2 inhibition and oocyte protection, AZD7762 strongly inhibited CHEK1 and impaired proliferation of ovarian somatic cells and ovarian growth. Accumulation of pCHEK1(S317) was detected by Western blot in all AZD7762-treated ovaries (Fig. 7C). CHEK1 is phosphorylated at serine-317 by ATR kinase in response to replication blocks and some forms of genotoxic stress (58). Immunostaining for DNA damage marker γ-H2AX and terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL) for apoptosis showed that AZD7762 treatment alone caused accumulation of DNA damage (Fig. 9A) and increased apoptosis in ovarian somatic cells (Fig. 9B), particularly granulosa cells of growing follicles. AZD7762 treatment did not cause DNA damage in primordial oocytes; therefore, its effects appear to be restricted to proliferating cells. Because Chek2^−/− ovaries show normal growth after drug treatment (Fig. 1), these results suggest that AZD7762 inhibited CHEK1 kinase that interfered with cell cycle progression in proliferating stromal and granulosa cells, which triggered cell death independently of CHEK1 and CHEK2 signaling. In summary, these data show that transient inhibition of CHEK2 is a feasible approach to protect ovarian PMF reserve during genotoxic chemotherapy treatments, providing that either less cytotoxic dual CHEK1/2 or highly selective CHEK2 inhibitor can be used.

Animals

All procedures used in this study were approved by the Institutional Animal Care & Use Committee (IACUC) at the Jackson Laboratory. C57BL/6J (JAX stock#000664), Tg(Pou5f1-EGFP)2Mnn/J (JAX stock#004654), and Trp53^tm1Tyj/J (JAX stock#002101) were obtained from the Jackson Laboratory. Chek2^{tm1b(EUCOMM)Hmgu} mice were obtained from KOMP program at JAX. The Trp63^S621A mutant line was generated using CRISPR-Cas9–mediated genome editing as described in fig. S2. For radiation experiments, 7-day-old females were irradiated using a Cesium-137 gamma irradiator. They were exposed to a total dose of 0.5 or 3 Gy administered at a rate of (~170 rad/min). Ovaries were collected for protein extracts for Western blot analysis or fixed in 4% PFA for immunostaining. For chemotherapy drug treatments, 7-day-old females received a single i.p. injection of saline, CDDP (5 mg/kg) (Millipore Sigma), or CTX (150 mg/kg) (Millipore Sigma). Doses were chosen based on previously published studies (64). Mouse weights were recorded 1 and 2 weeks after injection. Both ovaries from each female were collected 2 weeks after injection, fixed in Bouin’s solution for histological analyses.

Follicle quantification

Bouin’s fixed ovaries were embedded in paraffin and cut into 5 μm serial sections. Sections were stained with Periodic acid-Schiff (PAS) and follicle number was evaluated in every fifth section. Follicle stages were determined using standard methods. Briefly, PMFs were surrounded by a layer of squamous pre-granulosa cells; primary follicles had a single layer of cuboidal granulosa cells; secondary follicles had more than one granulosa cell layer but lacked antrum; antral follicles had visible antrum. Final average follicle count per ovary per female is represented as the sum of follicles in every fifth section multiplied by a correction factor of 5 (89). Percentage survival was calculated by normalizing oocyte counts per treated ovary to the average oocyte number of the vehicle control group for the same genotype [i.e., % survival = (WT Drug/mean WT Veh)*100%].

Organ culture and drug ex vivo treatments

Ovaries collected from 7-day-old pups were cultured on polycarbonate membrane (Whatman Nucleopore Polycarbonate membrane) in MEM (Gibco) supplemented with 10% fetal bovine serum (FBS, Seradigm), 25 mM Hepes (Lonza), and 1× Pen/Strep (Gibco). Ovarian explants were incubated at 37°C, 5% CO₂, and atmospheric O₂. Chemotherapy drugs and inhibitors were purchased in powder form and reconstituted following the manufacturer’s recommendations: CDDP (Selleckchem), MAFO (CTX-analog, US Biological), DOX (Selleckchem), ETO (Merck Millipore), AZD7762 (IC50 CHEK2 10 nM and CHEK1 5 nM) (Selleckchem), CCT241533 (IC50 CHEK2 3 nM and CHEK1 245 nM) (TOCRIS), LY2606368 (IC50 CHEK2 8 nM and CHEK1 < 1 nM) (Selleckchem), PF477736 [IC50 CHEK2 47 nM(Ki) and CHEK1 0.49 nM(Ki)] (TOCRIS). Stock solutions for all compounds were prepared in dimethyl sulfoxide (DMSO) except for CDDP, which was dissolved in dimethylformamide (DMF) (as recommended by the supplier). DMSO and DMF were used as vehicle controls and their resulting concentration in culture media was less than 0.1%. We used MAFO in ex vivo culture to imitate the in vivo activity of CTX, which is metabolized by the liver to generate metabolites with alkylating properties (28, 90). During drug treatments, ovaries were cultured with drugs for 2 days and then cultured in drug-free medium for 5 days before further experimentation (7 days total). The culture medium was changed every 2 days. During inhibitor treatments, ovaries were cultured with inhibitors for 2 hours before irradiation or drug treatment. After radiation (0.5 Gy), medium was changed to inhibitor-free ~24 hours after radiation. When ovaries were cultured with drugs, replacement to normal medium (drug and inhibitor-free) occurred at 48 hours after treatment start. Ovaries were cultured for a total of 7 days for oocyte survival analyses or for 3 to 24 hours for protein analyses. Culture experiments were repeated two to three times each with ≥1 ovary per genotype (depending on litter size and genotype availability). Typically, ovaries from females of the same genotype dissected at the same time were randomly distributed between vehicle and treatment samples to reduce the number of experimental animals. In some cases, one ovary per female was used for vehicle control and the other one for treatment. Drug and inhibitor concentrations used were as follows: CDDP: 0.1 μg/ml (0.3 μM), 0.25 μg/ml (0.8 μM), and 0.5 μg/ml (1.6 μM); MAFO: 0.1 μg/ml (0.4 μM), 0.5 μg/ml (1.2 μM), and 1 μg/ml (2.4 μM); DOX: 25 ng/ml (43.1 nM), 50 ng/ml (86.2 nM), and 100 ng/ml (172.5 nM); ETO: 50 ng/ml (84.5 nM), 100 ng/ml (170 nM), and 500 ng/ml (849.5 nM); AZD7762 (0.1, 1, and 10 μM), CCT241533 (0.05, 0.5, and 5 μM), LY2606368 (0.1, 1, and 10 μM), and PF477736 (0.5, 5, and 50 μM).

Immunohistochemistry and TUNEL staining

Slides with 5-μm ovarian sections were processed using standard methods. Briefly, sections were deparaffinized and re-dehydrated before antigen retrieval using sodium citrate buffer (10 mM sodium citrate and 0.05% Tween 20, pH 6.0). Sections were permeabilized in 0.2% Triton X-100 in phosphate-buffered saline (PBS) and blocked with 10% goat serum for 1 hour, before incubation with primary antibodies at 4°C overnight. Secondary antibodies were applied for 1 hour and slides were mounted with VectaShield (Vector Laboratories) with DAPI (4′,6-diamidino-2-phenylindole) or after Hoechst 33342 (Biotechne) staining. Primary antibodies used in this study were mouse anti-p63 (4A4, Biocare Medical, CM163A), rabbit anti-pCHEK2(T68) (Bioworld Technology, BS4043: validated in Chek2^−/− tissue), rabbit anti-phospho-p53(S15) (Cell Signaling #9284), rabbit anti-DDX4 (Abcam, ab13840), rabbit anti-FOXL2 (Abcam ab275153), mouse anti-γ-H2AX (Millipore, 05–636), rabbit anti-RAD51 (Abcam, ab176458), and rabbit anti-53BP1(Novus NB100–304). Secondary antibodies used were Alexa Fluor (Invitrogen). Immunostaining for phospho-p53(S15) was performed with Starr Trek reagent (Biocare Medical). Antibody co-immunostaining with TUNEL for apoptotic cells (DeadEnd Fluorometric system, Promega, G3250) was performed by standard immunofluorescence protocol followed by TUNEL staining according to the manufacturer’s protocol. After completion of TUNEL staining, sections were incubated again with secondary antibodies.

Whole mount staining of ovaries

Whole mount immunostaining of cultured ovaries was performed as described in (91). Briefly, explanted ovaries attached to the polycarbonate membrane were fixed in 2% PFA at 4°C overnight. Fixed ovaries were washed with 70% ethanol overnight and placed in PBS for at least 4 hours before immunostaining. Ovaries were permeabilized in solution containing 0.2% polyvinyl alcohol (PVA), 0.1% NaBH₄, and 1.5% Triton X-100 in PBS for 4 hours, and incubated in blocking solution [10% goat serum, 3% bovine serum albumin, 0.15% glycine, pH 7.4, 0.1% Triton X-100, 0.2% sodium azide, penicillin (100 U/ml), and streptomycin (100 ng/ml), in PBS] overnight. Primary antibody incubation was performed at room temperature with gentle rocking for 2 to 4 days. Primary antibodies used were mouse anti-p63 (4A4, Biocare Medical, CM163A) and rabbit anti-DDX4 (Abcam, ab13840). Ovaries were washed with wash solution (0.2% PVA and 0.15% Triton X-100, in PBS) for 1 to 2 days with buffer changes (2× daily). Ovaries were next incubated with Alexa Fluor secondary antibodies for 2 to 3 days followed by washing for 1 to 2 days with buffer changes.

Optical clearing, imaging, and quantification of oocytes

Optical clearing of immunostained ovaries was performed as described in (91). Briefly, ovaries were cleared using ScaleS4(0) [40% d-(−)-sorbitol (w/v), 4 M urea, 10% glycerol, and 20% DMSO, in PBS]. ScaleS4(0) solution was refreshed twice daily until tissues became cleared (2 to 3 days). The membranes with ovaries were mounted using CoverWell Incubation Chamber (Research Products International) and imaged using a Leica DM550 microscope. Images were collected as Z-stacks (5 μm) and used to generate maximum projection image in LAS X software. Oocytes were counted using markers DDX4 (cytoplasmic staining) and p63 (nuclear staining). Small oocytes with a diameter less than 30 μm (delineated by DDX4 staining) were categorized as primordial oocytes and those larger than 30 μm were categorized as growing oocytes (corresponding to growing follicles including primary and secondary follicles). Percentage survival was calculated by normalizing oocyte counts per ovary to the average oocyte number of the vehicle control group for the same genotype [i.e., % survival = (WT Drug/mean WT Veh)*100%].

Immunoblots

Protein extracts were prepared with radioimmunoprecipitation assay buffer supplemented with protease and phosphatase inhibitors (Sigma-Aldrich) using a minipestle or Bioruptor Pico (Diagenode). Proteins were resolved in 10% acrylamide gel and transferred to nitrocellulose membranes. Primary antibodies used in this study were rabbit anti-p63 (Cell Signaling Technology, 13109), rabbit anti-total p53 (Leica, CM5P), mouse anti-γ-H2AX (Millipore, 05-636), rabbit anti-pCHEK1 (S317) (Cell Signaling #12302S), rabbit anti-DDX4 (Abcam, ab13840), and mouse anti-ACTB (GeneTex, GT5512). After incubation with HRP secondary antibody, signals were detected using Luminata Forte/Crescendo Western HRP substrate (Millipore). For probing with multiple antibodies, membranes were stripped by using Western blot stripping buffer Restore (Thermo Fisher Scientific).

Statistics

Statistical analysis was performed using PRISM 9.0 (GraphPad Software). To analyze the difference between more than two independent groups (e.g., vehicle versus drug doses) statistical analysis was performed by one-way analysis of variance (ANOVA) and the significance was determined by Bonferroni multiple-comparison test or Kruskal-Wallis for nonparametric data with Dunn’s multiple-comparison test. For pairwise comparisons (e.g., wild type versus mutant) significance was determined by t test or Mann-Whitney nonparametric test. Values of P < 0.05 were considered statistically significant. Data are presented as means ± SEM.*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; n.s., nonsignificant.

This PDF file includes: