ICR mice (5–6 weeks female and 10 weeks old male) were purchased from Animal Care Facility of Nanjing Medical University and were housed in ventilated cages at constant temperature (22°C) and controlled humidity and light dark cycle. All animal experiments were approved by the Animal Care and Use Committee of Nanjing Medical University and were performed in accordance with institutional guidelines.

Rabbit polyclonal anti-OCT4 antibody (Cat#: ab181557) was purchased from Abcam (Cambridge, MA, United States); mouse monoclonal anti-CDX2 antibody (Cat#:AM392-5M) was purchased from BioGenex (Fremont, United States); mouse monoclonal PAR/pADPr antibody (Cat#:4335-MC-100) was purchased from R&D Systems (Minnesota, United States). Donkey anti-Mouse Alexa Fluor 488, 555 and Donkey anti-Rabbit Alexa Fluor 555 antibodies (Cat#: A21202, A31570, A31572) were purchased from Thermo Fisher Scientific (Rockford, IL).

To promote ovulation, female mice were intraperitoneally injected with 7 IU Pregnant Mare Serum Gonadotropin (PMSG) followed by 7 IU of Human Chorionic Gonadotropin (hCG) after PMSG priming. Cumulus-oocyte complexes (COC) were isolated from oviduct ampullae 14–15 h post-hCG injection and were cultured in drops of HTF fertilization medium under mineral oil. Sperm was collected from the tail of epididymis of adult male mice and incubated in HTF fertilization medium at 37°C in a 5% CO2 incubator for 1 h before fertilization. Then capacitated spermatozoon was added to HTF drops containing COC and co-cultured together for 6 h in the incubator to allow for fertilization. Zygotes were then washed and transferred into drops of KSOM (Aibei Biotechnology, Nanjing, M1450) medium supplemented with chemical inhibitors or not. Chemical inhibitors were used at the following concentrations: 2 μM Minocycline hydrochloride (MiH, Selleck, S4226); 20, 50 or 100 nM Olaparib (Ola, APExBio, A4154); 0.2, 0.5 or 1 nM Talazoparib (Tala, APExBio, A4153). Embryos were observed and imaged with an inverted phase-contrast microscopy (Nikon Ts2R, Japan).

Embryos were fixed in 4% paraformaldehyde for 15 min and then were permeabilized with PBS containing 0.2% Triton X-100 for 10 min at room temperature (RT). After being blocked for 1 h in blocking buffer, which comprised of PBS together with 0.1% BSA, 0.01% Tween-20 and 2.5% donkey serum, embryos were incubated with primary antibodies diluted in blocking buffer overnight at 4°C. Embryos were then washed for three times with PBS and labeled with secondary antibodies in the dark for 1 h at RT. Samples were then washed for three times with PBS and stained with 1 μg/ml DAPI for 5 min, and washed for three times before mounting on glass slides in small drops of antifade medium. Samples were then imaged using an inverted phase-contrast microscopy (Nikon Ts2R, Japan).

Zygotes were collected in drops of HTF medium after fertilization and place them on the warmed microscope stage for 1 h at 37°C. Zygotes were then transferred to the KSOM medium with or without MiH to culture.

After 100 h, the blastocysts of the above three groups were randomly selected and surgically transferred into the uteri of pseudopregnant female mice. 9 days after transferring, mice were euthanized to see whether they were pregnant.

An IVC assay was also carried out to observe the developmental potential in vitro by culturing the rest of blastocysts in the afore mentioned experiments according to a protocol we used before (Zhao et al., 2021). All the blastocysts were embedded in Matrigel drops and culture for further 120 h to see whether they could form egg cylinder structures.

Zygotes were incubated in HTF fertilization medium containing 0.1 mM H2O2 for 1 h at 37°C, then washed with fresh HTF fertilization medium three times to remove H2O2 and transferred into KSOM medium for further culture. NAC group of ROS measurement experiment were performed as control by addition of N-Acetyl-l-cysteine (Sigma-Aldrich, A9165) to a working concentration of 5 mM.

TUNEL assay was carried out to analyze apoptosis of embryos using One Step TUNEL Apoptosis Assay Kit (Beyotime, C1086) in accordance with the instruction manual. Embryos were fixed in 4% paraformaldehyde for 30 min and permeabilized in PBS containing 0.5% Triton X-100 for 5 min. Then embryos were incubated in the TUNEL reaction mixture (containing FITC-conjugated dUTP and terminal deoxynucleotidyl transferase) at 37°C in the dark for 1 h. The reaction was terminated by washing in washing buffer (containing 0.1% Tween-20 and 0.1% BSA in PBS) for three times. Finally, the embryos were stained with DAPI for 5 min at RT and washed before mounting on glass slides. The TUNEL labeling was observed using a fluorescence microscope (Nikon Ts2R, Japan).

2,7-dichlorodihydrofluorescein diacetate (DCFH-DA) was used to evaluate the intracellular ROS level in embryos. Embryos were incubated in KSOM medium supplemented with 10 μM DCFH-DA for 30 min at 37°C in a 5% CO2 incubator and stained with Hoechst 33342 for 10 min. Fluorescence was observed under a Laser Scanning Confocal Microscope (LSM 710, Zeiss, Germany) at a 488 nm excitation wavelength and analyzed with the Image J software.

Statistical analysis was performed using GraphPad Prism 7 software and all results were presented as means ± standard deviation from three independent experiments. Data were analyzed with the Student’s t-test. ∗0.01<p < 0.05; ∗∗p < 0.01; no labeling indicates no statistical significance.

Firstly, to decipher the effect of MiH on preimplantation mouse embryos, in vitro fertilization (IVF) was employed to obtain zygotes, which were randomized into two groups and cultured with KSOM medium in the absence (control group) or presence of 2 μM MiH (MiH group) till most of them developed into blastocysts at 100 h after fertilization (Figure 1A). We found that almost all embryos in MiH group progressed normally at 2-cell and 4-cell stages (Supplementary Figure S1A) and there were no statistical differences in the rates of 2-cell or 4-cell embryos to total zygotes at 16 h or 40 h, respectively (Supplementary Figures S1B,C).

However, MiH-treated embryos showed slower progression of development since 53 h, with only 50% of embryos being consisted of 7–8 cells while that of control group was nearly 65% (Figure 1B). Then nearly 90% of embryos in both groups would undergo compaction and continue to form morula at 70 h (Figure 1C). In order to count the cell number of each embryo, we stained nuclei with DAPI and found that the count of each morula in MiH group was significantly fewer than that in control group (Supplementary Figures S1C,D), indicating a slower developmental kinetics as well. The difference became more evident at 85 h, with only 45% of zygotes formed blastocysts, whereas the efficiency of control group was about 70% (Supplementary Figures S1A,C). Notably, the blastocyst formation efficiencies of the two groups were comparable at 100 h (Figure 1D).

In order to further determine whether there was any other difference in the blastocysts, we classified blastocysts into three types depending on their expansion level as cavitated, expanded and hatching. The cavitated blastocyst is one whose blastocele has already formed, but it later continues to fill with fluid so that the blastocyst can expand. When it is expanded, the blastocyst is larger and the zona pellucida is thinner, so hatching can begin (Figure 1D). We analyzed the frequencies of these three types of blastocysts in the two groups and found a higher ratio of cavitated blastocyst in the company with lower ratios of expanded and hatching blastocyst in MiH group without statical significance (Figures 1D,E), which could be attributed to the slower developmental rate. However, the total cell number per blastocyst was comparable between the ones in control and MiH group at 100 h (Figure 1F). Here, cavitated blastocysts in the two groups were excluded for statistics because of the great individual variation. Combined these data, we could not exclude the possibility that the above phenomenon was caused by slower pumping of fluid into the blastocyst cavity.

In brief, MiH would develop slower during cleavage stage but formed blastocyst with similar efficiency at last. These were not consistent with findings of T. Osadaet al. and Imamura T et al., who used 3-ABA, PJ-34 and 5-AIQ. The inconsistency perhaps was related with inappropriate does and different side effects among inhibitors.

For the maintenance of EPS self-renewal, MiH can be replaced by other PARP1 inhibitors and Parp1-deficient mouse EPS could contribute to both TE and ICM in the absence of MiH (Yang et al., 2017). We further examined whether the influence of MiH on embryonic development was through inhibiting PARP1 as well. Olaparib (Ola) and Talazoparib (Tala) were two Food and Drug Administration (FDA) approved canonical PARP1 inhibitors that recommended for the treatment of various cancers (Robson et al., 2017; Litton et al., 2018). Then we treated zygotes with 20, 50 and 100 nM Ola and 0.2, 0.5 and 1 nM Tala. Nearly half of the zygotes were impaired in higher concentrations groups while embryos in 20 nM Ola and 0.2 nM Tala-treated groups progressed normally (Supplementary Figure S2). To further evaluate effects of PARP1 inhibitors at lower concentrations on embryonic development, we traced embryos in the four groups at multiple timepoints during blastocysts formation. We noticed that embryos treated with 20 nM Ola or 0.2 nM Tala showed no difference in the formation efficiencies of 2-cell nor 4-cell at 16 and 40 h respectively (Supplementary Figure S3), similar to those in MiH group (Supplementary Figure S1). However, embryos treated with PARP1 inhibitors turned to develop slower at 53 h, with only 40% of zygotes in Ola group contained 7–8 blastomeres. The ratio of Tala group was about 35%, comparable to that of MiH group but significantly less than that of control group (Figures 2A,B). The delay turned to be obvious at 85 h that the blastocyst formation rates were significantly decreased in Ola- and Tala-treated groups (Figures 2A,C), whereas the final outcome was not impaired (Figures 2A,D). Moreover, embryos treated with Tala developed most slowly among those in four groups, which might be ascribed to the most potent effects of inhibition on PARP1 (Murai et al., 2012; Murai et al., 2014).

To further address whether PARP1 was inhibited in embryos treated with MiH, Ola and Tala, PAR (poly ADP-ribose polymer), a product of PARP1 activity, was detected with an anti-PAR antibody to assess inhibition. It was previously reported that distribution of PAR remained diffused and cytosolic during preimplantation development (Imamura et al., 2004). As shown in Figures 2E,F, the distribution of PAR was diffuse and cytosolic in some cells of blastocysts in control group. On the contrary, embryos cultured in KSOM supplemented with MiH, Ola and Tala showed no signal. These results suggest that PARP1 was inhibited by MiH and two other well-known PARP1 inhibitors as previous reports.

MiH is a critical chemical compound for EPS cells maintenance (Yang et al., 2017). We hypothesized that MiH played similar roles in first cell fate decision that generated populations of outside cells and inside cells, respectively. This is a critical developmental stage because the embryo must allocate its blastomeres into either TE or ICM. To determine whether supplementation of MiH, Ola or Tala would affect the specification of TE and ICM, immunostaining for CDX2 and OCT4 was applied (Palmieri et al., 1994; Niwa et al., 2005; Strumpf et al., 2005). As shown in Supplementary Figure S4A, OCT4 was almost expressed in each blastomere, while CDX2 was only expressed in the cells which lately develop into TE at 70 h. All of the PARP1 inhibitors-treated embryos contained much more CDX2+/OCT4+ cells than untreated ones but significantly fewer CDX2-/OCT4+ cells. When blastocyst forms, OCT4 and CDX2 is restricted to ICM and TE cells separately, with several CDX2+/OCT4+ cells exist. In PARP1 inhibitors-treated blastocysts, more CDX2+ cells emerged with fewer CDX2-/OCT4+ cells remaining (Supplementary Figure S4B), indicating that PARP1 inhibition might promote specification of TE identity, similar to the effect of MiH on EPS cells.

Thus, we supposed that MiH affected embryo development through inhibiting PARP1 and boosted a trophoblast bias.

In vitro fertilization (IVF) is an effective clinical strategy for the couple who fail to conceive. While the percentage of successes is much higher nowadays, one of the most plausible causes of the failure of IVF procedures is the poor quality of gametes leading to aberrant embryonic development. Previous Studies have shown that psychological stress can exert detrimental effects on reproduction in women. In vitro fertilization techniques, in particular, gametes collection, manipulation, and culture may generate stress environment which would cause oxidative stress. Furthermore, some studies suggested that inhibited PARP activity could protect against the loss of cell viability, preserve NAD + levels and improve cellular bioenergetics in in vitro experiments in U937 cells subjected to oxidative stress (Ahmad et al., 2019). Thus, we wonder whether MiH could restore the outcome of the embryos under stressful environment. To verify this hypothesis, we placed zygotes on the warmed microscope stage for 1 h, then cultured them in medium in the absence (in vitro exposure, IVE) or presence of MiH (IVE+MiH). The ratios of 2-cell and 4-cell embryos were strikingly resembling in all groups (Figures 3A,B, Supplementary Figures S5A,B). Since 53 h, zygotes in IVE and IVE+MiH groups developed slowly, with no more than 50% of them reaching 8-cell stage while the ratio of control group was 80% (Supplementary Figures S5A,C). Later, nearly 90% of zygotes in IVE+MiH group developed into morula, which was much higher than that of IVE group and was comparable to control group (91 vs. 96% control; 83 vs. 96% control) (Supplementary Figures S5A,D). Though long-term exposure to external environment significantly reduced blastocyst formation efficiencies of embryos in both IVE and IVE+MiH groups at 85 h (Supplementary Figures S5A,E), the supplement of MiH rescued the final blastocyst formation potential of embryos in IVE+MiH group, with equivalent ratio (93%) to that of control group, but significantly higher than that of IVE group (77%) at 100 h (Figures 3A,B). Additionally, the average cell number of blastocysts in IVE group was markedly less than that in control and IVE+MiH groups (Figure 3C). Analogously, the frequency of hatching blastocyst in IVE group decreased dramatically than that of control and IVE+MiH groups, while the ratio of cavitated blastocyst exhibited much higher in IVE group (Figure 3D). All these data above suggested that MiH played a protective role for preimplantation embryos in the process of suffering long-term exposure to environment.

In order to further address whether these recovered blastocysts possessed normal developmental potential beyond preimplantation stage, they were randomly selected from the three groups and surgically transferred into the uteri of pseudopregnant female mice. 9 days after transferring, only one which had been transferred with blastocysts in IVE+MiH group was pregnant, indicating that MiH may partly rescue the embryo developmental potential from long time exposure to the external environment (Supplementary Figure S6). However, we could not make a very solid conclusion that embryos in this group possessed better developmental potential beyond preimplantation stage, due to the limited sample size and failure of control group. Actually, a satisfyingly pregnant rate of embryo transplantation experiment will be inevitably affected by the uncertainty of embryo implantation even in the presence of what appears to be a receptive endometrium. It is well known that late blastocyst development is delayed in embryos in in vitro culture relative to their in vivo counterparts. However, we had to culture all of the embryos for 100 h to ensure the restoration of the suboptimal environmental exposure, perhaps resulting in the miss of appropriate transplanting time, especially for those unexposed ones. This perhaps in turn led to embryo-endometrial desynchronization of control group.

We also performed an IVC assay to observe the developmental potential in vitro by culturing the rest of blastocysts in the afore mentioned experiments according to a protocol we used before (Zhao et al., 2021). However, the egg cylinder formation efficiencies of all the three groups were as low as less than 10%, hardly to say whether IVE+MiH blastocysts obtained better further developmental potential. When the natural fertilized and in vivo developed E3.5 embryos were used, in the previous studies, the efficiencies were only 30–40% as well (Bedzhov et al., 2014; Zhao et al., 2021), which could partly explain the low efficiency in our experiments. Though both the transplantation and IVC assays were statistically not conclusive, they provide some hints of what to expect if the experiments were performed many times.

Clinical data collected in recent years proved that embryo culture impaired embryonic developmental potential including delaying cell cycle kinetics and reducing TE cells in blastocysts (Giritharan et al., 2007). We then asked whether IVE affected the specification of TE and ICM and observed significant reduction of CDX2 positive cell numbers of blastocysts in both IVE and IVE+MiH groups comparing with those in control and MiH groups. However, blastocysts in IVE+MiH group contained much more CDX2 positive cells than those in IVE group, implying that MiH might protect TE cells from environmental stress (Figure 3F). Meanwhile, blastocysts in all three groups had much less OCT4 positive cells when they were compared with ones in control group, with great variations between individuals in IVE group (Figure 3G). Hence, we supposed the protective effect of MiH might be partly attributed to its boost of cells into a TE fate.

Long term exposure to external environment may lead to unexpected pH and temperature shifts in the embryo culture medium. Alternation in pH has been shown to influence intracellular homeostasis, with particular effects, including protein synthesis, mitochondrial function, cellular metabolism, and cytoskeletal remodeling (Will et al., 2011). Temperature is another important factor for embryonic development, which will influence integrity of spindles and DNA fragmentation. It has been reported that over-activation of PARP1 led to apoptotic and necrotic cell death during Myocardial ischemia-reperfusion injury (Raedschelders et al., 2012). To confirm whether long term exposure to external environment would cause PARP1 over-activation, we performed immunodetection of PAR and found a significant stronger signal in embryos of IVE group, which exhibited weaker in embryos of both control and IVE+MiH groups (Figures 3H,J).

Besides, unavoidable environmental factors, such as light exposure, excess temperature and pH fluctuation of culture medium, which increase ROS production, have been recognized to negatively affect embryo developmental potential and result in suboptimal pregnancy rates (Agarwal et al., 2005a). For this reason, we wondered whether long term exposure to external environment induced the production of ROS. To verify this problem, we explored the intracellular ROS level at 100 h. Fluorescent analysis showed robust signal in the embryos of IVE group while supplementing culture medium with MiH could restore it to normal level (Figures 3I,K). Thus, MiH might reduce the adverse effects of long-term exposure to external environment on embryos through inhibiting the generation of ROS.

Having shown that MiH could partly reduce ROS level in embryos that exposed to external environment, we assessed whether PARP1 inhibitors would improve embryo viability against oxidative damage. Hydrogen peroxide (H2O2) is one of the strongest oxidants and will lead to overproduction of ROS. However, increased ROS production induces multiple cellular damages and mitochondrial alternation, which consequently disturbs embryonic development of preimplantation embryos in vitro (Liu et al., 2000; Kitagawa et al., 2004; Zhang et al., 2010). Accordingly, zygotes were randomly divided into untreated and treated groups. In the treated group, zygotes were exposed to 100 μM H2O2 for 1 h, washed extensively, and then cultured in medium with (H2O2+MiH) or without MiH (H2O2) while untreated zygotes were incubated in HTF for 1 h and then cultured in KSOM medium (Control) (Figure 4A). Since 40 h, H2O2-treated zygotes in both two groups (H2O2+MiH and H2O2) exhibited lower rates of 4-cell and 8-cell stages embryos (Supplementary Figure S7). Only zygotes in H2O2 group showed decreased rates of morula at 70 h (Figures 4B,C). Furthermore, though blastocyst formation efficiencies of H2O2+MiH, H2O2-treated zygotes were both much lower than those of non-treated ones at 85 h (Figure 4D), rate of H2O2+MiH group was restored to the equivalent level of control group at 100 h (Figure 4E). Morphologic analysis also revealed that treatment of zygotes with H2O2 induced fragmentation and developmental retardation during this process, while embryos in H2O2+MiH and control group exhibited less (Figure 4B). Moreover, the average cell numbers of blastocysts at 100 h significantly decreased in the H2O2 and H2O2+MiH groups than that of control group (Figure 4F). To further address the influence of H2O2 on cell fate decision, embryos were immunostained with CDX2 and OCT4 antibodies. We found that though CDX2 positive cell numbers decreased both in H2O2 and H2O2+MiH treated ones, blastocysts in H2O2+MiH group had much more CDX2 positive cells than those in H2O2 group. By contrast, no restoration of OCT4 positive cells were found in H2O2+MiH treated ones (Figures 4G–I), implying that the protective effect of MiH was associated with protection for TE cells.

Additionally, to confirm whether the antioxidative effect of MiH was through inhibition of PARP1, we firstly explored the level of PAR in the expanded blastocysts. As shown in Figures 5A,B, we found that H2O2 exposure induced nucleic localization signal of PAR. But in H2O2+MiH and H2O2+Ola groups, almost no signal was found in blastocysts, which was consistent with that in all non-treated groups. Combined with the data that no PAR signal emerged in H2O2-untreated groups, we indicated that MiH and Ola would inhibit PARP1 overactivation which was induced by H2O2 treatment. Then, we used N-acetyle-cysteine (NAC), a conventional antioxidant for ROS inhibition, as a positive control to investigate whether MiH could decrease ROS level in H2O2-treated embryos like NAC did. These results showed that H2O2 treatment increased ROS signal in embryos, but both PARP1 inhibitors and NAC supplementation could recover the ROS level (Figures 5C,D).

A TUNEL assay was also performed to establish whether MiH reduced apoptosis rate in blastocysts developed from H2O2-treated zygotes. As expected, H2O2 caused severe apoptosis in embryos in comparison with those in non-treated ones, however, blastocysts in H2O2+MiH and H2O2+Ola groups possessed much fewer apoptotic cells than those in H2O2 group, suggesting a protective role of MiH against apoptosis in embryonic development (Figures 5E,F).