Keap1-targeting microRNA-941 protects endometrial cells from oxygen and glucose deprivation-re-oxygenation via activation of Nrf2 signaling

Background Mimicking ischemia-reperfusion injury, oxygen and glucose deprivation (OGD)-re-oxygenation (OGDR) applied to endometrial cells produces significant oxidative stress and programmed necrosis, which can be inhibited by nuclear-factor-E2-related factor 2 (Nrf2) signaling. MicroRNA (miRNA)-induced repression of Keap1, a Nrf2 suppressor protein that facilitates Nrf2 degradation, is novel strategy to activate Nrf2 cascade. Methods MicroRNA-941 (miR-941) was exogenously expressed in HESC and primary human endometrial cells, and the Nrf2 pathway examined by Western blotting and real-time quantitative PCR analysis. The endometrial cells were treated with OGDR, cell programmed necrosis and apoptosis were tested. Results MiR-941 is a novel Keap1-targeting miRNA that regulates Nrf2 activity. In T-HESC cells and primary human endometrial cells, ectopic overexpression of miR-941 suppressed Keap1 3′-UTR (untranslated region) expression and downregulated its mRNA/protein expression, leading to activation of the Nrf2 cascade. Conversely, inhibition of miR-941 elevated Keap1 expression and activity in endometrial cells, resulting in suppression of Nrf2 activation. MiR-941 overexpression in endometrial cells attenuated OGDR-induced oxidative stress and programmed necrosis, whereas miR-941 inhibition enhanced oxidative stress and programmed necrosis. MiR-941 overexpression and inhibition were completely ineffective in Keap1−/Nrf2-KO T-HESC cells (using CRISPR/Cas9 strategy). Restoring Keap1 expression, using an UTR-depleted Keap1 construct, abolished miR-941-induced anti-OGDR activity in T-HESC cells. Thus Keap1-Nrf2 cascade activation is required for miR-941-induced endometrial cell protection. Conclusions Targeting Keap1 by miR-941 activates Nrf2 cascade to protect human endometrial cells from OGDR-induced oxidative stress and programmed necrosis. Video Abstract Graphical abstract


Chemical and reagents
Puromycin and ploybrene were purchased from Sigma-Aldrich Chemicals Co. (St Louis, Mo). The fluorescent dye JC-1 was provided by Invitrogen Thermo-Fisher (Shanghai, China). All cell culture reagents and qPCR reagents were obtained from Gibco BRL Co. (Grand Island, NY). The antibodies of the present study were provided by Santa Cruz Biotechnology (Santa Cruz, CA) and Cell Signaling Tech (Suzhou, China). The primers, sequences, constructs and virus were designed and provided by Shanghai Genechem Co. (Shanghai, China), unless otherwise motioned.

Culture of primary human endometrial cells
The fresh human endometrial tissues, acquired from a written-informed consent uterine-bleeding patient (31-year old, administrated at Changzhou Second People's Hospital, undergoing the partial hysterectomy surgery) were digested with 0.15% trypsin-EDTA (Sigma) and Collagenase I (Sigma) for 1 h, and then transferred to DMEM/Hams F-12 nutrient medium plus 15% FBS. Tissues were dissolved. Blood vessel cells, immune cells and other non-endometrial cells were abandoned through gravity sedimentation. The remaining human endometrial cells were resuspended and cultured in complete DMEM medium [5]. The primary human endometrial cells at passage 3-10 were utilized for biomedical assays. The protocols of using human tissues and cells were approved by the Ethics Committee at Changzhou Second People's Hospital.

miR-941 overexpression or inhibition
The pre-miR-941 nucleotide sequence and its anti-sense sequence were synthesized and sequence-verified by Shanghai Genechem Co. Each of the two was ligated to the GV248 construct (Shanghai Genechem Co.). The construct, along with the lentivirus-packing helper plasmids (psPAX2 and pMD2.G [33]), were co-transfected to HEK-293 T cells, forming the pre-miR-941-expressing lentivirus ("lv-pre-miR-941") and the pre-miR-941 antisense lentivirus ("antagomiR-941"). Virus were enriched, filtered, and added directly to human endometrial cells (in the polybrene-containing medium). When necessary puromycin (5.0 μg/mL) was included in the medium to select stable cells, with miR-941 levels examined by qPCR.

Keap1 3′-UTR activity
Keap1 3′-UTR reporter plasmid (containing the miR-941-binding sites, at position of 276-283) was generated using the same protocol described previously [31], which was transfected to human endometrial cells using the Lipofectamine 2000 protocol. Afterwards, cells were subjected to the applied genetic modifications, with the Keap1 3′-UTR luciferase activity tested through the Promega kit [40].

Transfection of miR-941 mimic
Human endometrial cells were seeded into the six-well tissue culture plates (at 1 × 10 5 cells in each well). Lipofectamine 2000 was utilized for the transfection of 500 nM of the wild-type ("WT") or the mutant ("Mut") miR-941 mimics (synthesized by Shanghai Genechem Co.). After 48 h, miR-941 levels were determined by qPCR.

RNA-pull down assay
The RNA-Pull down assay was carried out through the previously-described protocol [41,42], testing miR-941bound mRNA using the Pierce Magnetic RNA Pull-Down Kit, Shanghai, China). In brief, T-HESC cells were transfected with biotinylated miR-941 mimic or control mimic (100 nmol/L) for 48 h, and cells were harvested using the lysis buffer described early [42]. The biotincaptured RNA complex was pulled down by incubating the cell lysates (600 μg of each treatment) with the streptavidin-coated magnetic beads [41]. The bound mRNA was purified using the RNeasy Mini Kit (QIA-GEN, Shanghai, China), with expression of Keap1 mRNA, in the bound fractions examined by qPCR. Its levels were normalized to the input controls.

Cell viability
Human endometrial cells were seeded into the 96-well tissue culture plates (3000 cells in each well). Following the applied treatments, a cell counting kit-8 (CCK-8) kit (Dojindo Laboratories, Kumamoto, Japan) was utilized to examine cell viability [5], with CCK-8 optic density (OD) examined at test-wavelength of 450 nm.

Lactate dehydrogenase (LDH) assay
The human endometrial cells were seeded into the 12well tissue culture plates (at 0.5 × 10 5 cells in each well). LDH release to the cell medium, the quantitative marker of cell necrosis [43], was tested through a two-step LDH detection kit (Promega, Shanghai, China) [5]. LDH contents in the medium were always normalized to total LDH levels.
OGD/re-oxygenation (OGDR) As described previously [4,5], human endometrial cells with applied genetic treatments were initially placed into an airtight chamber (95% N 2 /5% CO 2 ) for 4 h, mimicking OGD. Thereafter, human endometrial cells were returned back to the complete medium and reoxygenated. Cells were further cultured for applied time periods.

Western blotting
Human endometrial cells were seeded into the six-well tissue culture plates (at 1 × 10 5 cells in each well). Following the applied treatments, cellular lysates were achieved and quantified [4,5]. The lysates proteins (40 μg per treatment into each lane) were separated by SDS-PAGE gels, and transferring to polyvinylidene difluoride (PVDF) blots [44]. The detailed protocols for Western blotting and data quantification (through the Image J software) were previously described [45,46]. Assaying of nuclear fraction proteins was described early [46].

Mitochondrial immunoprecipitation (Mito-IP)
T-HESC cells were harvested and resuspended [5], with the supernatants collected as the cytosolic fraction lysates. The pellets were resuspended to achieve mitochondrial fraction lysates and quantified [47]. Mitochondrial fraction lysates (300 μg per treatment) were pre-cleared (using anti-IgG Sepharose beads), and incubated with anti-CypD antibody (Santa Cruz Biotech). The mitochondrial complex was captured by anti-IgG. CypD-p53-ANT1 association was examined by Western blotting assaying of CypD-bound proteins.

JC-1 mitochondrial depolarization assay
The human endometrial cells were seeded into the 12well tissue culture plates (at 0.5 × 10 5 cells in each well). In OGDR-stimulated cells with mitochondrial depolarization ("ΔΨ") the JC-1 fluorescent dye shall aggregate, forming green monomers [48]. The detailed protocol of JC-1 protocol was discussed previously [5]. The JC-1 fluorescence intensity was examined at 530 nm (Titertek Fluoroscan, Germany). The representative JC-1 images, integrating both green and red fluorescence images, were also presented.

Superoxide detection
Endometrial cells with the applied genetic treatments were initially seeded onto 96-well tissue-culturing plates (at 3 × 10 3 cells of each well). Following the applied OGDR stimulation, the superoxide colorimetric assay kit (BioVision, Shanghai, China) was applied to examine the cellular superoxide contents. In brief, the superoxide detection reagent (50 μL/well) was added for 15 min under the dark, with the superoxide's absorbance tested at the test-wavelength of 450 nm [31].

Lipid peroxidation assay
As reported [31] endometrial cells with the applied genetic treatments were initially seeded onto the six-well tissue-culturing plates (at 1 × 10 5 cells per well). Following the applied OGDR stimulation, the lipid peroxidation assay kit (Abcam, Shanghai, China) was utilized to examine cellular lipid peroxidation levels, via the malondialdehyde method. The lipid peroxidation levels, determined by the thiobarbituric acid reactive concentration, were tested and quantified using the previouslydescribed protocol [31,49].

NQO1 activity
The detailed protocol for testing the relative NQO1 activity in human endometrial cells has been described elsewhere [50]. In brief, the inducer potency was quantified by using the NQO1 bioassay. T-HESC cells or the primary human endometrial cells (10 4 per well of a 96well plate), with applied genetic treatments, were cultured for 24 h. The NQO1 enzyme activity was tested and quantified in cell lysates using menadione as the substrate.

Keap1 re-expression
The Keap1 (with no 3′-UTR region) expression GV248 construct was designed and provided by Shanghai Genechem, transduced to T-HESC cells with miR-941 overexpression. Cells were then selected by puromycin for 10 days, with Keap1 re-expression verified by qPCR and Western blotting assays.

Statistical analysis
Data were presented as mean ± standard deviation (SD). The repeated-measures analysis of variance (RMA-NOVA) with Dunnett's post hoc test for multiple comparisons (SPSS 16.0, SPSS Co. Chicago, CA) were utilized to evaluate statistical significance. The twotailed unpaired T test (Excel 2013) was carried out to examine significance between two specific treatment groups. P < 0.05 was considered statistically significant.
Testing the potential effect of miR-941 on Nrf2 signaling, we found that although Nrf2 mRNA levels (Fig. 1f) were unchanged in lv-pre-miR-941-expresing cells, Nrf2 protein levels (Fig. 1e) were significantly elevated (vs. . T-HESC human endometrial cells were transduced with lentiviral pre-microRNA-941 ("lv-pre-miR-941"), with selection by puromycin the stable cells were established, with control cells transduced with lentiviral non-sense microRNA ("lv-miRC"); Expression of mature miR-941 and listed mRNAs was tested by qPCR assays (b, d, f and h); Keap1 3′-UTR activity was shown (c), with expression of listed proteins in total cell lysates (e) and nuclei lysates (g) tested by Western blotting; The relative NQO1 activity was tested as well (i). T-HESC cells were transfected with 500 nM of non-sense microRNA control ("miRC"), the wild-type ("WT") or the mutant miR-941 mimics (sequences listed in j), with Keap1 3′-UTR activity (k) and Keap1 mRNA/protein expression (k) tested after 48 h. RNA-Pull down assay confirmed the direct association between biotinylated-miR-941 and Keap1 mRNA in T-HESC cells (l). The primary human endometrial cells ("Endometrial cells", same for all Figures) were infected with lv-pre-miR-941 or lv-miRC, with expression of listed genes tested by qPCR (m-o, and q) and Western blotting (p) assays after 48 h. The relative NQO1 activity was tested as well (r). Expression of listed protein was quantified and normalized (e, g, k and p). "Pare" stands for the parental control cells (same for all Figures). Data were presented as mean ± SD (n = 5), and results were normalized. * P < 0.05 vs. "lv-miRC"/"miRC" cells. Experiments in this figure were repeated five times with similar results obtained control cells). Additionally, the stabilized Nrf2 protein translocated to T-HESC cell nuclei (testing nuclear lysates, Fig. 1g). Furthermore, mRNA and protein expression of two key Nrf2-dependent genes, including HO1 and NQO1 [33,[53][54][55], were significantly increased in miR-941-overexpressing cells (Fig. 1e and h), with the NQO1 activity significantly boosted (Fig. 1i). These results show that miR-941 overexpression can induce the Nrf2 signaling cascade activation in T-HESC cells.
In the primary human endometrial cells ("Endometrial cells"), infection with lv-pre-miR-941 led to an over 10 fold increase in the expression of mature miR-941 (Fig. 1m), causing downregulation of Keap1 mRNA (Fig. 1n). MiR-941 overexpression similarly induced Nrf2 signaling activation, leading to Nrf2 protein (but not mRNA) accumulation ( Fig. 1o and p), HO1 and NQO1 upregulation (both mRNA and protein, Fig. 1p and q) as well as an increase in NQO1 activity (Fig. 1r). Collectively, these results show that miR-941 targets Keap1 and activates Nrf2 signaling in human endometrial cells.
The lv-pre-miR-941-expressing T-HESC cells were protected from OGDR, as indicated by increased cell viability (Fig. 2e) and reduced necrosis (medium LDH release, Fig. 2f), when compared to control cells. Further, T-HESC cells transfected with the wild-type ("WT") or the miR-941 mutants (see Fig. 1), were subjected to OGDR stimulation. Results show that transfection of the WT-miR-941 potently inhibited OGDR-induced cytotoxicity in T-HESC cells (Fig. 2g and h), while the mutants were completely ineffective (Fig. 2G and H).
In the primary human endometrial cells OGDRinduced ROS production (tested by superoxide accumulation, Fig. 2i), mitochondrial depolarization (Fig. 2j) and cytosol cytochrome C release (Fig. 2k) were significantly attenuated by ectopic miR-941 overexpression. The latter also alleviated OGDR-induced cytotoxicity in the human endometrial cells (Fig. 2l). These results show that miR-941 overexpression inhibited OGDR-induced oxidative stress, programmed necrosis and cytotoxicity in human endometrial cells.

Discussion
The expression and potential function of miR-941 in human endometrial cells is unknown. The study by Hu et al., has shown that miR-941 is a human-specific miRNA, highly expressed in the human brain and in other human cells [56]. It has regulatory effects on gene expression, participating in hedgehog-and insulinsignaling pathways [56]. Zhang et al., have shown that miR-941 levels are significantly downregulated in hepatocellular carcinoma (HCC) tissues. Functionally, miR-941 negatively regulated KDM6B (lysine (K)-specific demethylase 6B) to inhibit HCC cell epithelialmesenchymal transition (EMT) and cell migration/invasion [57]. Bai et al., reported that miR-941 levels are significantly higher in acute coronary syndrome patients than those in healthy controls [58].
Recent studies demonstrate that miRNA-induced silencing of Keap1, the Nrf2 suppressor protein, is a novel strategy to regulate Nrf2activation in human cells. In breast cancer cells miR-200a targeted Keap1 to activate Fig. 4 miR-941 inhibition intensifies OGDR-induced programmed necrosis in human endometrial cells. The stable T-HESC cells with the pre-microRNA-941 anti-sense lentivirus ("antagomiR-941", two lines, "L1/L2") or microRNA anti-sense control lentivirus ("antaC") were subjected to OGD exposure for 4 h, followed by re-oxygenation ("OGDR") for applied time periods, ROS production (superoxide contents, a), mitochondrial CypD-ANT1-p53 association (tested by mito-IP assay, b) as well as mitochondrial depolarization (JC-1 green fluorescence accumulation, c) and cytochrome C release (d, testing cytosol proteins) were tested; Cell necrosis was tested by medium LDH release assay (e). The primary human endometrial cells were infected with antagomiR-941 or antaC for 48 h, afterwards cells were subjected to the same OGDR stimulation and cultured for applied time periods, ROS production (f), mitochondrial depolarization (g) and cytosol cytochrome C release (h) were tested similarly, with cell survival and necrosis tested by CCK-8 (i) and LDH release (j) assays, respectively. For the mito-IP assay, CypD-bound ANT1 and p53 were quantified (b), with expression of CypD, ANT1 and p53 tested in the "Input" controls (b). For the cytochrome C release assay, relative cytosol cytochrome C level was quantified (d and h). Data were presented as mean ± SD (n = 5), and results were normalized. * P < 0.05 vs. "Mock" treatment in "antaC" cells. # P < 0.05 vs. OGDR treatment in "antaC" cells. Experiments in this figure were repeated five times with similar results obtained Nrf2 signaling cascade [36]. In SH-SY5Y neuroblastoma cells and differentiated human neural progenitor cells, miR-7 activated Nrf2 signaling by targeting and silencing Keap1 [35]. In retinal pigment epithelium cells (RPEs) and retinal ganglion cells (RGCs), miR-141 activated Nrf2 signaling by targeting miR-141, protecting cells against ultra-violet (UV)-induced oxidative stress [59]. A recent study by Xu et al., has shown that Keap1- Stable T-HESC cells with the CRISPR/Cas9-Nrf2-KO construct ("Nrf2-KO") (a-d) or the CRISPR/Cas9-Keap1-KO construct ("Keap1-KO") (e-h) were further infected with the pre-microRNA-941 anti-sense lentivirus ("antagomiR-941"), the pre-microRNA-941 lentivirus ("lv-pre-miR-941") or non-sense microRNA lentivirus ("lv-miRC") for 48 h. These cells or the control cells with empty CRISPR/Cas9-KO construct ("Cas9-C") were treated with OGD for 4 h, followed by re-oxygenation ("OGDR") for 24 h, cell survival (a and e) and necrosis (b and f) were tested by CCK-8 and LDH release assays, respectively. Expression of mature miR-941 (c and g) and listed proteins (in total cell lysates, d and h) were shown. The lv-pre-miR-941-expressing T-HESC cells were further transfected with the UTR-depleted Keap1 construct ("+Keap1") for 48 h, expression of listed genes was shown (i and j); Cells were subjected to the same OGDR stimulation for 24 h, cell necrosis (testing medium LDH release, k) and mature miR-941 expression (l) were tested similarly. Keap1 and Nrf2 protein expression was quantified and normalized to Tubulin (d, h and j). Data were presented as mean ± SD (n = 5), and results were normalized.. # P < 0.05 vs. OGDR treatment in "Cas9-C" cells (a-c, e-g). # P < 0.05 (i and k). Experiments in this figure were repeated three times with similar results obtained targeting miR-626 activated Nrf2 signaling and protected RPEs from oxidative injury [31]. Therefore, Keap1targeting miRNAsinduceNrf2 activation to protect human cells from oxidative stress.
The results of this study confirm that miR-941 is a direct and specific Keap1-targeting miRNA, that regulates the Keap1-Nrf2 cascade in human endometrial cells. In T-HESC cells and primary human endometrial cells forced miR-941 overexpression inhibited Keap1 3′-UTR reporter luciferase activity and downregulated Keap1 mRNA/protein expression, subsequently leading to Nrf2 activation. Conversely, Keap1 3′-UTR reporter luciferase activity and expression were elevated in endometrial cells with miR-941 inhibition, whereas Nrf2 activation was inhibited. RNA-Pull down experiments confirmed that miR-941 directly binds Keap1 mRNA in T-HESC cells. Functional studies confirmed that miR-941 exerted significant endometrial cytoprotection against OGDR. Ectopic miR-941 overexpression in endometrial cells largely inhibited OGDR-induced ROS production and programmed necrosis. Thus, miR-941 promotes Nrf2 cascade activation by targeting and silencing Keap1. Therefore, miR-941 expression provides a novel strategy to protect human endometrial cells from OGDR and possible ischemia-reperfusion injuries.

Conclusion
MiR-941 negatively regulates Keap1 to activate Nrf2 signaling cascade, thus protecting human endometrial cells from OGDR-induced oxidative stress and programmed necrosis.