MCF7, T47D, MDA-MB-231, and HCC1937 cells were obtained from the American Type Culture Collection. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) or Roswell Park Memorial Institute-1640 (RPMI-1640) (Macgene, Beijing, China) with 10% fetal bovine serum (FBS) (LONSA SCIENCE, Shanghai, China) and maintained in a humidified atmosphere containing 5% CO₂ at 37°C. Cells were transfected with specific plasmid by JetPRIME (Polyplus, Strasbourg, France) according to the manufacturer’s protocol. The BC tissue chip was purchased from Guge Biotechnology Company (Wuhan, China).

Anti-MOF (sc-81765) and breast-cancer susceptibility gene 1 (BRCA1) (Santa Cruz, sc-6954) were obtained from Santa Cruz Biotechnology. Antibodies including CUL4A (14851-1), CUL4B (12916-1), glyceraldehyde-3-phosphate dehydrogenase(GAPDH) (60004-1-Ig), and Flag (66008-2) were purchased from Proteintech (Wuhan, China). Other antibodies were listed as follows: H4K16ac (Epitomics, EPR1004), ERα (Cell Signaling Technology, Inc (CST), #8644), Ki67 (Abcam, ab16667), murine double minute 2 (MDM2) (Wanleibio, WL01906), HSP90 (Sangon Biotech, D120009), HSP90 K294ac (Rockland, 600-401-981), and acetylated lysine (CST, #9441). Inhibitors including MG149, cycloheximide (CHX), and MG132 were purchased from MedChemExpress (MCE, Princeton, NJ, USA). TAM was purchased from Sigma. CHX, MG149, MG132, and TAM were dissolved in dimethyl sulfoxide (DMSO).

Immunohistochemistry (IHC) staining was performed to detect the expression of MOF and ERα in BC tissue chips. Following deparaffinization and quenching of endogenous peroxidase, the tissue section was treated by deparaffinization and quenching of endogenous peroxidase and then subjected to antigen retrieval with sodium citrate buffer. Then, the section was incubated with 5% FBS and then incubated with ERα (1:100) and MOF (1:100) antibodies overnight at 4°C. After incubation with secondary antibody at 37°C, the section was subjected to staining by the DAB Detection Kit (Polymer) (GeneTech, Shanghai, China) and counterstaining with hematoxylin (Solarbio, Beijing, China) for the observation with a light microscope (Nikon, Tokyo, Japan). All slides were scored in an open discussion by two experienced pathologists, who were blinded to the outcome. Immunostaining was scored on the basis of the intensity score and quantity of positive cell score. Intensity score: negative, 0; weak, 1; moderate, 2; and intense, 3. Quantity of positive cell score: <5%, 0; 5%–25%, 1; 26%–50%, 2; 51%–75%, 3; and >75%, 4. The product of intensity score and quantity of positive cell score was used as the total score.

Cells were seeded onto coverslips in 24-well plate for growth to 70% cell confluence. Cells were fixed using 4% paraformaldehyde and blocked by 5% FBS and then subjected to incubation with the primary antibody for MOF or ERα (1:100) overnight at 4°C. After incubation with corresponding secondary antibody, cells were mounted with DAPI (4',6-diamidino-2-phenylindole) (C0060, Solarbio), and images were taken from a DP74 color fluorescence camera (Olympus, Tokyo, Japan)

Total RNA was isolated using RNAiso Plus (TaKaRa, Kyoto, Japan). RNA was reverse-transcribed by the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA). Quantitative reverse transcriptase PCR (qRT-PCR) was performed using the SYBR qPCR Mix (TOYOBO). GAPDH was used as an internal control. Primers for qRT-PCR were listed in Supplementary Table 1. Then, relative quantitation of gene expression was calculated using the 2^−ΔΔCT method.

Total protein was extracted using the sodium dodecyl sulfate (SDS) lysis buffer (1% sodium dodecyl sulfate, 5% glycerol, 1 mM ethylenediamine tetraacetic acid (EDTA), 25 mM Tris, 150 mM NaCl, and 1 mM phenylmethylsulfonyl fluoride (PMSF)). Protein samples were separated by SDS-polyacrylamide gel electrophoresis (PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, Bedford, MA, USA). After blocking with 5% non-fat milk powder and incubation with specific antibody at 4°C overnight, membranes were subjected to corresponding secondary antibody and then visualized by an ECL detection kit (Wanleibio, Dalian, China).

Proteins were extracted from cells using BC-200 lysis buffer (20 mM 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES), 200 mM KCl, 10 mM β-mercaptoethanol, 1 mM EDTA, 10% glycerol, and 0.1% NP-40) containing protease inhibitor cocktail (APExBIO, Houston, TX, USA). Extracted proteins were immunoprecipitated by incubation with 1 μg of antibody and followed by binding with Protein A/G magnetic beads (Bimake, Shanghai, China). After washing with lysis buffer, proteins were extracted by SDS sample buffer and detected by Western blotting.

Chromatin immunoprecipitation (ChIP) assay was conducted using the SimpleChIP Plus Sonication Chromatin IP Kit (Cell Signaling Technology). Cells were cross-linked by 1% formaldehyde followed by sonication. The immunoprecipitated DNA was analyzed by qPCR with specific primers listed in Supplementary Table 2.

For cell proliferation assay, cells were seeded into 96-well plate and then treated with 10 µl of CCK8 (Biosharp) per 100 µl of culture medium at specific time points. After incubation at 37°C for 4 h, the absorbance value was determined at 450 nm by a SPECTROstar Nano instrument for calculating cell proliferation curves. In the colony formation assay, cells were seeded in 6-cm culture dish (200 cells per dish) and culture in 37°C incubator for 10–14 days. After termination of culture, cells were fixed with methanol and stained by crystal violet. The colonies with more than 50 cells per colony were counted.

Female NSG mice aged 6–8 weeks were prepared. MCF7 cells (5 × 10⁷ cells) with stable MOF transfection or control vector were subcutaneously injected into one flank of each mouse, respectively. The tumor growth was observed every 2 days. After 3 weeks, mice were sacrificed, and tumors were taken out for size measurement. Tumors were fixed with 4% paraformaldehyde and followed by IHC staining. Animal experiments were performed with the approval from the Animal Research Ethical Inspection Form of Shandong University School of Life Sciences (SYDWLL‐2018‐19).

Data were statistically analyzed using GraphPad Prism software (San Diego, CA, USA) and were shown as means ± S.D. in three independent experiments. A Chi-square test was applied for analyzing pathological data. One-way ANOVA analysis was performed for time-course studies, and Student’s t-test was applied for comparisons of two groups. P < 0.05 was considered to be statistically significant.

MOF is reported as a critical suppressor in BC by inhibiting EMT and tumor invasion, suggesting a favorable prognosis (19); whereas ERα functions as the crucial oncogenic driver for the progression of BC (4). However, the relationship between the expression of MOF and ERα still remains elusive. To evaluate the correlation between MOF and ERα, we examined the protein expression level of MOF and ER in BC tissues from 78 patients. The immunohistochemical staining (IHC) results demonstrated the staining of MOF/ERα defined as either low/negative (weak or none) or high/positive (strong or moderate) based on the relative intensity of staining (Figure 1A). Statistical analysis of IHC results showed that around 64.5% of BC tumors with low MOF expression exhibited ERα-positive staining, whereas the majority (68.8%) of tumors with high MOF expression displayed ERα-negative staining, indicating that there is a negative correlation between MOF and ERα expression in BC tissues (Table 1, Figure 1B). Moreover, we examined the MOF and ERα expression in multiple BC cell lines. Western blot analyses revealed that the protein abundance of MOF with histone H4K16 acetylation exhibited a remarkably attenuated expression in ERα+ BC cells (MCF7 and T47D) compared with that in ERα− BC cells (MDA-MB-231 and HCC1937), and histone H4K16 acetylation also showed a similar pattern with MOF expression (Figure 1C). Taken together, these results suggested that MOF functions as a tumor suppressor in BC tumors and that the expression of MOF was negatively associated with that of ERα in BC tissues and cells.

To investigate whether and how MOF plays roles in the expression of ERα, plasmids of Flag-HA-MOF (for MOF overexpression) and pGPU6-shMOF (for MOF knockdown) were transfected into ERα+ BC cell lines (MCF7 and T47D), respectively. qRT-PCR analysis showed that the mRNA level of MOF was significantly upregulated or reduced in these cell lines, whereas the mRNA level of ERα had indistinguishable change (Figures 2A, C), suggesting that the expression of MOF did not regulate ERα expression at the transcriptional level. However, the protein abundance of ERα was obviously influenced by MOF overexpression or knockdown in a negatively regulatory manner (Figures 2B, D). Namely, both MCF7 and T47D cells transfected with Flag-HA-MOF plasmid (MOF overexpression) showed an increased amount of MOF and a decreased level of ERα protein (Figure 2B). Conversely, in the pGPU6-shMOF–transfected cells (MOF knockdown), the protein abundance of ERα was elevated (Figure 2D). In addition, when increased doses of Flag-HA-MOF plasmid were transfected into MCF7 cells, the ERα protein levels showed the corresponding downward trend with the gradually advanced expressions of MOF and H4K16ac (Figure 2E).

As a steroid hormone nuclear receptor, ER could be activated by estrogen 17β-estradiol (E2) (25). After binding with E2, homodimerized ER would translocate into the nucleus and functions as a transcription factor to regulate target gene transcription (25). Hence, we explored whether MOF-mediated regulation of ERα expression would be affected by E2 stimulation. After transfection with Flag-HA-MOF plasmid or empty vector as control for 48 h, cells were treated with E2 for 3 h. As depicted in Figure 2F, MOF overexpression could induce a significant reduction of ERα protein in the presence or absence of E2, suggesting that MOF downregulates the ERα protein level in an estrogen-independent manner. In addition, nuclear and cytoplasmic separation assay demonstrated that protein abundance of ERα in both cytoplasm and nucleus obviously decreased after MOF overexpression regardless of the presence of E2 (Figure 2G). Similar results were also observed by immunofluorescence staining that the distribution of ER in cytoplasm/nucleus was reduced in MOF-overexpressed MCF7 cells with or without E2 treatment (Figure 2H). These results suggest that MOF overexpression inhibited ERα protein levels in both cytoplasm and nucleus with or without E2 treatment.

We next determined the effect of MOF on ERα-mediated transactivation upon E2 stimulation. MOF-overexpressed MCF7 cells were treated with E2 for specified incubation time, and the results showed that the mRNA expression of the three endogenous target genes (TFF1, CCND1, and GREB1) of ERα were significantly upregulated by E2 after 3 h of incubation in the control group cells. Whereas MOF overexpression abrogated this expression raise of ERα target genes by E2 (Figure 3A), suggesting that the transactivation abilities of ERα on target genes upon E2 treatment was prohibited by MOF overexpression. ChIP analysis further demonstrated that under the stimulation of E2, the recruitment of ERα at the promoters of TFF1, CCND1, and GREB1 was inhibited by ectopic expression of MOF (Figure 3B).

We further investigate the biological function of MOF in ERα+ BC cells. Moreover, stable cell lines with MOF overexpression were established by lentivirus infection of MCF7 cells, and CCK8 assay was performed to determine the functional role of MOF in BC cell proliferation. As shown Figure 3C, E2 stimulation significantly promoted cell proliferation of ERα+ BC MCF7 cells in the control group, whereas in MOF-overexpressed cells, this E2-stimulated raise was abolished, suggesting that MOF overexpression inhibited E2-induced proliferation of BC cells. In addition, the inhibitory effect of MOF on MCF7 cell prol9iferation occurred regardless of the presence or absence of E2 (Figure 3C), indicating that MOF prohibits cell proliferation of ERα+ BC cells in an E2-independent manner. In addition, MCF7 cells with shMOF transfection showed that MOF knockdown led to increased cell proliferation (Figure S1A). Colony formation assay further showed that MOF overexpression restrained the colony formation ability of MCF7 cells whenever E2 is present (Figure 3D). In vivo tumor formation experiments revealed that the size of tumor formed by MOF-overexpressed MCF7 cells was obviously smaller than that of control group cells (Figure 3E), indicating that MOF overexpression significantly impeded the growth of subcutaneous tumors formed by ERα+ BC cells in mice. In addition, IHC staining showed that, compared with the control tumor, reduced expression of proliferation marker Ki67 was observed in the tumor tissue with MOF overexpression (Figure 3F). Taken together, these results demonstrated that MOF overexpression prevented cell proliferation and tumorigenicity of ERα+ BC cells through the inhibition on ERα function.

MOF downregulates ERα protein abundance in MCF7 cells. We speculated that the protein stability of ERα might be affected by MOF for the negative effect on ERα expression. By using CHX, an inhibitor of protein translation, to block de novo protein synthesis, we found that ERα protein stability was attenuated by MOF overexpression (Figure 4A). Under CHX treatment, the degradation of ERα protein was overtly accelerated by MOF overexpression compared with the control group. The half-life of ERα was reduced down to around 4 h in the MOF-overexpressed MCF7 cells compared with that to around 9 h in the control group (Figure 4A). Conversely, we found that the application of MG132 (an inhibitor of proteasome function) could strikingly prevent ERα protein degradation induced by MOF. With the time extension of MG132 treatment, ERα protein expression increased gradually in MOF-overexpressed MCF7 cells and reached a similar level as the control group at 6 h (Figure 4B), suggesting that MOF-induced ERα protein degradation occurred by the proteasome pathway. Furthermore, polyubiquitination of ERα protein by MOF was observed in co-immunoprecipitation (Co-IP) assay. MOF overexpression strengthened the polyubiquitination of ERα, as shown by more intense ladder band of polyubiquitin-conjugated ERα protein in Flag-MOF–transfected cells (Figure 4C). In addition, MOF knockdown resulted in the abrogation of ERα polyubiquitination to promote ERα protein stability (Figure S1C). These results indicated that MOF promoted ERα protein degradation through ubiquitin–proteasome pathway.

We further explored the E3 ubiquitin ligase responsible for the MOF-induced ERα ubiquitination and degradation. Several E3 ligases like MDM2, CHIP, RNF31, and BRCA1 have been reported to trigger polyubiquitination of ERα for ubiquitin/proteasome-mediated proteolysis (26–29). Our RNA-seq raw data suggested a possible upregulation of CUL4A, which belongs to the Culling-Ring E3 ligase subfamily, in MOF-overexpressed MCF7 cells. After investigating the expression of several E3 ligase candidates in MCF7 cells harboring overexpression or knockdown of MOF, it was found that CUL4A and CUL4B can be positively modulated by MOF in a qRT-PCR assay (Figure 5A). In addition, the protein abundance of CUL4A and CUL4B could be upregulated by MOF overexpression (Figure 5B). To further confirm the involvement of CUL4A or CUL4B in MOF-mediated ERα degradation, cells were co-transfected with FH-MOF plasmids and CUL4A or CUL4B small interfering RNA (siRNA). It was demonstrated that blockage of CUL4B but not CUL4A could abrogate MOF-induced ERα protein degradation (Figure 5C). Moreover, Co-IP assay revealed that CUL4B and ERα proteins could physically interact with each other (Figure 5D). In addition, CUL4B knockdown abolished MOF-encouraged ubiquitination of ERα as revealed by the reduced amount of polyubiquitin-conjugated ER in CUL4B siRNA-transfected cells (Figure 5E). These results indicated that MOF promoted the ubiquitination and protein degradation of ERα via upregulated CUL4B functioning as an E3 ligase.

Molecular chaperone HSP90 binds with ERα to maintain the conformational stability of ER for ligand binding and to protect ERα from protein degradation, whereas hyperacetylation of HSP90 inhibits its chaperone function for ERα (9). By Co-IP assay, it was shown that the acetylation level of HSP90 was overtly raised in MOF-overexpressed MCF7 cells, whereas the acetylation level of ERα was not obviously affected (Figure 6A). In addition, MOF knockdown also did not overtly affect the acetylation level of ERα but markedly decreased that of HSP90 (Figure S1B) It was further confirmed that MOF-induced hyperacetylation of HSP90 occurred through the K294 acetylation site (Figure 6B), which was reported to be determinant for the chaperone binding of HSP90 with its client proteins (30). Co-IP assay further revealed that when MOF was overexpressed in MCF7 cell, the interaction between HSP90 and ER was undermined, whereas more association of ER with CUL4B was observed instead (Figure 6C). Taken together, it is suggested that MOF enhanced the acetylation level of HSP90 at K294 site to attenuate the chaperone association of HSP90 with ERα, thereby liberating ERα to more interact with CUL4B for ubiquitin-mediated proteasomal degradation of ERα.

In addition to the negative regulation of MOF overexpression on ERα protein stability in ERα+ BC cells, we also examined the effect of MOF inhibition on ERα− BC cells. As shown in Figure 7A, knockdown of MOF in ERα-negative HCC1937 cells resulted in a recovery of ERα protein expression. As an HAT inhibitor, MG149 could inhibit MOF within a certain concentration range (47 ± 14 µM) because higher concentration would work on other histone acetylases (like 74 ± 20 µM for Tip60) (31, 32). First, the inhibitory effect of MG149 was verified in ERα+ MCF7 cells, where increased doses of MG149 could result in the raised abundance of ERα protein and reduced H4K16ac (Figure 7B). Then, treatment of 35 μM MG149 in ER− HCC1937 cells could obviously restore ERα protein expression similar as the effect of MOF knockdown (Figure 7C). Because reactivation of ERα expression could restore endocrine therapy sensitivity in patients with ERα− BC (12, 13), we further investigated the effect of MG149 on the sensitivity of ER− BC cells to TAM. Compared with the inhibition concentration IC50₅₀ of TAM at 41.06 μM in HCC1937 treated with TAM alone for 24 h, IC₅₀ was reduced to 21.26 μM with a combinatory treatment of TAM and MG149 for 24 h (Figure 7D), suggesting that MOF inhibitor MG149 could effectively improve the response of ERα− BC cells to TAM treatment by the restoration of ERα protein abundance.