Introduction
Breast cancer is the deadliest form of carcinoma affecting women, with nearly a quarter-million cases diagnosed in 2016 [
1]. Although there are effective treatments against some types of breast carcinoma, such as those for subtypes with abnormal overexpression of the HER2/Neu oncogene, the majority breast cancers remain incurable. Breast cancer stem cells (BCSCs) possess self-renewal and differentiation capabilities, leading to tumor recurrence, metastasis and therapeutic resistance [
2,
3]. CD44
+/CD24
− or aldehyde dehydrogenase1 (ALDH1) phenotypes are efficient in the identification of BCSCs from breast cancer populations. However, there is a small overlap between CD44
+/CD24
− and ALDH1 stem phenotypes, as well as less stem markers in differentiation of different breast cancer subtypes [
4]. Therefore, it is necessary to identify more discriminatory biomarkers of distinct molecular subtypes for the isolation and identification of the BCSCs subpopulation.
EMT describes the process by which epithelial cells detach from neighboring cells and are transferred to other tissue sites via dissolution of basement membrane and passage through the extra-cellular matrix [
5,
6]. EMT could also facilitate the generation of cancer stem cells from more differentiated cancer cells [
7]. The EMT process during breast carcinogenesis is considered to be controlled by a series of signaling pathways, including Notch, Wnt/β-catenin and Hedgehog [
8,
9]. The Hedgehog (Hh) pathway is responsible for the maintenance of stem cells and EMT, which can contribute to the evolution of breast cancer [
10].
Ubiquitination describes a highly conserved and reversible modification process of protein degradation, which is involved in nearly all aspects of cell biology [
11]. Deubiquitinating enzymes (DUBs) can prevent ubiquitin-mediated degradation of target proteins [
12]. Importantly, dysregulated DUBs expression is frequently associated with the tumorigenesis process, specifically cell self-renewal, apoptosis and EMT [
13,
14]. It has been confirmed that DUBs are essential for the regulation of stem cell-related markers and controlling various steps of metastatic progression, including invasion, dissemination and eventual metastasis to distant organs [
14,
15]. Ubiquitin specific peptidase 37 (USP37), a novel DUB, is a member of ubiquitin-specific processing proteases family. Human USP37, localized mainly in the cytoplasm, is composed of 979 amino acids harboring three ubiquitin-interacting motifs between the Cys box and His box of the primary sequence [
16,
17]. The function of USP37 was initially identified as a potent regulator of the cell cycle where it could accelerate the G1/S transition with exceptional high expression [
18,
19]. Previous studies have found that USP37 could regulate the stem cell-related marker SOX2 by binding with its promoter region at the transcriptional level [
20]. Pan et al. reported that high levels of USP37 gene expression in lung cancer promoted cell viability as well as the Warburg effect via deubiquitination and stabilization of pluripotent factor c-Myc protein [
21]. These advances implicated that USP37 gene may be associated with the stemness of tumor cells facilitating cancer progression. Recent studies indicate that USP37 is a potential factor involved in breast cancer progression [
16]. However, the biological function of USP37 in the direct regulation of BCSCs and EMT remains unexplored.
In this study, we found that USP37 expression was upregulated in breast cancer tissues compared with surrounding tissues and its overexpression was significantly correlated with increased rates of mortality. We demonstrated that USP37 was highly expressed in BCSCs. The knockdown of USP37 could inhibit the stemness, cell invasion and EMT via downregulation of Hedgehog pathway. USP37 also interacted with and stabilized glioma-associated oncogene 1 (Gli-1) protein. Additionally, USP37 knockdown enhanced the sensitivity of breast cancer cells to cisplatin in vitro and in vivo. We postulate that USP37 may represent a novel molecular target for breast cancer treatment.
Materials and methods
Gene data extracted from invasive breast tumor samples was obtained from the TCGA Data Portal. Results of analysis were used to create figures via GraphPad Prism. According to PAM50 gene expression signature, information about four breast cancer subtypes (basal-like, Luminal A, Luminal B, and enriched Her-2) were classified. A scatter plot diagram, where each dot indicated an individual sample, was synthesized using the All Complete Tumors of Breast Invasive Cancer dataset [
22,
23]. For survival analysis, clinical data related to invasive breast carcinoma were downloaded from the TCGA database [
22]. Kaplan–Meier curves were analyzed by GraphPad Prism.
Gene set enrichment analysis (GSEA) was utilized to evaluate potentially biological mechanism associated with USP37 mRNA expression levels in the TCGA breast cancer samples. GSEA software which was obtained from the Broad Institute draws the result pictures automatically.
Cell culture and animals
Human normal breast epithelial cells (MCF-10A) and breast cancer cell lines (MCF-7, MDA-MB-231, BT549 and T47D) were obtained from the Laboratory of Pathology at Dalian, Medical University. The composite culture of MCF-10A cells used DMEM/F12 media supplemented with insulin (10 μg/mL), cholera toxin (100 ng/mL) (Sigma-Aldrich), EGF (20 ng/mL) (R&D systems, Wiesbaden, Germany), hydrocortisone (500 ng/mL), and L-glutamine (Invitrogen, Gaithersburg, MD, USA). MCF-7 cells were cultured in DMEM/HIGH GLUCOSE (Hyclone, Logan, UT, USA) with 10% FBS. MDA-MB-231 cells were cultured in MEM Alpha Modification medium (Hyclone) supplemented with 10% FBS. In addition, BT549 and T-47D cells were cultured in RPMI-1640 medium (Hyclone) supplemented with 10% FBS. Growth condition for all cells was 37 °C at 5% concentration of CO2.
BALB/c Nude mice (6 to 8 weeks old) were purchased from Vital River Laboratory Animal Technology Company (Beijing, China) and reared according to the Animal Care and Use Committee of Dalian Medical University.
Plasmids, antibodies and reagents
The pEZ-M35-USP37 plasmid was created by FulenGen Co. (Guangzhou, China). Anti-Gli-1 and anti-Smoothened antibodies were obtained from Abcam (Cambridge, MA, USA); all other antibodies were obtained from Proteintech Group, Inc. (Wuhan, China). Secondary antibodies were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Cycloheximide (CHX), MG132, cisplatin and purmorphamine reagents were obtained from MedChemExpress (Shanghai, China).
siRNA, shRNA, lentivirus
The design and synthesis of siRNAs were completed by RiboBio company (Guangzhou, China), and transfection of siRNAs against USP37 was performed with Lipofectamine 2000. MCF-7 and MDA-MB-231 cells were transfected with USP37 siRNAs or a negative control (NC) siRNA. Lentivirus vectors including short hairpin RNA against USP37 (shUSP37#2) and the negative control (shScramble) were purchased from GenePharma Company (Shanghai, China). The transfected cells were treated with puromycin (Clontech, USA) for selection. To overexpress USP37 in MCF-7 cells, the pEZ-M35-USP37 plasmid (2 μg) were infected into MCF-7 cells via Lipofectamine 2000. Expression was confirmed by RT-qPCR and western blotting. The siRNA or shRNA sequences were listed in Additional file
1: Table S1.
BCSCs isolation
Here, 1 × 107 MCF-7 cells were grown with primary antibody against CD24 following the manufacturer’s protocol for the CD24 Microbead Kit (Miltenyi Biotec, Bergisch Gladbach, Germany). Labeled cells were then incubated with goat anti-mouse IgG microBeads (Miltenyi Biotec) and magnetically separated with MiniMACS columns (Miltenyi Biotec). Acquired CD24− cells were incubated with CD44 microbeads (Miltenyi Biotec) at 4 °C for 15 mins. Cells were again washed and magnetically separated.
Colony formation assay of breast cancer cells was carried out by plating infected cells at a density of 1000 cells/well in a 6-well plate and then different concentrations of cisplatin were added. After 2 weeks of incubation, cells were washed three times with phosphate buffered saline, fixed with methanol and stained with 1% crystal violet. Visible colonies were stained violet and counted for data analysis. The detection of cell viability was performed in accordance with the Cell Counting Kit 8 assay (Dojindo Laboratories, Kumamoto, Japan).
Cell invasion assay
Cells were similarly cultured in MEM Alpha Modification medium for 24 hours. Briefly, 1 × 105 cells were seeded without serum into 24-well insert Transwell chambers (8 μm pore size, Corning, USA) and pretreated with Matrigel (BD, Bioscience, San Jose, CA, USA). Medium supplemented with 20% serum was added into the lower chamber. After 12–16 hours of seeding, cotton swabs were used to clean the upper cells. The cells on the other side of the membrane were stained with 1% crystal violet. A randomly selected area was counted with an optical microscope.
Wound healing assay
Cells were incubated in 12-well plates. When cellular density reached nearly 100%, the cell monolayer was wounded with a 200 μl micro-pipette tip. The wound areas were washed three times with phosphate buffered saline (PBS). Then the medium was changed to MEM Alpha Modification without FBS. The wound areas were micrographed at 0, 12, 24 and 48 h. All assays were performed in triplicate.
Western blotting
Cold PBS was used to wash the transfected cells. Then cells were harvested and treated with a mixture of lysis and 1 × RIPA buffers (Sigma Chemicals, St. Louis, MO, USA). Protein concentrations were estimated using the Easy IIProtein Quantitative Kit (Transgen, China). Thirty to forty microgram of each protein lysate was separated by 8–12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Samples were transferred to PVDF membranes (Millipore, Billerica, MA, USA). The PVDF membranes were blocked with 5% non-fat milk dissolved in TBST (TBS and 0.01% Tween-20) for 2 h at room temperature, then incubated with the specific antibodies solution overnight at 4 °C. Finally specific antibodies were diluted as follows: USP37, 1:500; Gli-1, 1:500; Smoothened, 1:500; E-cadherin, 1:1000; Snail1, 1:500; N-cadherin, 1:1000; Vimentin, 1:500 and GAPDH (loading control), 1:1000. Finally, the transfer membranes were incubated with anti-IgG secondary antibodies (1:16000 in TBST) for 1 h at 37 °C. The protein band images were captured with a ODYSSEY infrared imaging system.
RNA extraction and real-time quantitative PCR analysis
Total RNA was extracted from cultured cells with Trizol reagent (Transgen, China). In total, 1 μg of total RNA was reverse-transcribed with the All-in-one First-Strand cDNA Synthesis SuperMiX kit (Transgen, China). The mRNA expression levels of USP37, ALDH1, CD24 and CD44 were quantified by real-time PCR. RT-qPCR was analyzed with the iCycler™ Real Time System and a SYBR Premix EX Tag Master mixture kit (Transgen, China) according to the manufacturer’s instructions. The relative expression levels of mRNA were evaluated by using the 2
−ΔΔCt method. Primer sequences are listed in Additional file
1: Table S1.
Immunofluorescence analysis
Cells were grown to 3 × 103 in 24-well plates at 37 °C and fixed with 100% methanol. Next, cells were cleaned two to three times using PBST (PBS and 0.5% Triton X-100). After blocked with 3% BSA for 2 h at 4 °C, cells were incubated with the following primary antibodies (San Ying Biotechnology, China) overnight at 4 °C: USP37, 1:100; Gli-1, 1:100; Flag, 1:100; E-cadherin, 1:100 and N-cadherin, 1:100. Subsequently, cells were incubated with fluorescence conjugated secondary antibody (Sigma, 1:200) for 1 h and stained the cell nuclei with DAPI (Beyotime, China). Images were viewed with CKX41 Inverted Microscope (Olympus, Japan).
CHX chase assay and co-immunoprecipitation (co-IP) assay
For the CHX chase assay, cells were treated with CHX (50 μg/ml) and harvested at the indicated timepoints. Treated cells were lysed, and the lysates were analyzed by western blotting with anti-USP37 or anti-GAPDH antibodies. For the Co-IP assay, 8 × 106 MCF-7 cells were harvested and lysed in IP lysis buffer containing protease inhibitor. The experiment based on manufacturer’s protocols (IP Kit, Proteintech Group) was performed.
Cells (MCF-7, MDA-MB-231, BT549 or T47D) were inoculated into ultra-low attachment 6-well plates (Corning, New York, USA) at a density of 1000 cells/well, and then grown in DMEM/F12 supplemented with B27 (1:50, Invitrogen), 20 ng/ml human recombinant EGF (Sigma-Aldrich, St. Louis, Missouri, USA), 20 ng/ml bFGF (Sigma-Aldrich, St. Louis, Missouri, USA), 4 μg/ml heparin (Sigma-Aldrich, St. Louis, Missouri, USA), and 5 μg/ml insulin (Sigma-Aldrich, St. Louis, Missouri, USA) for 14 days. Cell colonies larger than 60 μm in diameter were counted under an inverted microscope (Olympus Corporation, Tokyo, Japan). After cultivation for 28 days, spheroid cells were collected for western blotting.
Mouse xenograft mode
Animal experiments and procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (NIH). Twenty BALB/C nude mice were randomly divided into four groups. The negative control group (shScramble) and shUSP37#2-transfected MCF-7 cells (5 × 105) were resuspended in 100 μl PBS and injected into mammary fat pads. When the size of tumor reached approximately 100 mm3, animals were randomly treated with cisplatin (2 mg/kg) or 0.9% saline. Mice were then retreated with their assigned treatment once every 2 days. The tumor sizes were measured with a Vernier caliper and recorded every other day. The tumor volume was calculated using the formula: tumor volume [mm]3 = (length [mm]) (width [mm])2 × 0.5. After the inoculation of tumors for 3 weeks, xenografted tumors were excised from sacrificed mice then analyzed by immunohistochemistry and western blotting.
Immunohistochemical analysis
Tumor tissues were obtained from mouse xenograft mode. Paraffin-embedded tissue was cut into 5 μm thick slices that were fixed onto glass slides. All human breast cancer tissue arrays were purchased from Shanghai Outdo Biotech (Shanghai, China), including 60 cases of cancer tissues and surrounding tissues (HBre-Duc060CS-03; HBre-Duc060CS-04). These tissue sections were immunostained with corresponding antibodies, and then deparaffinized in xylene and rehydrated with ethanol. Tissue sections were preincubated with 10% normal goat serum, followed by incubation with primary antibody solution overnight at 4 °C. After washing with PBS, slides were incubated with the secondary antibody at 37 °C for 10 min, then cleaned with cold PBS and treated with peroxidase conjugated-biotin streptavidin complex for 10 min. Finally stains were examined with 3,3′-diaminobenzidine and hematoxylin. As previously described [
23], the immunostained tissues were scored by multiplying the intensity (0–3) and extent (0–100) of staining.
Ethical approval
This study was conducted with the approval of the Ethical Committee and Institutional Review Board of Dalian Medical University.
Statistical analysis
Statistical analysis was performed with SPSS software version 11.0. Data are expressed as mean ± SD. Differences between two groups were evaluated by Student’s t-test. One-way ANOVA was used when comparing multiple groups. P < 0.05 was considered statistically significant. Clinical data analysis of survival and relevant correlations were performed with GraphPad Prism.
Discussion
DUBs have been shown to participate in ubiquitin cleaving from ubiquitin conjugated protein substrates [
29,
30]. USP37, a novel DUB, contains an insert between its catalytic lobe and its ubiquitin-binding lobe while its function could prevent 14-3-3γ degradation, which might contribute to malignant transformation by MAPK signaling [
31,
32]. Subsequently, it was verified that USP37 expression modulated the oncogenic fusion protein PLZF/RARA stability and cell transformation potential in PLZF/RARA-associated acute promyelocytic leukemia [
33]. Recently, clinicopathological analysis confirmed that USP37 was a poor prognostic factor in breast cancer [
16]. However, there was no direct evidence to identify the carcinogenic mechanism of the USP37 gene in breast cancer. In this study, we have demonstrated overexpression of USP37 in Luminal B subtype breast cancer was a predictor of poor outcomes in breast cancer. Supporting our results, previous clinicopathological analysis has demonstrated that overexpressed USP37 is considered to be a poor prognosis in breast cancer [
16]. Additionally, GSEA analysis on the TCGA dataset implicated that USP37 expression was positively associated with metastasis, cell growth and anti-apoptosis. Therefore, USP37 levels could potentially serve as a specific oncogene involved in breast carcinoma progression.
Cancer stem cells (CSCs) are endowed with stem cell properties, which include the ability to self-renew and differentiate by symmetrical or asymmetrical cell division. CSCs self-renewal of the cellular population and generation of progenitor cells have been shown to resist radiation and chemotherapy. Therefore, these cells are commonly deemed to be crucial target for cancer therapy [
34,
35]. It was verified that CSCs could be maintained in an undifferentiated status and induce tumor-sphere formation in defined serum-free medium [
36]. However, limited markers were utilized for the identification of breast cancer stem cells. While researchers have demonstrated that CD44 and ALDH1 are critical biomarkers to identify BCSCs from breast cancer populations [
37,
38], our data suggested that the USP37 gene was significantly associated with CSC properties, such as self-renewal, treatment resistance and EMT phenotype as well. Interestingly, elevated mRNA expression of USP37 was detected in CD24
−/CD44
+ cells and ALDH1
+ cells; the protein expression of USP37 was obviously elevated in spheres compared to adherent cells. Knockdown of USP37 suppressed mammospheres formation and inhibited cancer stem markers, such as ALDH1and OCT4 in breast cancer cells. It has been reported that USP37 participates in regulation of CSCs-related proteins, such as SOX2 and c-myc [
20,
21]. Our data further indicated that knockdown of USP37 by constructed lentiviral system weakened the stemness in MCF-7 and MDA-MB-231 cells via inhibiting the expression levels of ALDH1 and OCT4. CSCs are considered to be “bad seeds” due to their drug resistance caused by imbalanced pathway and epigenetics in cancer [
39]. As shown above, we detected that knockdown of USP37 in breast cancer cells could promote the sensitivity to cisplatin-induced cell death by the CCK-8 assay and colony formation assays. The Bax/Bcl-2 ratio is a commonly used method to determine whether intracellular apoptosis system is activated [
40]. We detected that USP37#2 shRNA combined with cisplatin treatment induced cell apoptosis with an underlying decrease in Bcl-2/Bax ratio. On the contrary, the upregulation of USP37 reversed this phenomenon. USP37 knockdown suppressed stemness and chemo-resistance, which may assist clinical oncologists in designing and testing novel therapeutic strategies. In summary, we suggest that USP37 gene expression confers the stemness and potentially acts as a critical marker of CSCs in breast cancer.
The process of EMT was involved in the acquisition of aggressive cellular traits, including motility, invasiveness and anti-apoptosis, resulting in the dissemination of cells and colonization in distance tissues [
6,
41,
42]. Here, we showed that USP37 siRNA treatment stimulated the expression of epithelial marker (E-cadherin) but decreased the expression of three known inducers of EMT (Snail1, N-cadherin and Vimentin) (Fig.
6a). Moreover, we found that knockdown of USP37 suppressed cell migration and invasion in breast cancer cells. On the contrary, upregulation of USP37 promoted EMT, migration and invasion. Accumulation evidence has indicated that tumor cells undergoing EMT process are endowed with the trait of cancer stem-like cells [
43], which further speculated USP37 as a CSC marker of breast cancer. In summary, our data showed that USP37 could regulate the migration, invasion, EMT of breast cancer cells.
The Hh pathway was found to be required for the maintenance of breast CSCs traits, tumor formation and the EMT [
10]. Normally, the ligand binding of Patched (Ptch1), a 12-pass transmembrane receptor of Sonic Hedgehog (SHH), can activate zinc finger transcription factor (Gli-1). It has been suggested that ectopic expression of Gli-1 upregulates expression of the transcription factor Snail1 accompanied with a decrease in E-cadherin, a characteristic of EMT [
24]. In fact, Snail expression promotes EMT via repressing E-cadherin [
44]. Hh signaling pathway modulates EMT indirectly via Snail [
45]. Therefore, it is a promising research to investigate whether USP37 promoted the activation of the Hh pathway.
Our findings suggested that the post transcriptional levels of Smo and Gli-1 were decreased in vitro and in vivo. Upregulation of USP37 could enhance protein expression levels of Smo and Gli-1. In order to understand the mechanism of USP37-induced EMT and stemness via the Hh pathway, we treated MCF-7 cells with PM. We found that activation of Hh signaling pathway was accompanied by elevated expression of USP37 gene as visualized by western blotting and immunofluorescence assay. Furthermore, the effect USP37 downregulation on CSCs traits including the formation of spheroids, BCSCs markers and cell invasion were impaired after PM treatment. Meanwhile, knockdown of USP37 reversed the effect of PM on the EMT markers. These data confirm that USP37 mediates breast cancer stem-like properties, cell invasion and EMT via the Hh pathway. Gli-1 acts exclusively as a transcriptional activator in the Hh pathway and signal outcomes are determined by the balance of activated and inhibitive Gli proteins [
46]. Remarkably, we found that USP37 could regulate and stabilize the protein level of Gli-1 (Additional file
2 : Figure S1). As a deubiquitinase, USP37 is involved in the regulation of multiple proteins by deubiquitination, including P27, Cdt1, PLZF/RARA and 14-3-3γ [
31,
33,
47,
48]. Future work should aim to determine whether USP37 could stabilize Gli-1 through deubiquitination.