Background
Breast cancer (BC) has become “the world’s largest cancer” shown in “2020 Global Cancer Report” and its incidence rate is growing rapidly recent year [
1]. Notably, triple-negative breast cancer (TNBC) accounting for 15–20% of all BC subtypes is the most malignant subtype with severe proliferative and aggressive phenotype [
2]. TNBC is characterized by deficient expression of estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor type 2 (Her-2) [
3]. Therefore, unlike the hormone receptor-positive and Her-2-positive subtypes, TNBC still lacks valid clinical biomarkers and therapeutic strategies [
4]. The effort to screen out reliable drug targets to improve the prognosis of TNBC patients is urgent.
circRNAs were covalently closed noncoding RNAs without 5’cap or polyadenylated tail. circRNAs are resistant to exonucleolytic degradation due to their closed-loop structure. Thus, circRNAs are more stable than their linear counterparts. Amount of studies have demonstrated that circRNAs can be detected in plasma, saliva, and exosomes [
5‐
8]. For example, Ju et al., [
9] developed a prognostic tool based on four circRNAs to improve the prognostic stratification of patients with curatively resected colon cancer at stage II/III. All these provide strong support for the practice of circRNAs as biomarkers or therapeutic targets in clinical treatment.
Majority of studies indicate that circRNAs mainly function as the “sponges” of microRNA/protein in various cancers [
7,
10,
11], including TNBC [
5,
12,
13]. Apart from these, both our group [
7,
14] and other teams [
6,
15] have demonstrated the reality of circRNA-encoded proteins, which suggests their potential oncogenic molecular actions in disease development. The study of the protein-coding ability of circRNAs provides novel insight for clarifying the etiology of tumors and is becoming a hot topic. For example, EIF6-224aa, encoded by circ-EIF6, promotes TNBC progression via stabilizing MYH9 and activating Wnt/beta-catenin pathway [
6]. HER2-102, encoded by circ-HER2, promotes TNBC progression via activating EGFR/HER3 signaling [
15]. However, whether other protein-coding circRNAs involve in TNBC and how they work in TNBC remain elusive. Considering the unique translatable functions of circRNAs in oncogenic pathways, a better understanding of the molecular action of circRNAs in TNBC could help us figure out creative drug targets for the clinical treatment. After mining the public database, we screened out a circRNA called circCAPG (circBase ID: hsa_circ_0055412), which is highly expressed in TNBC and positively correlated with both advanced clinical stage and poor prognosis. More importantly, we found that circCAPG encoded a 171-aa peptide (CAPG-171aa) which plays crucial roles in tumor growth and metastasis. Therefore, in this study, we aimed to elucidate the molecular action of CAPG-171aa, the mechanism of circCAPG biogenesis, and explore the potential therapeutic capacity of this novel polypeptide CAPG-171aa in the TNBC.
Materials and methods
Patients and tissue samples
The paraffin-embedded (FFPE) TNBC (n = 132) and adjacent tissues of TNBC (n = 40) used in this study were provided by TNBC patients undergoing surgical resection at the Chinese PLA General Hospital. Clinical data was obtained from medical records and follow-up information was derived from telephone consultations or death certificates. The whole study was authorized by the Ethics Committee of the Chinese PLA General Hospital and all research complied with the principles of the Declaration of Helsinki. The study was approved by the Ethics Committee of Chinese PLA General Hospital and all the patients had signed the informed consent.
circRNA expression data in TNBC (GEO: GSE113230, GSE101123) was obtained from the NCBI GEO database and annotated in circBase. TNBC mRNA data obtained from the RTCGA package, EdgeR, and DEseq2 package (version 3.12.1) was used to detect differentially expressed circRNAs or mRNAs. Multiple changes > 1.3 and P values < 0.05 were considered as significant differences. The ggboxplot, ggplot2 and pheatmap packages in R were used to display boxplot, volcano, and heatmaps.
Cell culture
All cell lines were purchased from Procell Life Science & Technology Co., Ltd (Wuhan, China) and characterized by DNA fingerprinting and passaged < 6 months. MDA-MB-231 and MDA-MB-468 cells were grown in Leibovitz’s L-15 Medium (KeyGEN, KGM41300N-500) supplemented with 10% fetal bovine serum (FBS, Gibco, USA). HEK293T, Hs578T, MCF7, and T47D cells were cultured with Dulbecco’s modified Eagle’s medium (DMEM, Gibco, USA) supplemented with 10% FBS. ZR-75-1 cells were cultured with Minimum Essential Medium (KeyGEN, KGM41500N-500) supplemented with 10% FBS. MCF-10 A cells were grown in DMEM/F12 medium (Gibco, USA) supplemented with 5% horse serum (Solarbio, China) and growth supplements. All cell lines were cultured in 5% CO2 at 37 °C in a humidified atmosphere.
Human TNBC organoid model
Fresh specimens of TNBC tissues were cut into ~ 1mm3 fragments on ice and digested for 40 min at 37 ℃ with 220 rpm. The digestion medium was added DMEM/F12 supplemented with Y-27,632 (10 µM), primocin, and collagenase II (1 mg/mL). The tissues were terminated by FBS after digestion. The suspension liquid was filtered with a 100 μm filter strainer and centrifuged at 450 g for 5 min at 8℃. Red blood cells in the cell pellet are lysed using red blood cell lysate and centrifuged at 450 g for 5 min at 8 ℃. Subsequently, wash the pellet three times using D-BSA. Resuspended with organoid medium and homogeneously seeded in microwell molds.
Reverse transcriptase PCR (RT-PCR) and quantitative real-time PCR (qRT-PCR)
Total RNA was extracted from TNBC cell lines and tissue specimens with Trizol (Yeasen, 19211ES60) and a total of 1.5 µg/sample of RNA was used for cDNA synthesis with M-MLV RT Kit with gDNA Clean for qPCR (Accurate Biotechnology(Hunan)Co., Ltd, AG11705) and TransScript® Fly First-Strand cDNA Synthesis SuperMix for RT-PCR (TransGen Biotech, AF301) following the procedure of 42 °C for 90 min and 95 °C for 5 min. RT-PCR and qRT-PCR were performed with 2×Taq PCR StarMix (GENSTAR, A012) and 2X SYBR Green qPCR Master Mix (APExBIO, K1070), respectively. The relative expression levels of target genes were calculated following the formula of 2
−ΔΔCt. All primers were synthesized by Tsingke Biological Technology (Tsingke, China). The primer sequences were listed in Table
S1.
Small interfering RNA (siRNA) and plasmid transfection
The siRNA specifically targeting STK38, EIF4A3, RBM38, HNRNPL, and circCAPG, as well as non-specific si-control were synthesized at RiboBio (Guangzhou, China). Two specific shRNAs, sh-circCAPG and sh-scramble, were cloned into PLKO.1-TRC plasmid (Tsingke, China), respectively, to silence the circCAPG, and sh-scramble works as a negative control.
To construct the circCAPG expression vector, the sequence of circCAPG was amplified from HEK293T cDNA using KeyPo Master Mix (Vazyme, PK511) and cloned into circular RNA expression plasmid PLO5-ciR (Geneseed, China). The circCAPG-flag plasmid was derived by inserting the flag sequence immediately to the upstream stop codon of the putative open reading frame (ORF) of circCAPG. The mutant of the internal ribosome entry site (IRES) of circCAPG (the IRES was deleted) was cloned into PLO5-ciR plasmid to generate circCAPG-flag-IRES-mut. To construct a circCAPG vector containing flanking intron sequence, the genomic region of circCAPG with its flanking introns was synthesized by Tsingke Biological Technology (Tsingke, China) and inserted into the pCDNA3.1 vector. Full length of UBA52, PRPF31, ATP5A1, TUFM, STK38, GTF21, PRPFKB3, UQCRC2, PPM1B, KHDC1, SNRPA1, NPM1, NCBP2, LSM5 and SLU7 was amplified from HEK293T cDNA using 2×TransStart® FastPfu PCR SuperMix (TransGen Biotech, AS221) and cloned into pCDNA3.1 vector. To assay IRES activity of circCAPG, the promoter region of F-Luc in the psiCHECK2 vector was deleted and the IRES sequences of circCAPG were cloned behind R-Luc. All plasmids were extracted using an endotoxin-free plasmid extraction kit (Shandong Sparkjade Biotechnology Co., Ltd., AD0105). Primer sequences were listed in Table
S1. All expression vector has been verified by sequencing (Sangon Biotech, China). All transfection experiments were conducted with Lipofectamine™ 3000 Transfection Reagent (Invitrogen, USA) following the manufacturer’s instructions.
Actinomycin D (ACTD) assay
For the ACTD assay, the extracted total RNAs were treated with 1 µg/mL actinomycin D (Sigma-Aldrich, USA) against new RNA synthesis for 0, 2, 4, 8, 12 and 24 h, respectively.
Nucleocytoplasmic separation
The RNA of nuclear and cytoplasmic fractions of TNBC cells were extracted using a Paris kit (Invitrogen, USA) according to the manufacturer’s instructions.
Virus production, infection, and puromycin selection
Vectors (sh-circCAPG-1, sh-circCAPG-2, sh-scramble, PLO5-ciR, circCAPG, circCAPG-flag, circCAPG-flag-IRES-mut, circCAPG-flag-ATG-mut) were transfected into 293T cells with Lipofectamine™ 3000 based on the manufacturer’s instructions. Infectious supernatant was collected twice after 48 and 72 h and then filtered through 0.45 μm filters (Hangzhou Cobetter Filtration Equipment Co., Ltd). MDA-MB-231 and MDA-MB-468 cells were infected with the appropriate amount of the virus recombinant lentivirus for 48 h and then selected by 1 and 2 µg/mL puromycin (Sigma-Aldrich, USA), respectively, for 72 h.
Cell proliferation, colony formation, and wound-healing assay
MDA-MB-231 and MDA-MB-468 cells (1 × 103) were seeded into 96-well plates and cell viability was determined by absorbance at 450 nm after 0, 24, 48, 72 and 96 h with Cell Counting Kit-8 (Elabscience, E-CK-A362). For colony formation assay, cells were seeded in 6-well plates (SORFA Life Science) at a density of 3 × 103 cells per well and incubated in L15 containing 10% FBS for two weeks. Then, the cells were washed with ice-cold PBS twice, fixed with formaldehyde, and stained with a crystal violet staining solution (Beyotime, China). For the wound-healing assay, cells were seeded in a 24-well plate (SORFA Life Science) with a serum-free medium. Then, a sterile 10 µL plastic pipette tip (CellProBio) was used to scratch through the single-cell layer, and images were captured after 0 and 48 h at the same place with a microscope (Nikon, Japan).
Migration and invasion assays
Migration and invasion assays were performed using the Transwell system (Nest, 723,001). For the 48 h migration assay, MDA-MB-231 and MDA-MB-468 cells (1 × 105) were seeded in small chambers (Nest, 723,001) with a serum-free medium, and a 700 mL medium with 20% FBS was added to the bottom wells. Likewise, for invasion assay, 1 × 105 cells were seeded in Matrigel-coated chambers with serum-free medium and cultured for 48 h along with L15 supplemented with 20% FBS in the bottom wells. The images were captured by microscope (Nikon, Japan).
Immunofluorescence assay
For immunofluorescence assay, TNBC cells were collected and fixed with ice cold acetone. The cells were fixed with 4% paraformaldehyde for 30 min at room temperature and then treated with 0.2% Trition X-100 for 10 min. Cells were incubated with 10% goat serum for 60min at 37°C and incubated with the KI67 antibodies at 4°C overnight. After washing with PBS, cells were incubated with corresponding fluorescent secondary antibody and counterstained for nuclei using 4’, 6-diamidino-2-phenylindole (DAPI) (Solarbio, China) for 15 min at room temperature.
Luciferase reporter assay
For IRES activity analysis, 293T cells were transfected with psiCHECK2 containing circCAPG IRES. After 72 h transfection, the luciferase activity of circCAPG IRES was assessed using the dual-luciferase reporter kit (TransGene, China) based on the manufacturer’s instructions.
RIP assays
RIP assay was carried out with the PureBinding®RNA Immunoprecipitation Kit (GENESEED, P0101), and the anti-SLU7 antibody (Proteintech, USA) and IgG antibody (Millipore, USA) were utilized according to the manufacturer’s instructions. Briefly, protein A/G magnetic beads were incubated with anti-SLU7 or IgG, and negative control antibodies, respectively, and then incubated with cells. Subsequently, RNA was co-precipitated and extracted, and finally quantitated by RT-qPCR.
Ubiquitination assay
MDA-MB-231 and MDA-MB-468 cells were treated with 10 µg/mL MG132, respectively, (Solarbio, IM0310) for 12 h. Cell lysates were obtained using PierceTM IP lysis buffer (Thermo Fisher Scientific, USA) supplemented with a cocktail (Thermo Fisher Scientific, USA) and then incubated with anti-MEKK2 antibody (Proteintech, USA) and Protein A/G beads overnight at 4℃. Finally, protein A/G beads with the bound immunoprecipitates were collected and analyzed using Western blotting.
Cycloheximide (CHX) chase assay
The MDA-MB-231 and MDA-MB-468 cells in each group were treated with 10 µg/mL CHX (Sigma-Aldrich, USA) for 0, 1, 2, 4, and 8 h. The expression level of MEKK2 protein was determined by Western blotting.
Co-immunoprecipitation
Immunoprecipitation and co-immunoprecipitation were carried out with Pierce Classic Magnetic IP/ Co-IP Kit (TermoFisher Scientific, USA) according to the manufacturer’s instructions. Cells were washed with ice-cold PBS and lysed in cold PierceTM IP lysis buffer. Then, the supernatant was collected and incubated with a specific IP antibody or negative control IgG at 4 °C overnight. Next, the co-immunoprecipitates were separated and analyzed by SDS–PAGE, MS, or Western blotting. VeriBlot (Abcam, USA) was used to avoid the detection of heavy and light chains. Finally, protein A/G beads with the bound immunoprecipitates were collected and analyzed using Western blotting.
Protein isolation and western blotting
Cells were lysed with protein lysis buffer (Solarbio, R0020) containing cocktail and then separated by SDS-PAGE gels, and transferred onto the polyvinylidene fluoride (PVDF) membrane (Millipore, USA). After being blocked with 5% nonfat milk, the membrane was incubated with a specific primary antibody at room temperature for 1 h, and then washed and incubated with a secondary antibody (Proteintech, USA) for 1 h. Western blotting was performed by anti-MEKK2 (Proteintech, 55106-1-AP), MEK1/2 (ABclonal, A4868), SMURF1 (Bioss, bs-9391R), STK38 (Huabio, ER60144), ERK1/2 (Proteintech, 11257-1-AP), UB (Cell Signaling Technology, #3936), Phospho-MEK1/2 (Ser221) (Cell Signaling Technology, #2338), Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (Cell Signaling Technology, #8544), CAPG (Bioss, bsm-60,451 M), CDK4 (AtaGenix Laboratories Co., Ltd.(Wuhan), Wuhan, PR China, PA9437H), Cyclin D1 (Yeasen, 30026ES50), BCL2 (Shanghai hengyuan biological technology co., LTD, A-01203), CAPG (Proteintech, 10194-1-AP), FLAG (SMART, /smart-lifesciences, SLAB0101), HA (Yeasen, 30702ES20), HIS (Yeasen, 30401ES10). β-ACTIN, GAPDH, or TUBULIN (Proteintech, USA) were used as loading controls. All antibodies were diluted as required by the manufacturer (New Cell & Molecular Biotech, WB100D). The antibodies used in this study are listed in Table
S2.
In vivo xenograft assay
The 4–5 weeks-old female BALB/c nude (nu/nu) mice purchased from SPF (Beijing) Biotechnology Co., Ltd. (Beijing, China) were housed under specific pathogen-free (SFP) conditions. Approximately 1 × 107 / mL MDA-MB-231 cells in PBS supplemented with Matrigel (Yeasen, 40183ES10) (1:1) were injected subcutaneously into the BALB/c nude (nu/nu) mice after one week. All animal studies were approved by the Ethics Committee for Animal Experimentation of China Agricultural University.
Statistical analysis
Two-tailed Student’s t-test was performed to calculate statistical significance with GraphPad Prism 8 (version 8.0.1, USA). Asterisks denote statistical significance (*P < 0.05, **P < 0.01, ***P < 0.001) and ns indicate no significance among groups. All data were validated in at least three independent experiments and represented as mean ± standard error of the mean (SEM).
Discussion
TNBC is a heterogeneous and fatal disease in women with limited treatment options [
24]. Therefore, the discovery of a new targeted therapy strategy and biomarker is urgent. Herein, we found a novel circRNA, circCAPG, which was formed by the exons 6–8 of the CAPG gene and has not been studied so far. In the present study, we found that the expression of circCAPG was upregulated in TNBC, and was tightly associated with tumor invasion, metastasis and poor overall survival. We further demonstrated that circCAPG encodes a novel polypeptide named CAPG-171aa driven by an active IRES. In vitro, overexpression studies confirmed the tumor promoting role of CAPG-171aa rather than circCAPG in the progression of TNBC. Mechanism studies revealed that CAPG-171aa inhibited the ubiquitination of MEKK2 by interacting with STK38 and thereby activated the downstream MEK1/2-ERK1/2 signaling pathway. Most importantly, we found that the RNA-binding protein
SLU7 regulates the generation of circCAPG by binding the ICSs of the circRNA’s pre-mRNA. Furthermore, we demonstrated that the expression levels were positively correlated with overall survival of the TNBC in a 10-year follow-up study with 132 patients, the area under the curve of receiver operating characteristic was 0.8723 with 100% specificity and 80% sensitivity. Altogether, results demonstrated that circCAPG might serve as a novel biomarker for diagnosis and prognosis, and a potential therapeutic target for TNBC as well.
The vast majority of circRNA studies were focused on its function as ‘microRNA sponges’ not only in TNBC but also in some other diseases [
25,
26], whereas only few studies from our group [
7,
14] or others reported its translational function [
13,
15]. Therefore, we first explored whether circCAPG regulates TNBC progression through sponge function. However, we noticed that circCAPG barely interacts with miRNAs in sponge manner in TNBC. With the development of biotechnology and online bio-software, circRNAs driven by certain IRES [
6,
7] or N6-methyladenosine (m6A) [
27] with potential protein-coding functions could be predicted and some have been reported recently. Our previous work also proved the existence of protein-encoded circRNAs [
7,
14]. All this prompted us to explore the possibility and the molecular effects of circCAPG-derived translative proteins or polypeptides on the malignant behaviors of TNBC. After evaluating the sequence of circCAPG at the circRNADb website [
17], both putative IRES structure and ORF of circCAPG were predicted and obtained, suggesting its potential protein-coding capacity. Since IRES is a key promoter for circRNA translation, a circCAPG-flag-IRES-mut plasmid was constructed to test the expression of CAPG-171aa in comparison with the putative circCAPG-IRES construct. The results demonstrated that the sequence of circCAPG owns the IRES-derived translational initiation capability and polypeptide-coding ability. By a series of experiments, we further confirmed that this circRNA-encoded polypeptide CAPG-171aa rather than circCAPG itself promoted the proliferation and metastasis of TNBC cells. Although our results showed the tumor promoting effect of CAPG-171aa on TNBC in BALB/c nude mice and TNBC cell lines, and its enhancing role in the migration and invasion of TNBC cells in vitro, a serial of suitable TNBC-metastasis models to further confirm whether CAPG-171aa-induced TNBC metastatic outcome and tumorigenesis in vivo was missing in this study. To figure out the molecular action of how CAPG-171aa in TNBC metastasis and tumorigenesis would help better understanding the mechanism of TNBC metastasis, which is also our task and is carried out in our ongoing study. In addition, no observed malignant changes in a polypeptide CAPG-171aa stable overexpressed MCF-10A normal control cell line is still an open question (Fig.
S4I and J). The reason for this might be due to the complexity and the context of tumorigenesis of TNBC, that massive oncogenic factors and the micro environment should be co-operated together at the same time to prompt the development of TNBC, rather than a single one can orchestra the carcinogenesis, which is also another task for our future study.
In our study, STK38 was identified as an important binding partner of CAPG-171aa proven by MS analysis and IP experiments. Interestingly, STK38 does not interact with maternal CAPG protein. This further reflects that CAPG-171aa involved in the progression of TNBC as a novel oncoprotein. STK38 is a serine-threonine protein kinase that belongs to a subfamily of the AGC kinase family [
18]. It can regulate the protein levels of MYC in a kinase activity-dependent manner [
20] and promote SMURF1-mediated polyubiquitination of MEKK2 [
21]. Consistent with the study by Ji et al. [
21], we found that STK38 inhibits the protein levels of MEKK2 through SMURF1-mediated polyubiquitination and degradation. Most importantly, we found that interaction between CAPG-171aa and STK38 could inhibit the ubiquitination of MEKK2 protein and then activate the MEK1/2-ERK1/2 pathway, which in turn promotes the progression of TNBC.
Notably, investigating the biogenesis mechanisms of circRNAs could help us better understand how circRNAs were generated and accumulated specifically in TNBC progression [
28]. However, only a few literatures reported the biogenesis mechanisms of circRNA [
27]. Recently, Li et al. reported that ICS and RBP were critical for the formation of circRNAs [
11]. Pairing between ICSs at different introns was often thought to be able to bring the distal splice sites closer together and thus promote the biogenesis of circRNA [
29‐
32]. Zheng et al. [
32] reported that the inverted repeat Alu elements (IRAlus) of the intron flanking homeodomain interacting protein kinase 3 (HIPK3) Exon2 mediated the formation of circHIPK3 which plays an important role in the etiology and carcinogenesis. Except for ICS, some RNA-binding proteins (RBPs) also engaged in the biogenesis of circRNA. For example, Zeng et al. [
33] reported that the splicing factor epithelial splicing regulatory protein 1 (ESRP1) can increase circANKS1B production. However, it is still unclear how RBPs compete or cooperate with the inverted-repeat Alu elements to balance the production of circRNAs. In our study, we found that the expression of circCAPG was negatively correlated with the expression of
SLU7 in TNBC tissues (n = 132). Furthermore, overexpression of
SLU7 can reduce the expression of circCAPG. To examine the SLU7-induced downregulation of circCAPG in TNBC, we carried out a series of experiments and found that the formation of circCAPG requires the binding between
SLU7 and ICS (AluSp and AluSz). In addition, our in vitro studies further demonstrated that
SLU7 OE suppressed the malignant progression of TNBC, while the subsequent supplement of circCAPG OE could not fully promote the malignant phenotype of TNBC. One RBP can regulate many circRNAs, which means
SLU7 could also interfere with the expressions of many other circrRNAs except for circCAPG. For example, the knockdown of
HNRNPL can lead to 139 significantly upregulated and 93 downregulated circRNAs [
34]. The issue that we have not fully addressed is that although
SLU7 regulates the formation of circCAPG in TNBC, the cognate mRNA of the circRNA seems to be unimpacted, which is inconsistent with the previous research [
35]. This might be related to the special role of circCAPG in TNBC. Further studies are required to fully elucidate the transcriptional regulation as well as the transcriptional output on circRNAs.
By Kaplan-Meier survival analysis, we found that TNBC patients who had higher circCAPG expression also showed worse prognosis. The receiver operating characteristic (ROC) curve of circCAPG showed that circCAPG yields an excellent diagnostic ability. This suggests that circCAPG has the potential to work as a biomarker in TNBC patients. Unfortunately, we did not get a clear positive band of CAPG-171aa in the serum of TNBC patients by Western blotting with anti-CAPG antibody. One reason for this failure might be due to the lack of a CAPG-171aa-specific antibody. To increase the detection specificity and sensitivity for CAPG-171aa, we are going to make the antibody for ELISA assay in our future study. We do believe that CAPG-17aa determination in a larger scale in TNBC patients would be necessary and enable our better understanding for its diagnostic value.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.