miR‐335 modulates Numb alternative splicing via targeting RBM10 in endometrial cancer

Abstract

Numb is a conserved protein plays important roles in the development of cancer. Two Numb isoforms have been found produced by alternative splicing and play contrast roles in regulating cellular functions. It is reported that the expression of Numb long isoform (Numb‐L) was increased in various kinds of cancers, but in endometrial cancer, the condition is still unknown. The level of two Numb transcripts and protein isoforms were detected by semiquantitative polymerase chain reaction and immunoblotting in 47 paired endometrial tumor and adjacent non‐tumor control tissues. The level of three alternative splicing related proteins: RBM5, RBM6, and RBM10 was determined by immunoblotting. MiRNAs targeting RBM10 were predicted by bioinformatics tools and their interaction with RBM10 was confirmed by luciferase assay and immunoblotting. The function of miR‐335 in endometrial cancer was examined in xenograft mouse model. Numb‐L level was increased in tumors and negatively correlated with RBM10 protein level. RBM10 mRNA level was not significantly altered in endometrial tumors suggesting its expression may regulated by post transcriptional regulators such as miRNAs. We identified miR‐133a, miR‐133b, and miR‐335 directly target RBM10, but only miR‐335 level increased in tumors and negatively correlated with RBM10 protein level. miR‐335 overexpression promoted tumor growth by downregulating RBM10 and upregulating Numb‐L level in xenograft mouse model. miR‐335 overexpression promoted Numb‐L expression via targeting RBM10 in endometrial cancer, which may provide new biomarkers for EC diagnosis.

1. INTRODUCTION

Numb is a conserved protein which regulates cells fate and differentiation in normal physiological conditions. Numb also plays important roles in the development of cancer, but its function varies in different type of cancers. It is reported that, Numb functions as a tumor suppressor through controls p53 activity and repressing Notch signaling in breast cancer cells.1 Meanwhile, low Numb level has been proven to be related to the poor prognosis of breast cancer patients.1, 2 However, in glioma, Numb overexpression does not exert a tumor suppressor function and does not impair cell proliferation in vitro.3 In endometrial cancer, Numb is found overexpressed and accumulates in the cell nucleus.4, 5

Alternative splicing (AS) is a process enables a single gene produce variant mRNAs, which contributes to the generation of proteomic and functional diversity. According to the high‐throughput sequencing data, there are more than 95% human genes may undergo AS.6 In human cells, AS is mediated by spliceosome, which is a dynamic enzymic complex consist of five ribonucleoprotein particles (U1, U2, U4, U5, and U6 snRNPs).7 It is considered that all the components of AS system and the AS process are tightly controlled during every development stages. Meanwhile, their dysregulation is closely related to various human diseases including cancers.8, 9 It is identified that two main Numb transcripts can be produced by AS which translated into two major Numb protein isoforms that differ in the length of their proline‐rich region.10 These two Numb isoforms play contrast roles in regulating cellular functions. The longer isoform Numb‐L promoted proliferation but not differentiation in murine embryonic carcinoma cells, whereas the shorter isoform Numb‐S promoted differentiation but not proliferation.11 It is reported that expression of Numb‐L was increased in lung cancer, bladder cancer, breast cancer, and colon cancer, but in endometrial cancer, the condition is still unknown.12, 13

In the present study, we detected the level of two Numb transcripts and protein isoforms by semiquantitative polymerase chain reaction (PCR) and immunoblotting in 47 paired endometrial tumor and adjacent non‐tumor control tissues. We found Numb‐L was increased in tumor tissues which was regulated by disturbed miRNAs.

2. MATERIALS AND METHODS

2.1. Subjects

A total of 47 endometrial cancer patients were recruited from The First Affiliated Hospital of Zhejiang Chinese Medical University between January 2016 and May 2019. Tumor tissues and adjacent normal tissues were collected after surgery. No patient had undergone any chemotherapy or radiotherapy treatment before samples collection. Samples were collected immediately after surgical resection, clinical stage and histologic classification were based on the revised International Federation of Gynecology and Obstetrics classification system.14 Histologic diagnoses were reviewed by three independent pathologists. This study was approved by the ethics committee of The First Affiliated Hospital of Zhejiang Chinese Medical University and written informed consents were obtained from all the patients. The characteristics of patients were provided in Table 1.

Table 1.

Clinical characteristics of patients

Characteristics	Number of cases	Numb‐L level	P value	RBM10 level	P value	miR‐335 level	P value
Age, years
>45	31	10.56 ± 7.52	.43	0.29 ± 0.21	.059	5.98 ± 2.73	.27
≤45	16	8.23 ± 6.27		0.40 ± 0.18		7.24 ± 5.19
Lymph node metastasis
Yes	18	12.87 ± 6.85	.0087	0.28 ± 0.20	.014	8.22 ± 4.95	.032
No	29	5.87 ± 5.21		0.45 ± 0.18		5.37 ± 2.55
BMI
≤25	26	10.62 ± 6.95	.41	0.32 ± 0.21	.55	4.76 ± 3.94	.059
>25	21	8.15 ± 7.65		0.36 ± 0.18		8.02 ± 3.14
FIGO stages
I	22	6.15 ± 6.42	.076	0.45 ± 0.14	.020	4.71 ± 2.79	.042
II	11	5.02 ± 3.86		0.30 ± 0.18		6.47 ± 3.73
III	14	12.81 ± 9.36		0.22 ± 0.21		7.92 ± 4.92
Histological type
Endometrioid	34	9.41 ± 8.07	.84	0.30 ± 0.20	.37	5.69 ± 4.15	.10
Non‐endometrioid	13	9.87 ± 7.47		0.36 ± 0.22		7.77 ± 2.82

2.2. RNA extraction, RT‐PCR, and quantitative RT‐PCR

Total RNA was extracted from tissues and cells using Trizol Reagent (Thermo Fisher Scientific) following the manufacturer's instructions. One microgram total RNA was subjected to reverse transcription using the RevertAid first‐stand cDNA synthesis kit (Thermo Fisher Scientific). Numb cDNA were amplified using DreamTaq master mix (Thermo Fisher Scientific) with PCR cycle conditions of 95°C 5 minutes, 95°C 15 seconds, 60°C 15 seconds, 72°C 1 minute, for 30 cycles. PCR products were separated by electrophoresis using 1.5% agarose gel.

Primers: NUMB‐F GAAGTAGAAGGGGAGGCAGA; NUMB‐R GTCGGCCTCAGAGGGAGTA.

Quantitative reverse transcription‐polymerase chain reaction (RT‐PCR) analysis was used to determine the relative level of selected miRNAs. The levels were detected by TaqMan miRNA RT‐Real Time PCR. Single‐stranded cDNA was synthesized using TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems, Foster City, California) and then amplified using TaqMan Universal PCR Master Mix (Applied Biosystems) together with miRNA‐specific TaqMan MGB probes (Applied Biosystems). Cel‐miR‐39 was used for normalization. Each sample in each group was measured in triplicate and the experiment was repeated at least three times.

2.3. Western blotting

Protein extracts were boiled in sodium dodecyl sulfate/β‐mercaptoethanol sample buffer, and 15 μg samples were loaded into each lane of 10% polyacrylamide gels. The proteins were separated by electrophoresis, and the proteins in the gels were blotted onto a polyvinylidene fluoride membrane (Amersham Pharmacia Biotech, St. Albans, Herts, United Kingdom) by electrophoretic transfer. The membrane was incubated with rabbit anti‐Numb polyclonal antibody (Abcam, Cambridge, Massachusetts), or rabbit anti‐Norch1 polyclonal antibody (Abcam) or mouse anti‐β‐actin monoclonal antibody (Santa Cruz BiotechnologyInc., Santa Cruz, California) overnight at 4°C. The specific protein‐antibody complex was detected by using horseradish peroxidase conjugated rabbit anti‐mouse IgG. Detection by the chemiluminescence reaction was carried using ECL kit (Pierce, Appleton, Wisconsin). The β‐actin signal was used as a loading control. The relative protein level was quantified by Quantity One (BioRad).

2.4. Cell culture

Human endometrial cancer cell line HEC‐1‐A were purchased from ATCC and RL95‐2 was purchased from National Infrastructure of Cell Line Resource. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum (FCS; Hyclone), 20 mmol/L HEPES, 100 IU/mL penicillin, and 100 μg/mL streptomycin. To construct Numb or miR‐335 overexpression stable cells line, HEC‐1A cells were transfected with Numb or miR‐335 expression vectors or empty control vector, and then cultured in puromycin containing medium 3 days after transfection. Monoclones were separated and the overexpression of miR‐335 was confirmed by qRT‐PCR.

2.5. Cell proliferation assay

Cell proliferation was estimated by the 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium bromide (MTT) assay. HEC‐1A and RL95‐2 cells were seeded in wells of 96‐well plates at low density (2 × 10³) and allowed to attach overnight. The cells were then transfected with Numb‐L or Numb‐S expression vector. Twenty microliters of MTT (5 mg/mL) (Sigma‐Aldrich) were added to each well 72 hours after transfection, and the cells were incubated for a further 4 hours. The absorbance was recorded at 570 nm with a 96‐well plate reader after addition of dimethyl sulfoxide.

2.6. Xenograft mouse model

1 × 10⁶ HEC‐1A stable cells were injected subcutaneously into the left and right shoulder flanks of female nude mice (five mice each group). The sizes of xenograft were measured once a week. Mice were sacrificed 30 days after injection and the tumors were subjected to RNA and protein extraction.

2.7. Dual luciferase assays

To generate 3'‐UTR luciferase reporter, the full length RBM10 3'UTR was cloned into the pmirGLO plasmid (Promega, Madison, Wisconsin) downstream of firefly luciferase coding region. miRNA mimics were synthesized by GenePharma Co., Ltd (Shanghai, China). For the luciferase reporter assay, HEC‐1A cells were seeded in 48‐well plates. Luciferase reporter vectors were co‐transfected with miRNA mimic using lipofectamine 2000 (Invitrogen). Two days after transfection, cells were harvested and assayed with the Dual‐Luciferase Assay kit (Promega). Each treatment was performed in triplicate in three independent experiments. The results are expressed as relative luciferase activity (Firefly Luciferase/Renilla Luciferase).

2.8. Statistical analysis

Data were analyzed using SPSS Statistical Package version 19 (SPSS Inc., Chicago, Illinois). The level of proteins and miRNAs in tumors, luciferase activities and proliferation assay were analyzed using paired t test. Pearson correlation analysis was performed to investigate the linear relationship between protein level and miRNA levels. The findings were considered significant at a P value < .05.

3. RESULTS

3.1. Numb‐L is upregulated and RBM10 is downregulated in endometrial tumors

Two Numb isoforms have been found produced by AS and play contrast roles in regulating cellular functions. It is reported that the expression of Numb‐L is increased in endometrial cancer, but the underlying mechanism is still unknown. RNA binding protein including RBM5, 6, and 10 have been found to regulate Numb AS in lung cancer 15. So we hypothesize that these three RNA binding proteins may be involved in dysregulated Numb‐L AS in endometrial cancer. To explore the expression pattern of Numb in endometrial cancer, we amplify the Numb mRNA segment by RT‐PCR. The results indicated that the Numb‐L mRNA level was dramatically increased in tumor tissues (Figure 1A). Moreover, the pan protein of Numb was detected by immunoblotting and we found remarkably increased protein level of long isoform Numb in tumors compared with adjacent non‐tumor controls (Figure 1A). It is reported that the AS regulators including RBM5, RBM6, and RBM10 may control the expression pattern of Numb gene.15 So we detected the expression of these proteins by immunoblotting. We found only RBM10 level was significantly reduced in tumors (Figure 1A) suggesting RBM10 may regulate Numb AS in endometrial cancer. Meanwhile, as the downstream gene of Numb, Notch1 level was increased in tumors (Figure 1A). Subsequently, we quantified the immunoblotting results of Numb and Notch1 from 47 paired tumors and adjacent non‐tumor control tissues. The level of Numb‐L and Notch1 was significantly increased in tumors (Figure 1B,C).

Figure 1.

Numb‐L is overexpressed in endometrial cancer. Twenty six patients with endometrial cancer were include in this research. RNA and protein were extracted from all tumor and adjacent non‐tumor control tissues samples from. The RNA level of Numb was determined by RT‐PCR. The protein level of Numb, RBM5, RBM6, RBM10, and Notch1 was determined by immunoblotting. The results were quantified by Quantity One software and analyzed by paired t test. *P < .05, **P < .01

3.2. Numb‐L functioned as an oncogene in endometrial cancer cells

Numb plays important roles in the development of cancer, but its function varies in different type of cancers.1, 3 Since two isoforms were found in endometrial tumors, we first examined their functions in endometrial cancer cells. Vector control and Numb overexpression for two isoforms of Numb were established in EC cell lines. The overexpression of Numb‐L and Numb‐S was confirmed by immunoblotting (Figure 2A) and the cell viability was monitored by MTT assay. As shown in Figure 2B, a significantly increased cell viability was found in Numb‐L overexpressed HEC‐1A cells. Oppositely, Numb‐S overexpression inhibited HEC‐1A cells viability. It is reported that Numb is a negative regulator of the conserved Notch signaling pathway which is deregulated in cancer.16 To determine the contribution of Numb isoforms to the Notch signaling pathway, the mRNA expression levels of Hes1 and Hey2, two confirmed Notch targets, were subsequently detected by qRT‐PCR (Figure 2C). The results indicated that the Notch signaling was activated by Numb‐L but repressed by Numb‐S. Similar results were observed in RL95‐2 cells. Numb overexpression RL95‐2 cells were established (Figure 2D). We found Numb‐L promoted RL95‐2 cells viability and Numb‐S had the opposite function (Figure 2E). Meanwhile, Numb‐L promoted the expression of Hes1 and Hey2, but Numb‐S repressed their expression (Figure 2F).

Figure 2.

The opposite function of numb‐L and numb‐S in endometrial cancer cell lines. Numb‐L or Numb‐S overexpression cell lines were established and MTT assay were used to determine the cells viability. The relative expression of HES1 and HEY2 was determined by qRT‐PCR. Results were analyzed by paired t test and P < .05 was considered statistically significant. *P < .05, **P < .01. qRT‐PCR, quantitative reverse transcription‐polymerase chain reaction

3.3. MiR‐335 regulated RBM10 expression on posttranscriptional level in endometrial tumors

To investigate the mechanism of how of RBM10 is overexpressed in EC tumor tissues, we first analyzed the RBM10 protein and mRNA level in EC or control tissues. The results indicated that although the RBM10 protein level is reduced in tumors, the RBM10 mRNA level was not significantly changed, suggesting the overexpression of RBM10 in tumors were induced in post transcription level (Figure 3A,B). Since microRNAs are a group of very important post transcriptional gene expression regulators, we predicted the candidate microRNAs which have the potential to target RBM10 mRNA, by using bioinformatics tool: miRanda (http://www.microrna.org). We found nine microRNAs may target RBM10 mRNA 3'UTR directly (miR‐22, ‐29a, ‐29b, ‐122, ‐133a, ‐133b, ‐179, ‐335, ‐382). To confirm the interaction between microRNAs and RBM10 mRNA, we constructed the reporter vector by cloning RBM10 3'UTR sequence into pmirGLO plasmid (Figure S1A). The RBM10 reporter vector transfected into HEC‐1A cells with one of the night candidate microRNAs for 48 hours, and the cells were lysed and subjected to dual luciferase assay. We found the relative firefly luciferase activity was repressed by miR‐133a, miR‐133b, and miR‐335 significantly (Figure 3C). When three nucleotides in the predicted miR‐133a/b target region or four nucleotides in the miR‐335 predicted target region were mutated, the luciferase activity was not repressed by miR‐133a/b and miR‐335, indicated that these three microRNAs inhibit firefly luciferase expression via targeting RBM10 3'UTR sequence (Figure S1C).

Figure 3.

RBM10 expression was repressed by miR‐335 in tumors. The protein and mRNA level of RBM10 in tumors were determined by immunoblotting and qRT‐PCR. Dual luciferase were used to examine the interaction between miRNAs and RBM10 3'UTR. qRT‐PCR was used to determine miR‐335 level in tumors and the relationship between miR‐335 and RBM10 protein level was analyzed by Pearson correlation analysis. P < .05 was considered statistically significant. *P < .05, **P < .01. qRT‐PCR, quantitative reverse transcription‐polymerase chain reaction

To investigate whether endogenous RBM10 expression is regulated by miR‐133a/b and miR‐335, we overexpressed these miRNAs in HEC‐1A cells for 48 hours, and the RBM10 protein level was determined by immunoblotting. We observed significantly repressed RBM10 protein level in miR‐133a/b and miR‐335 overexpressed cells, indicating RBM10 is a direct target of miR‐133a/b and miR‐335 (Figure 3D). Moreover, we detected the level of miR‐133a/b and miR‐335 in the normal and tumor tissues from 26 EC patients. As shown in Figure 3E, miR‐335 was significantly upregulated in the tumors when compared with adjacent non‐tumor control tissues. However, no significant difference was found of miR‐133a/b level in tumors (Figure S2A,B). After correlation analysis, we found a negative correlation between miR‐335 level and RBM10 protein level (Figure 3F). Meanwhile, no significant correlation was found between miR‐133a/b and RBM10 protein level in tumors from EC patients (Figure S2C,D).

3.4. miR‐335 promoted tumor growth, repressed RBM10 expression, and upregulated Numb‐L expression in xenograft mouse model

To investigate the function of miR‐335 in vivo, we established miR‐335 overexpression HEC‐1A cell line. The cells were injected subcutaneously into nude mice, and the tumor volumes were monitored every week. As shown in Figure 4A, the miR‐335 overexpression contributed to a faster growth of HEC‐1A cells in vivo. Meanwhile, the overexpression of miR‐335 was confirmed by qRT‐PCR (Figure 4B), and we observed increased NUMB‐L and reduced RBM10 (Figure 4C).

Figure 4.

miR‐335 promoted tumor growth in xenograft mouse model. The miR‐335 overexpression or control HEC‐1A cells were injected subcutaneously into nude mice, and the tumor volumes were monitored every week. Mice were sacrificed 30 days after injection. The miR‐335 level was determined by qRT‐PCR and the Numb expression pattern was examined by immunoblotting. Results were analyzed by paired t test and P < .05 was considered statistically significant. *P < .05, **P < .01. qRT‐PCR, quantitative reverse transcription‐polymerase chain reaction

4. DISCUSSION

Numb is a cell fate regulator and it function was first identified in Drosophila. As a negative regulator, Numb binds Notch receptor and represses downstream genes expression of the Notch signaling pathway.17 It is reported that the AS of Numb is strictly controlled during the mammalian development, with the longer isoform containing exon12 at early development stages which promote proliferation, and the shorter isoform lacking exon 12 promotes differentiation18, 19, 20). Increasing evidences indicate that altered Numb expression pattern relates to the progression of tumors including lung cancer, bladder cancer, breast cancer and colon cancer.12, 13 The present findings indicated the opposite role of Numb‐L and Numb‐S in endometrial cancer cells, and Numb‐L is overexpressed in endometrial cancers which relates to repressed Notch1 signaling. These findings provided candidate markers for patients diagnosis.

RBM5, RBM6, and RBM10 are nuclear localized RNA binding proteins which involved in mRNA AS. Bechara EG and colleagues identified that RBM5, 6, and 10 regulated Numb AS in lung cancer cells.15 In the present study, we detected the expression of RBM5, RBM6, and RBM10 in tumor tissues from patients with endometrial cancer and found only RBM10 protein level is reduced in tumors which contributed to increased Numb‐L level, suggesting Numb AS is mainly regulated by RBM10 in endometrial cancer.

miRNA is group of short non‐protein coding RNA which regulates gene expression through binding with target mRNAs. In this study, we identified three miRNAs target RBM10 directly, but only miR‐335 was negatively correlate with RBM10 protein level in tumors, indicating miR‐335 is a key regulator of RBM10 in ECs. Altered miR‐335 level has been found in various cancers, but its function is still controversial. miR‐335 has been found reduced in gastric cancer,21 papillary thyroid cancer,22 and osteosarcoma23 and relates to drug resistance in lung cancer.24 However, overexpression of miR‐335 was identified in colorectal carcinoma,25 acute myeloid leukemia,26 and uterine sarcoma.27 In the present study, we confirmed in vivo that miR‐335 functioned as an oncomiR promote endometrial tumors growth via repressing RBM10 and upregulating Numb‐L.

In conclusion, we identified that overexpressed miR‐335 promoted Numb‐L expression via targeting RBM10 in endometrial cancer, which may provide new biomarkers for EC diagnosis.

CONFLICT OF INTEREST

All authors declare no conflict of interest.

Supporting information

Fig. S1 Identify the target sites of miRNAs in RBM10 3'UTR. A. The schematic diagram of the RBM10 3'UTR reporter vector. B. The schematic diagram of mutant RBM10 3'UTR. C. Dual luciferase assay. Results were analyzed by paired t test and P < .05 was considered statistically significant. *P < .05, **P < .01.

Fig. S2 MiR‐133a/b expression is not changed in tumors and not correlates with RBM10 level. The expression of miR‐133a/b in tumors was determined by qRT‐PCR, and the their correlations with RBM10 protein level were analyzed by Pearson correlation analysis. P < .05 was considered statistically significant.

Dou X‐Q, Chen X‐J, Zhou Q, Wen M‐X, Zhang S‐Z, Zhang S‐Q. miR‐335 modulates Numb alternative splicing via targeting RBM10 in endometrial cancer. Kaohsiung J Med Sci. 2020;36:171–177. 10.1002/kjm2.12149