Circ-SMARCA5 suppresses progression of multiple myeloma by targeting miR-767-5p

Background We aimed to investigate the correlation of Circ-SMARCA5 with disease severity and prognosis in multiple myeloma (MM), and its underlying mechanisms in regulating cell proliferation and apoptosis. Methods Bone marrow samples from 105 MM patients and 36 healthy controls were collected for Circ-SMARCA5 expression measurement. And the correlation of Circ-SMARCA5 expression with patients’ characteristics and survival was determined. In vitro, the effect of Circ-SMARCA5 on MM cell proliferation and apoptosis was evaluated by altering Circ-SMARCA5 expression through transfection. Rescue experiments and luciferase assay were further performed to explore the mechanism of Circ-SMARCA5 as well as its potential target miR-767-5p in regulating MM cell activity. Results Circ-AMARCA5 was downregulated in MM and presented a good value in distinguishing MM patients from controls and it was also negatively correlated with Beta-2-microglobulin (β2-MG) level and International Staging System (ISS) stage. Additionally, Circ-SMARCA5 high expression was associated with higher CR as well as better PFS and OS. As for in vitro experiments, Circ-SMARCA5 expression was lower in MM cell lines compared with normal cells, and Circ-SMARCA5 overexpression inhibited cell proliferation but promoted cell apoptosis in RPMI8226 cells. Rescue experiments disclosed that the effect of Circ-SMARCA5 on cell activity was attenuated by miR-767-5p, and luciferase reporter assay revealed direct binding between Circ-SMARCA5 and miR-767-5p. Conclusions Circ-SMARCA5 is downregulated and correlated with lower β2-MG level and ISS stage as well as better prognosis in MM patients, and it inhibits proliferation but promotes apoptosis of MM cells via directly sponging miR-767-5p. Electronic supplementary material The online version of this article (10.1186/s12885-019-6088-0) contains supplementary material, which is available to authorized users.


Background
Multiple myeloma (MM), also called plasma cell myeloma, is a hematological malignancy characterized by proliferation of malignant monoclonal plasma cells, and its clinical manifestations include multiple bone destruction, pathological fracture, hypercalcemia, anemia, renal dysfunction and so on [1]. MM is the second most common hematological malignancy that ranks after non-Hodgkin's lymphoma and is age-related, suggesting that with the population aging, the incidence of MM will greatly increase in the future [2]. Similar to other hematological malignancies, the common therapeutic methods for MM include chemotherapy, targeted therapy, immunotherapy and hematopoietic stem cell transplantation, while the treatment efficacy is subject to relapse of disease and refractoriness to treatments, leading to poor survival in most of the MM patients [3]. Therefore, exploring more potential targets is necessary for development of new therapies and improving patients' prognosis in MM.
Non-coding RNAs, a category of RNA molecules not translated into proteins, are widely abundant in human genome and engage in numerous cellular processes including transcription, post-transcriptional modification and signal transduction [4]. Circular RNAs (CircRNAs) comprise a class of non-coding RNAs whose 3′ and 5′ ends are joined together to form a covalently closed loop [5]. Benefiting from the structure, circRNAs are resistant to exonuclease-mediated degradation and are more stable compared with linear RNAs [6]. Based on evidence from the past decades, circRNAs play critical roles in gene regulation by sponging micro RNAs (miRNAs) and regulating protein translation, thereby involve in pathogenesis of various diseases, and their expression patterns and mechanisms in some solid tumors have been revealed [7]. As for hematological malignancies, circRNAs participate in development and progression of diseases and serve as valuable biomarkers for diagnosis and prognosis in patients, whereas little is known with respect to circRNAs in MM [8][9][10].
CircRNA SWI/SNF-related matrix-associated actindependent regulator of chromatin subfamily A member 5 (Circ-SMARCA5) is encoded by SMARCA5 gene that is located on chromosome 4 (location: NC_000004.12 (143,513,463..143557489)) [11]. Circ-SMARCA5 is previously shown to participate in carcinogenesis by inhibiting cell proliferation, migration and invasion in several solid tumors including hepatocellular carcinoma and cervical cancer [12][13][14]. Whereas in hematological malignancies, the role of Circ-SMARCA5 is unknown. One previous study reports that Circ-SMARCA5 is involved in the regulation of cell cycle and alterations of chromatin structures, and dysregulation of SMARCA5 gene is presented in CD34+ hematopoietic progenitors of hematological malignancy, and we speculated that Circ-SMARCA5 might participate in etiology and work as a potential biomarker of MM, a typical hematological malignancy [11,15]. Therefore, the purpose of this study was to investigate the correlation of Circ-SMARCA5 with disease severity and prognosis in MM patients as well as its underlying mechanisms in regulating cell proliferation and apoptosis in MM cells.

Participants
One hundred and five de novo MM patients treated at the Affiliated Hospital of Nantong University between Jan 2015 and Dec 2016 were consecutively enrolled in this study. The inclusion criteria were: (1) newly diagnosed as MM according to 2014 International Myeloma Working Group (IMWG) updated criteria for the diagnosis of multiple myeloma; (2) age above 18 years old; (3) able to be regularly followed up; (4) life expectancy over 12 months. The exclusion criteria were as follows: (1) relapsed/refractory MM; (2) received stem cell transplantation (SCT), chemotherapy, radiotherapy or other systematic treatments before enrollment; (3) with a history of solid tumors or hematological malignancies other than MM; (4) pregnant/lactating women. In addition, there were 36 healthy bone marrow donors being recruited as controls during the same period. The present study was approved by the Ethics Committee of the Affiliated Hospital of Nantong University, and all participants provided written informed consents.

Baseline data collection
After enrollment, baseline characteristics were collected from all patients, which included: information of age, gender, immunoglobulin subtype, hemoglobin (Hb), calcium, serum creatinine (Scr), albumin (ALB), Beta-2microglobulin (β2-MG), Durie-Salmon Stage, the International Staging System (ISS) Stage, bone lesion, lactate dehydrogenase (LDH) and cytogenetics abnormality. Durie-Salmon Stage and ISS Stage were assessed according to the Durie-Salmon Criteria and ISS Criteria respectively [16,17], and cytogenetics abnormalities were analyzed by fluorescence in situ hybridization (FISH).

Sample collection and Circ-SMARCA5 detection
Bone marrow samples of all patients were collected prior to initiation of therapy, and bone marrow samples of controls were also obtained after enrollment. Mononuclear cells were separated from bone marrow by gradient density centrifugation, then plasma cells were purified using CD138-coated magnetic beads according to the manufacturer's instructions (Miltenyi Biotec, Germany) to ensure greater than 90% plasma cell purity. Subsequently, Circ-SMARCA5 expression was determined by quantitative polymerase chain reaction (qPCR) assay.

Treatment and assessment
All patients received induction therapies after enrollment, and the selections of chemotherapy regimens entirely depended on patients' willingness and disease conditions. Fourty-eight patients received conventional chemotherapy, 36 patients underwent bortezomib-based regimen and 21 received immunomodulatory drug-based regimen, details were as follows: 26 treated with VAD regimen (doxorubicin/vincristine/dexamethasone); 22 treated with DVD regimen (liposomal doxorubicin/ vincristine/dexamethasone); 36 treated with BCD regimen (bortezomib/cyclophosphamide/dexamethasone), 21 treated with MPT regimen (melphalan/prednisone/thalidomide). After induction treatment, 31 patients underwent autologous SCT, and others received maintenance therapies or repeated induction regimens according to the response to induction treatment. The treatment responses including complete response (CR), very good partial response (VGPR), partial response (PR) were assessed according to the IMWG criteria [18], and the overall response rate (ORR) was defined as the percentage of patients with the achievement of CR, VGPR or PR. All MM patients were regularly followed up to 2017/12/31, and the median follow-up duration was 24.0 months (range: 5.0-36.0 months). Furthermore, progression free survival (PFS) and overall survival (OS) were calculated, which were defined as: PFS, the duration from the treatment to disease progression or death; OS, the duration from the treatment to the death.
Measurement of Circ-SMARCA5 in MM cell lines and normal plasma cells.
The qPCR was performed to detect the expression of Circ-SMARCA5 in wild-type MM cell lines: NCI-H929, RPMI8226, U226, OPM2 and JJN3, and normal plasma cells.

Transfection, groups and followed measurements in RPMI8226 cells
Blank overexpression, Circ-SMARCA5 overexpression, blank shRNA and Circ-SMARCA5 shRNA plasmids were constructed by Shanghai GenePharma Bio-Tech Company (Shanghai, China), and then transfected into RPMI8226 cells as NC (+), Circ (+), NC (−) and Circ (−) groups. Then in each group, qPCR was performed to measure Circ-SMARCA5 expression at 24 h, CCK8 was performed to measure cell proliferation at 0 h, 24 h, 48 h and 72 h, AV/PI was performed to measure cell apoptosis rate at 72 h, and Western Blot was performed to measure apoptotic markers Cleaved Caspase 3 and Bcl-2 expressions at 72 h post transfection. The detailed procedure of each detection was described in the following subsections.

Validation in NCI-H929 cells
In order to validate the effect of Circ-SMARCA5 on cell proliferation and apoptosis, the transfections of blank overexpression, Circ-SMARCA5 overexpression, blank shRNA and Circ-SMARCA5 shRNA plasmids were repeated in NCI-H929 cells, as well as qPCR at 24 h for detection of Circ-SMARCA5, CCK8 at 0 h, 24 h, 48 h and 72 h for cell proliferation measuement, AV/PI at 72 h for measurement of cell apoptosis, and measurement of apoptotic markers Cleaved Caspase 3 and Bcl-2 expressions at 72 h post transfection by Western Blot.

Regulation of miR-767-5p by circ-SMARCA5 in several MM cell lines
In order to further validate the regulation of miR-767-5p by circ-SMARCA5 in MM, blank overexpression, Circ-SMARCA5 overexpression, blank shRNA and Circ-SMARCA5 shRNA plasmids were transfected into NCI-H929, U226, OPM2 and JJN3 cell lines, and miR-767-5p was detected at 24 h post transfection by qPCR.

Rescue experiment in RPMI8226 cells
In order to explore the underlying mechanism of Circ-SMARCA5 on regulating cell functions in MM cells, rescue experiment was then conducted. Blank overexpression, miR-767-5p overexpression, Circ-SMARCA5 overexpression, and miR-767-5p overexpression & Circ-SMARCA5 overexpression plasmids were constructed by Shanghai GenePharma Bio-Tech Company (Shanghai, China) and then transfected into RPMI8226 cells as Control (+), miR-767 (+), SMARCA5 (+), and SMARCA5 (+)/miR-767 (+) groups. Then in each group, qPCR was performed to measure miR-767-5p and Circ-SMARCA5 expressions at 24 h, CCK8 was performed to measure cell proliferation at 0 h, 24 h, 48 h and 72 h, AV/PI was performed to measure cell apoptosis rate at 72 h, and Western Blot was performed to measure apoptotic markers Cleaved Caspase 3 and Bcl-2 expressions at 72 h post transfection. The detailed procedure of each detection was described in the following subsections.

Luciferase reporter assay
In order to further validate the interaction between Circ-SMARCA5 and miR-767-5p, luciferase reporter assay was subsequently conducted as follows: (1) the RPMI8226 cells were digested and inoculated on a culture plate and left in 5% CO2, humidity saturated incubator (37°C) over night. (2) When cell density exceeded 70%, co-transfecting the cells with Circ-SMARCA5 wild type/ mutant type luciferase reporter plasmids and miR-767-5p overexpression/control plasmids. Reporter plasmids and overexpression plasmids were constructed by Shanghai GenePharma Bio-Tech Company (Shanghai, China). (3) At 24 h after transfection, the culture solution was removed and the cells were washed with cold PBS. (4) Add 350 μl pre-cooled harvest buffer into each culture dish and the cells were lysed at 4°C or on ice for 10 mins. (5) Centrifuge for 5 s at 12,000 rpm. Transfering the supernatant to a new tube and discard the pelleted cell debris. (6) 5 ml microcentrifuge tubes were prepared by adding 100 μl of reactant solution of ATP buffer mixed with luciferin buffer (in ratio of 1:3.6) to each tube. (7) 100 μl of cell lysate supernatant from step 5 was added into each of the microcentrifuges prepared in step 6, blending quickly and read the absorbance value on a luminometer. (8) Ensure to read the absorbance value of all samples with the same procedure, then measured the activity of LacZ from the rest of the cell lysate and used it as an internal reference to correct the readings of luciferase.
qPCR Total RNA was extracted using TRIzol reagent (Invitrogen, USA) and then applied for the synthesis of cDNA using PrimeScript™ RT reagent Kit (with random primers for Circ-SMARCA5) (TAKARA, Japan). Following that, cDNA was used for qPCR with SYBR Green kit (TaKaRa, Japan). Amplification of PCR was conducted under the following condition: 95°C for 3 min, 40 cycles of 95°C for 5 s, 61°C for 10 s, and then 72°C for 30s. The result was calculated using 2 -ΔΔCt method and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal reference. The primers used in qPCR were listed in Additional file 1: Table S1.

Western blot
After lysing with RIPA lysis buffer (Thermo Scientific, USA), RPMI8226 cells were centrifuged, then total protein concentration was measured using BCA™ protein assay kit (Pierce, USA). Electrophoresis was used for separating proteins on sodium dodecyl sulfatepolyacrylamide (SDS) gel. Subsequently, PVDF membrane (Millipore, USA) was used to transfer proteins and then blocked with 5% non-fat dried milk in PBST for 1 h at room temperature. After incubation with primary antibodies, the PVDF membrane was finally incubated with the secondary antibody. ECL advanced Western blot analysis detection kit (BD, USA) was then used to visualize the bands indicating the abundance of proteins. The grey intensity was measured using ImageJ (NIH, USA). The antibodies used in Western Blot were listed in Additional file 2: Table S2.

CCK8
Ten ul CCK-8 (Dojindo, Japan) and 90 ul medium were added to each group of RPMI8226 cells, then the cells were incubated at 37°C under 95% air + 5% CO2. Cell proliferation ability was presented by optical density (OD) value, which was measured using a microplate reader (BioTek, USA).

Av/pi
RPMI8226 cells were digested using pancreatin and washed with phosphate buffer solution, then suspended in blinding buffer (100 ul). 10 ul AV (Invitrogen, USA) was subsequently added and the cells were left on ice for 15 min in dark. Following that, 5 ul PI (Invitrogen, USA) was added and the apoptosis rate was analyzed by flow cytometry (FCM) (Becton Dickinson, USA).

Statistics
Statistical analysis was performed using SPSS 22.0 software (IBM, USA), and figures were made with the use of GraphPad Prism 7.00 software (GraphPad Software Inc., USA). Count data were expressed as count (percentage); Continuous data were described as mean ± standard deviation (clinical data) or mean ± standard error (experimental data) if normally distributed, and as median (25th-75th quantiles) if not normally distributed. Comparison was determined by Chi-square test, t test or Wilcoxon rank sum test. Receiver operating characteristic (ROC) curve was used to evaluate the ability of Circ-SMARCA5 relative expression to discriminate between MM patients and controls. Survival curves were made using the Kaplan-Meier method and significant differences between the curves were determined by log-rank test. Univariate and multivariate logistic regression analyses with Forward Stepwise (Conditional) method were performed to assess the predictive factors of CR; univariate and multivariate Cox's proportional hazards regression analyses with Forward Stepwise (Conditional) method were performed to determine the factors affecting PFS and OS. P value < 0.05 was considered as significant.

Circ-SMARCA5 relative expression in MM patients
The median Circ-AMARCA5 relative expression was 0.778 (0.377-1.421) in MM patients, which was lower than that in control group (1.407 (0.864-2.763)) (P < 0.001) (Fig. 1a). ROC curve illustrated that Circ-AMARCA5 was of good value in distinguishing MM patients from the controls with area under curve (AUC) of 0.714 (95%CI: 0.614-0.814). The sensitivity and specificity at the best cut-off point (the point where the largest sum of sensitivity and specificity occurred) were 93.3 and 41.7% respectively and the cut-off value of Circ-AMARCA5 relative expression was 2.242 ( Fig. 1b).

Association between Circ-SMARCA5 relative expression and patients' baseline characteristics
All patients were divided into high expression and low expression groups depending on whether their Circ-SMARCA5 relative expression was greater than the median value in MM patients or not, which showed that Circ-SMARCA5 high expression was correlated with lower β2-MG level (P = 0.001) as well as less advanced ISS stage (P < 0.001) ( Table 2).
Correlation between treatment response of MM patients and Circ-SMARCA5 expression.
There were 25 patients who achieved CR and 80 patients who did not. The expression of Circ-SMARCA5 was higher in CR patients compared to non-CR patients (P = 0.045) (Fig. 1c). As for ORR, 80 patients achieved ORR but 25 patients did not, and Circ-SMARCA5 expression was similar between ORR patients and non-ORR patients (P = 0.458) (Fig. 1d).
Comparison of survivals between Circ-SMARCA5 high expression and low expression groups.

Circ-SMARCA5 relative expression in MM cell lines and normal plasma cells
In order to get a deeper understanding of the underlying mechanism of Circ-SMARCA5 in MM, in vitro experiments were performed to investigate the effect of Circ-SMARCA5 on cell activities as well as its possible target Wilcoxon rank sum test was used to compare Circ-SMARCA5 relative expression between MM patients and controls, and the ROC curve was used to evaluate the ability of Circ-SMARCA5 relative expression to discriminate between MM patients and controls. c-d: Association of Circ-SMARCA5 expression with treatment response in MM patients. The cutoff value of Circ-SMARCA5 was the median value of Circ-SMARCA5 in MM patients. Comparison between two groups was performed using the Chi-square test. e-f: Comparison of PFS and OS between Circ-SMARCA5 high expression and low expression groups. The cutoff value of Circ-SMARCA5 was the median value of Circ-SMARCA5 in MM patients. Survival curves were made using the Kaplan-Meier method and differences between the curves were determined by log-rank test. P < 0.05 was considered significant. PFS, progression free survival; OS, overall survival; Circ-SMARCA5, Circular RNA SWI/SNF-related matrix-associated actin-dependent regulator of chromatin subfamily A member 5; ROC, Receiver operating characteristic. CR, complete response; ORR, objective response rate in MM cells. Circ-SMARCA5 relative expression in MM cell lines and normal plasma cells was detected by qPCR, which revealed that Circ-SMARCA5 was downregulated in MM cell lines including NCI-H929 (P < 0.05), RPMI8226 (P < 0.01), U226 (P < 0.05), OPM2 (P < 0.05) and JJN3 (P < 0.01) compared to normal plasma cells (Fig. 2). The lowest expression of Circ-SMARCA5 was presented in RPMI8226 cells, thus, RPMI8226 cells were chosen to be applied in the following experiments.

Circ-SMARCA5 relative expression after transfection
To verify the transfection effects of plasmids, we used qPCR to detect the expression of Circ-SMARCA5 in RPMI8226 cells at 24 h after transfection, and the data showed that Circ-SMARCA5 was upregulated in SMARCA5 (+) group compared to Control (+) group (P < 0.001) but downregulated in SMARCA5 (−) group compared with Control (−) group (P < 0.001), which indicated successful transfection (Fig. 3a).

Relative luciferase activity of Circ-SMARCA5
In order to further validate the targeting effect of Circ-SMARCA5 on miR-767-5p, luciferase reporter assay was subsequently conducted to confirm the binding site of Circ-SMARCA5 with miR-767-5p. The sequences of Circ-SMARCA5 wild/mutant type and miR-767-5p were shown and the binding sites were marked in Fig. 8a. For wild type Circ-SMARCA5, the relative luciferase activity was lower in miR-767-5p group compared to miR-NC group (P < 0.01), whereas for mutant Circ-SMARCA5, there was no difference in relative luciferase activity between miR-767-5p group and miR-NC group (P > 0.05) (Fig. 8b). This validated that Circ-SMARCA5 interacted with miR-767-5p by direct binding in RPMI8226 cells.

Discussion
Our study discovered that: (1) Circ-SMARCA5 expression was downregulated in MM patients, and its high expression was correlated with lower β2-MG level and ISS stage.
(2) Circ-SMARCA5 expression was positively associated with treatment response and survival in MM  (3) In vitro experiments disclosed that Circ-SMARCA5 reduced cell proliferation but improved cell apoptosis via sponging miR-767-5p in MM cells. MM is currently incurable, and its clinical manifestations are not apparent until the oncological events initiate and progress for a decade in most cases, leading to late diagnosis and unfavorable treatment outcomes [2]. Conventional therapy for MM depends on chemotherapy using cytotoxic agents, while the native resistance to drugs is common due to multiple chromosomal abnormities in the pathogenesis of MM, resulting in unfavorable treatment response and survivals in MM patients [1]. Therefore, identifying biomarkers that could predict the risk as well as prognosis of MM will help with timely diagnosis and treatment for MM patients. Moreover, the detailed mechanism underling the development and progression of MM is still not fully understood, thus, studies investigating the mechanism of MM etiopathology are greatly needed.
CircRNAs are widely expressed in human cells with relatively higher expression and more stable structure compared with their linear counterparts, suggesting that CircRNAs could be the ideal biomarkers for human diseases [7]. Accumulating studies have disclosed that Cir-cRNAs might hold missing triggers of tumorigenesis of various cancers including hematopoietic cancers by regulating gene expression [15,[19][20][21]. For instance, Circ-ANAPC7 is predominantly upregulated in AML by CircRNA microarray analysis and might serve as a promising diagnostic marker for AML [9]. Another study illustrates that Circ_0004277 is notably downregulated in acute myeloid leukemia (AML) and its high expression predicts better treatment response to chemotherapy, acting as one of the diagnostic and treatment targets for AML [10]. In addition, Circ-CBFB has been exhibited to be overly expressed in chronic lymphocytic leukemia (CLL) and serves as a diagnostic and prognostic biomarker for CLL [8]. These previous studies suggest that some CircRNAs play important roles in hematological malignancies and potentially function as promising diagnostic and prognostic markers. Circ-SMARCA5 is a typical CircRNA derived from exons of SMARCA gene, and very limited studies investigating its role in cancer have been carried out until now [22]. For instance, Circ-SMARCA5 is initially reported to be upregulated in prostate cancer and serves as an oncogene by promoting cancer cell proliferation and inhibiting cell apoptosis [23]. Another research reveals that Circ-SMARCA5 level is reduced in hepatocellular carcinoma (HCC) tissues compared with adjacent tissues and its low expression is correlated with poor clinicopathological features in HCC patients [14]. According to these studies, the role of Circ-SMARCA5 in development of solid tumors is controversial. Besides, it is still unknown whether Circ-SMARCA5 is involved in MM or not. In vitro studies have shown that SMARCA5 gene is dysregulated in hematopoietic progenitors that could also potentially develop into MM cells, therefore, we speculated that Circ-SMARCA5, which is encoded from SMARCA5 gene, might play critical roles in MM etiology [15,24]. In this study, we compared the expression of Circ-SMARCA5 in MM patients with that of controls, which revealed that Circ-SMARCA5 was downregulated in MM patients compared to controls. Moreover, we also found that Circ-SMARCA5 high expression was correlated with lower β2-MG level and ISS stage in MM patients. Here are several possible explanations: (1) Circ-SMARCA5 might be one of the antioncogenes whose expression was inhibited by the onset of MM, therefore, Circ-SMARCA5 expression was downregulated in MM patients. (2) Circ-SMARCA5 might participate in cell cycle regulation by inhibiting cell proliferation and promoting cell apoptosis in MM cells to attenuate disease development and progression, thereby reduced disease severity in MM patients, which was validated in our followed in vitro experiments.
Additionally, we also investigated the prognostic value of Circ-SMARCA5 in MM patients and discovered that Circ-SMARCA5 high expression was correlated with better treatment response to chemotherapy and longer survival, indicating that patients with Circ-SMARCA5 high expression were more sensitive to clinical treatment. These could be due to that: (1) Circ-SMARCA5 might function as an anti-tumor gene whose high Comparison of Circ-SMARCA5 expression between control cells and each MM cell line was performed by t test, and P < 0.05 was considered significant. *P < 0.05, **P < 0.01. Circ-SMARCA5, Circular RNA SWI/SNF-related matrix-associated actin-dependent regulator of chromatin subfamily A member 5 expression was associated with lower disease severity. Thus, MM patients with higher Circ-SMARCA5 presented with less advanced disease (lower ISS stage or Durie-Salmon stage), thereby leading to better treatment outcomes, which was further supported by results from Multivariate Cox's regression analysis in our study. (2) Circ-SMARCA5 might affect cells sensitivity to cytotoxic drugs, thereby resulting in better treatment response to chemotherapy and prolonging survival in MM patients. However, further validation is needed focusing on the influence of Circ-SMARCA5 on chemoresistance. Moreover, the sample size of this study was relatively small, which might influence statistical power. Also, the followup time for evaluating survivals was relatively short, therefore the effect of Circ-SMARCA5 on long-term prognosis in MM patients was still not clear in this study.
MiRNAs are single-stranded non-coding RNAs with length of 19-22 nucleotides, which are common targets of CircRNAs and are involved in cells growth, proliferation, apoptosis and differentiation by incomplete base pairing with messenger RNAs (mRNAs) [28]. MiR-767-5p is a member of miR-767 that are located on chromosome X (location: NC_000023.11 (152,393,421..152393529, complement)) [22]. In human melanoma, miR-767 is shown to promote cell proliferation by suppressing cylindromatosis [29]. Another functional experiment displays that aberrant activation of miR-767 devotes to tumor epigenesis in lung cancer via repressing TET1/3 mRNA and regulating genomic 5methylcytosines to 5-hydroxymethylcytosines levels [11]. These previous studies disclose the oncogenic role of miR-767 in several human cancers, whereas in MM, the function of miR-767 still lacks investigation, and only two studies have been disclosed: the first study observed that miR-767-5p was upregulated in both tissue (two folds) and plasma samples (four folds) from MM patients compared to normal control, and it promotes MM cells (H929 and MM.1S) progression via regulating MAPK4 pathway [30]; the second study observed that miR-767-5p was overexpression in extramedullary relapse MM samples compared to non-extramedullary relapse MM samples [31]. In order to explore the potential target miRNA of Circ-SMARCA5 in MM, we referred to Circular RNA Interactome database (https://Circinteractome.nia.nih.gov/) and miRanda database (http://www.microrna.org/micr orna/home.do) and found that miR-767-5p might be a potential target for Circ-SMARCA5 in etiology of MM. To testify this hypothesis, we further performed rescue experiments, which illuminated that Circ-SMARCA5 suppressed Fig. 8 Luciferase reporter assay of Circ-SMARCA5. The sequences of wild type Circ-SMARCA5, mutant type Circ-SMARCA5 and miR-767-5p were shown and the binding sites were lined up (a). For wild type Circ-SMARCA5, the relative luciferase activity was decreased in miR-767-5p group compared to miR-NC group, whereas for mutant type Circ-SMARCA5, the relative luciferase activity was similar between miR-767-5p group and miR-NC group (b). Comparison of relative luciferase activity between the two groups was conducted using t test, and P < 0.05 was considered significant. **P < 0.01, NS, non-significant. Circ-SMARCA5, Circular RNA SWI/SNF-related matrix-associated actin-dependent regulator of chromatin subfamily A member 5; miR, micro RNA cell proliferation but facilitated cell apoptosis by adversely regulating miR-767-5p in RPMI8226 cells. Moreover, the direct binding between Circ-SMARCA5 and miR-767-5p was illustrated by luciferase reporter assay.

Conclusions
In conclusion, Circ-SMARCA5 is downregulated and correlated with lower β2-MG level and ISS stage as well as better prognosis in MM patients, and it inhibits proliferation but promotes apoptosis of MM cells via directly sponging miR-767-5p.