Exosomal circRNA as a novel potential therapeutic target for multiple myeloma-related myocardial damage

Myocardial damage is a mostly incurable complication of multiple myeloma (MM) that seriously affects the treatment outcome and quality of life of patients. Exosomal circular RNAs (exo-circRNAs) play an important role in tumor occurrence and development and are considered key factors in MM pathogenesis. However, the role and mechanism of action of exo-circRNAs in MM-related myocardial damage are still unclear. This study aimed to investigate correlations between exo-circRNAs and MM and to preliminarily explore the role of exo-circRNAs in MM-related myocardial damage. Six MM patients and five healthy controls (HCs) were included in the study. High-throughput sequencing and qRT-PCR verification were used to obtain a profile of abnormally expressed exo-circRNAs. GO, KEGG, miRanda, TargetScan and Metascape were used for bioinformatics analyses. H9C2 cells treated with exosomes from U266 cells were used in cell experiments. CCK-8, PCR, immunofluorescence and western blotting assays were used to detect cell proliferation and expression of autophagy-related indicators. Electron microscopy was used to observe the number of autophagic vesicles. Bioinformatics analysis showed that circRNAs with upregulated expression had the potential to promote MM-related myocardial damage. In addition, PCR results confirmed that circ-G042080 was abundantly expressed in the serum exosomes of 20 MM patients. Correlation analysis showed that the expression level of circ-G042080 was positively correlated with the clinical level of MM and MM-related myocardial damage and that circ-G042080 might interfere with MM-related myocardial damage through a downstream miRNA/TLR4 axis. Cell experiments demonstrated that the circ-G042080/hsa-miR-4268/TLR4 axis might exist in H9C2 cells incubated with exosomes and cause abnormal autophagy. Abnormal expression of serum exo-circRNAs was found to be associated with MM-related myocardial damage, suggesting that exo-circRNAs might become a new diagnostic marker of MM-related myocardial damage and a therapeutic target.


Introduction
Multiple myeloma (MM) is the second most common hematological neoplastic disease after non-Hodgkin's lymphoma, accounting for 1% of all cancers and 10% of all hematological malignancies [1]. MM is characterized Sun et al. Cancer Cell Int (2021) 21:311 by the clonal proliferation of malignant plasma cells and the deposition of monoclonal immunoglobulin (M protein) in the bone marrow. M protein light chains and polysaccharide complexes are deposited in tissues and organs and can damage corresponding organ functions, leading to hypercalcemia, renal failure, anemia, bone lesions and cardiac comorbidities [2]. Among them, cardiac comorbidities are one of the most serious complications and can lead to cardiomyopathy and heart failure caused by cardiac amyloidosis and/or anemia. In addition, certain treatments used for MM can affect cardiac function [3]. Clinical data have confirmed that up to 50% of MM patients have heart damage [4]. In addition, cardiac comorbidities often have an insidious onset and lack specific clinical manifestations, and clinicians are prone to miss the diagnosis, thus delaying diagnosis and treatment. Therefore, we urgently need new targets for MMrelated myocardial damage comorbidities to diagnose or treat the disease.
Circular RNAs (circRNAs) are a group of endogenous noncoding RNAs that are present in exosomes and plasma and are characterized by a covalent closed loop structure lacking a polyadenylated tail [5]. The latest research has revealed that circRNAs can act as micro-RNA (miRNA) sponges, protein translation templates and immunomodulators [6] and play a key role in regulating gene expression. Because circRNAs are closely related to disease development and are highly conserved and stable, they have become a hot spot in current research [7], especially their role as competitive endogenous RNA (ceRNA) and miRNA sponges. A large number of studies have confirmed that circRNAs can sequester miRNAs through complementary RNA base pairing and prevent miRNAs from binding to their mRNA targets [8], thus forming a circRNA-miRNA-mRNA network that interferes with disease occurrence and development. Studies have shown that the sponge mechanism of circRNAs is widely present in many diseases. For example, in gynecologic cancers [9,10], liver cancer [11], bladder cancer [12], gastric cancer [13], breast cancer [14], glioma [15], hematological malignancies [16][17][18], and diabetes [19], among other diseases, circRNAs have been reported as promising tumor biomarkers and therapeutic targets. At present, upregulation or downregulation of circRNA expression has also been found in MM, and miRNAs have been indicated to be functionally deregulated or abnormally expressed in MM cells [20], suggesting that circRNAs can interfere with myeloma cell proliferation and MM development through a sponge mechanism [21,22]. In addition, studies have found that circRNAs play an important role in regulation of cardiac function [23,24]. Therefore, circRNAs have broad application prospects in MM and MM-related cardiac damage.
Exosomes are small extracellular vesicles (EVs) [25] with a diameter between 30 and 100 nm. In humans, exosomes are widely present in various body fluids, such as saliva, serum, plasma, urine and malignant tumor effusion. Exosomes express various characteristic surface markers, proteins and other substances and can transport circRNAs, mRNAs and other noncoding RNAs [26]. Exosomes transfer the molecules or biological information they carry to target cells through endocytosis or direct fusion with the target cell membrane [27], thereby mediating intercellular communication between the parental cell and neighboring cells and distant tissues. This effect is currently considered to be an important promoter of tumor disease progression [27]. In addition, exosomes can induce the activation, proliferation, differentiation and death of target cells. Therefore, they have high value in clinical research [28,29]. MM disease progression largely depends on the bone marrow microenvironment, and communication mediated by EVs is an important aspect of this environment. Studies have found that exosomal miRNAs play an important role in MM progression and targeted therapy [20,[30][31][32]. Unfortunately, research on the role of exosomal circRNAs (exo-circRNAs) in MM and MM-related heart damage is still relatively rare. Therefore, we sequenced exo-circRNAs in the serum of MM patients, constructed a circRNA profile and combined these data with patient clinical symptoms to explore the relationship between exo-circRNAs and MM-related myocardial damage. In addition, this work provides a new potential exo-circRNA target for MMrelated myocardial damage.

Clinical serum specimens
Twenty MM patients who were admitted to the Department of Hematology, Affiliated Hospital of Shandong University of Traditional Chinese Medicine, were enrolled in the study. Five healthy adult volunteers were included in the experiment as healthy controls (HCs). The peripheral blood of the subjects was collected with a vacuum coagulation tube, and the serum was obtained after centrifugation (5 min, 2500 rpm) and stored at − 80 °C. The study was conducted in accordance with the Helsinki Declaration revised in 2008, and informed consent was obtained from all participants. All experimental methods were approved by the Ethics Committee of the Affiliated Hospital of Shandong University of Traditional Chinese Medicine.

Extraction of exosomes
Exosomes were isolated using an ExoQuick exosome precipitation kit (System Bioscience, Palo Alto, CA, USA; EXOQ5A-1) according to the manufacturer's instructions. Briefly, 250 µl of sample was mixed with 63 µl ExoQuick exosome precipitation solution, incubated at 4 °C for 30 min, and then centrifuged (30 min, 1500 × g). The supernatant was aspirated, the sample was centrifuged again (5 min, 1500 × g), and the obtained pellet was resuspended in 100 µl PBS and stored at − 80 °C.

Morphological identification of exosomes
The exosomes were fixed with the same amount of 4% paraformaldehyde and washed with phosphate-buffered saline (PBS; pH = 7.4). Then, the samples were placed on formvar/carbon 200-mesh copper grids for 20 min at room temperature and fixed in 1% glutaraldehyde for 5 min. Next, the samples were stained with uranyl oxalate for 5 min, and 4% uranyl acetate and 2% methyl cellulose were added at a ratio of 1:9 on ice. Finally, the grids were imaged via transmission electron microscopy (TEM) (Hitachi, HT7700).

Identification of exosome marker proteins
The exosomes were dissolved in RIPA lysis buffer (Sparkjade, Jinan, China, EA0002), and total protein was extracted. The protein concentration was determined using a bicinchoninic acid (BCA) protein detection kit (Beyotime, Shanghai, China, P0011). The protein was separated via sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE; Sparkjade, Jinan, China; EC0004) and transferred to a polyvinylidene fluoride (PVDF) membrane (Sparkjade, Jinan, China, ED0004). The PVDF membrane was soaked in 5% skimmed milk for 2 h and then incubated with anti-calnexin, anti-TSG101 and anti-CD63 antibodies overnight at 4 °C. After incubation with secondary antibody for 1 h, the images were analyzed using an Alpha Innotech Fluor Chem Q Imaging Analysis System. The relative TSG101 and CD63 protein levels were normalized to the optical density of calnexin.

Western blotting
Cells were lysed in RIPA lysis buffer (Sparkjade,), and the protein concentration was measured with a BCA protein assay kit (Beyotime). Equal amounts of sample protein were separated in 10% SDS-PAGE gels and transferred onto PVDF membranes (Sparkjade). The samples were blocked with 5% skimmed milk at room temperature for 2 h and then incubated with anti-calnexin, anti-TSG101, anti-CD63 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-Toll-like receptor 4 (TLR4; Abcolonal; A5258), anti-LC3B (Abcolonal; A19665), anti-SQSTM1/ p62 (Abcolonal; A19700) or anti-Beclin 1 (Abcolonal; A7353) antibody overnight at 4 °C. After washing with PBS 3 times, diluted secondary antibody was added and incubated with the membrane at room temperature for 1 h, and the membrane was washed 3 consecutive times with PBS. Finally, an Alpha Innotech Fluor Chem Q imaging analysis system was used to visualize the bands. The relative TLR4, LC3, Beclin 1 and P62 protein levels were normalized to the optical density of Actin-1.

RNA extraction and real-time qPCR verification
Total RNA was extracted from exosomes using an RNA extraction kit (Sparkjade, TAKARA.9767), and the concentration and purity of extracted RNA were measured with a NanoDrop 2000 spectrophotometer (Thermo). RT-PCR and SYBR Green qPCR kits (AH0401) were used to verify the RNA (Sparkjade), and PCR was run on a Roche LightCycler 480 system. According to the procedures provided in the instructions, the circ-G024080 and hsa-mir-4268 expression levels, and the TLR4, LC3 and GAPDH mRNA levels were detected. Each sample was tested at least 3 times, and the changes in relative gene expression were calculated using the formula 2 −(ΔΔCt) . GAPDH was used as an internal control, and the results for each sample were normalized to GAPDH expression (See Table 1). The primer sequences were as follows: Table 1 Gene primer sequences Cell viability H9C2 (XYXB-1190) and U266 (EY-X1023) cells were obtained from ATCC (Rockville, MD, USA). H9C2 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), and U266 cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium supplemented with 10% FBS. The cells were used in subsequent experiments after reaching the logarithmic growth phase. U266 cell exosomes were extracted for the following experiment.
H9C2 cells were cultured in 96-well plates at 1 × 10 5 cells per well. The cells were divided into three groups: a control group (H9C2 only), exosome group (H9C2 cells incubated with U266 exosomes) and GW4869 group (H9C2 cells incubated with U266 exosomes and then with GW4869). The cells in the three groups were incubated at 37 °C and 5% CO 2 for 12 h. Subsequently, according to the manufacturer's instructions, a Cell Counting Kit-8 (CCK-8) assay (Beyotime, Shanghai, China, C0039) was used to determine cell viability. Each group was tested three times.

Bioinformatics analysis
Briefly, after identification of the significantly differentially expressed circRNAs between the MM group and the HC group, Cluster and TreeView software were used to perform hierarchical cluster analysis. GO and KEGG analyses were used to evaluate the functions and pathways of differentially expressed genes in the MM and HC groups. TargetScan and miRanda were used to predict interactions between circRNAs and downstream miR-NAs, and Cytoscape software (v2.8.0) was used to analyze and construct a ceRNA network.

High-throughput sequencing
Total RNA was isolated from the samples, and the quality of the library was determined using a BioAnalyzer 2100 system (Agilent Technologies, Palo Alto, CA, USA). According to the supplier's instructions, ribosomal RNA (rRNA) was removed from the samples using an NEBNext rRNA Depletion Kit (New England Biolabs, Ipswich, Massachusetts, USA), and then, a sequencing library was constructed with an NEBNext ® Ultra ™ II Directional RNA Library Prep Kit (New England Biolabs). The BioAnalyzer 2100 system was used for library quality control and quantification, and an Illumina HiSeq instrument (Illumina) was used for 150-bp paired read sequencing.

Luciferase reporter constructs and luciferase activity assay
Direct interactions between the partners in the circ-G042080-mediated ceRNA network were evaluated using a luciferase activity assay. A luciferase reporter vector (GP-mirGLO Dual-Luciferase miRNA Target Expression Vector; Promega) was used to produce luciferase constructs. circ-G042080 and the 3'UTR of TLR4 were cloned via RT-PCR. circ-G042080-WT, circ-G042080 mutant and TLR4-3'UTR (WT and mutant) were constructed. 293 T cells were seeded into 24-well plates and allowed to grow for 24 h without antibiotics before transfection. The cells were transfected with the constructed reporter vectors (300 ng) together with the miRNA (hsa-miR-4268) mimics or negative control mimics (100 nM) using Lipofectamine 2000 (2 μl). The cells were lysed ~ 24 h after transfection, and the luciferase activity was assayed using a Dual-Luciferase Reporter Assay System (Promega). Firefly luciferase activities were normalized to Renilla luciferase activities. All experiments were performed independently in triplicate.

Cell counting kit-8 (CCK-8) assay
The experimental grouping was the same as described in Section "Cell viability". After reaching the logarithmic growth phase, the cells were plated in a 96-well plate at 1 × 10 5 cells/well and cultured for 2 h in a 37 °C incubator. Then, 10 µl of CCK-8 solution was added to each well and incubated with the cells for 1 h. The OD value was detected at 450 nm using a microplate reader to calculate the cell proliferation rate. Relative cell proliferation rate = (experimental OD-control OD)/control OD.

Statistical analysis
The experimental data are presented as the mean ± standard deviation (SD). SPSS 26.0 software was used for statistical analysis. A t test was used for comparisons between two groups, and correlation analysis was used to evaluate the relationship between every two factors. Receiver operating characteristic (ROC) curves were used to determine the diagnostic sensitivity and specificity of disease-related indicators. Univariate and multivariate Cox regression models were used to determine factors affecting prognosis. Except for the univariate Cox regression model with P < 0.1 indicating statistical significance, in other statistical analysis methods, P < 0.05 indicated a significant difference.

Identification and analysis of exosomes from MM patients
In this study, 6 MM patients and 5 HCs were enrolled for high-throughput sequencing of serum exo-circRNAs. The serum exosomes were identified by TEM, western blotting (WB) and NTA. TEM showed that the exosomes were round and less than 120 nm in diameter. (Fig. 1A). Common protein expression markers, including CD9, CD63, CD81, and Tsg101, can be used to identify exosomes [33]. WB results demonstrated that what we collected from MM patients' serum were exosomes (Fig. 1B). Furthermore, NTA results indicated that the average diameter of exosomes was 63.84 ± 20.32 nm (Fig. 1C). Together, the above results confirmed that our samples were exosomes. High-throughput sequencing showed that there were 10,106 differentially expressed circRNAs in the MM group compared with the HC group, of which 10,321 were upregulated and 5785 were downregulated (Fig. 1D). Next, fold change (FC) ≥ 2.0 and P ≤ 0.05 were used as the criteria for significant differences, and a total of 1265 upregulated circRNAs and 787 downregulated circRNAs were identified (Fig. 1E).
We analyzed and summarized the identified circRNAs based on origin, length, location and category. Among all the differentially expressed circRNAs, only 604 were known and were available in circBase and other databases (Fig. 1F). The median lengths of upregulated and downregulated circRNAs in the MM group were 168 nt and 345 nt, respectively (Fig. 1G), indicating that MM exosomes mainly carry small circRNAs less than 500 nt in length. The results also showed that the differentially expressed circRNAs were distributed in 22 autosomes, sex chromosomes and mitochondria in humans. No specific chromosome of origin was found (Fig. 1H). In addition, most of the significantly upregulated circR-NAs in the MM group were sense overlapping circRNAs, while most of the downregulated circRNAs were exonic (Fig. 1I).

Bioinformatics analysis of differentially expressed circRNAs
GO and KEGG analyses were used to predict the functions and signaling pathways of exo-circRNAs in MM. Previous results showed that the number of upregulated circRNAs was significantly larger than that of downregulated circRNAs in the MM group. Therefore, we mainly analyzed the upregulated circRNAs in the MM group. GO annotations were used to identify the top 10 biological process (BP), cellular component (CC) and molecular function (MF) enriched terms, which may reveal the main roles of upregulated circRNAs ( Fig. 2A-C).
The results showed that activation of transmembrane receptor protein tyrosine kinase activity identified in the BP analysis was closely related to the resistance of myeloma cells. This activation can promote the survival and proliferation of myeloma cells and protect myeloma cells from apoptosis induced by bortezomib [34]. In addition, the CC term pericentral material was also closely related to MM. Excessive centrosome surrounding materials can cause abnormal mitosis and then affect myeloma cell apoptosis and cell cycle progression [35]. Moreover, histone methyltransferase activity (H3-K4 specific) identified in the MF analysis can also interfere with the occurrence of MM. Overexpression of the histone methyltransferase MMSET in t(4;14)+ MM patients is thought to be the driving factor in the pathogenesis of this myeloma subtype [36,37]. KEGG analysis results showed that exo-circR-NAs that were significantly upregulated in MM might regulate or participate in 38 signaling pathways. We selected the top 10 predicted signaling pathways based on the P value and selection counts (Fig. 2D). Among them, the "PI3K-Akt signaling pathway" plays an important role in myeloma cell proliferation, migration and apoptosis and can interfere with drug resistance in MM [38,39]. The results showed that 3 of the top 10 KEGG pathways were related to cardiac disease. In addition, the descriptions "glutamate receptor pathway" and "calcium ion pathway" appeared repeatedly in GO and KEGG analyses. Previous studies have confirmed that calcium ions play an important role in regulating heart function [40] . Abnormal levels of intracellular calcium ions can induce mitochondrial dysfunction and lead to heart failure [41,42]. Glutamate receptors can promote apoptosis of ischemic human cardiomyocytes through the p38/MAPK pathway mediated by calcium influx [43]. Moreover, the ionotropic glutamate receptors NMDA and NMDAR1 can cause lipid peroxidation damage through intracellular calcium overload, which induces cardiomyocyte death [44]. The above results indicate that exo-circRNAs in the serum of MM patients are likely to participate in the process of MMrelated myocardial damage through calcium ion pathways and glutamate receptor channels.

PCR verification of exo-circRNA expression in MM patients
Based on the previous exo-circRNA sequencing results in MM patients and HCs and using the P value and fold changes, we selected 5 upregulated circRNAs and 8 downregulated circRNAs for PCR verification.
The results were consistent with the high-throughput sequencing results. Through bioinformatics analysis, we found that the upregulated circRNA chr2: 2744264-2744443 + was closely associated with MM-related myocardial damage. Because it is an unknown circRNA, we named it circ-G042080 based on its gene location.
The PCR results showed that the average relative circ-G042080 expression level in the MM group was more than two times that in the control group, which was consistent with the high-throughput sequencing results (Fig. 3). This result further confirmed the high circ-G042080 expression in the serum exosomes of MM patients.

Construction of a circRNA-miRNA-mRNA interaction network and verification with a luciferase assay
circRNAs can affect miRNA function and miRNA expression by competing for miRNA binding sites. We predicted the downstream miRNAs and target genes of circ-G042080 using miRanda and TargetScan. The top 3 miRNAs with the strongest ability to interact with circ-G042080 were hsa-miR-4268, hsa-miR-764 and hsa-miR-3907, which we subjected to GO and KEGG analyses. Through KEGG analysis, among the top 10 enrichment pathways, we found that the Toll-like receptor signaling pathway was closely related to myocardial damage. Toll-like receptors can induce local heart inflammation and further aggravate myocardial damage [45] . In addition, arrhythmogenic right ventricular cardiomyopathy (ARVC), hypertrophic cardiomyopathy (HCM) and dilated cardiomyopathy, which were also among the top 10 enriched pathways, are also closely related to myocardial damage. Cytoscape software was used to construct a ceRNA network diagram consisting of circ-G042080, miRNAs and downstream target genes. Figure 4A shows that 28 downstream target genes may bind to the top 3 miRNAs and that 4 of these target genes are involved in the Toll-like receptor signaling pathway. Among downstream mRNAs, TLR4, a member of the Toll-like receptor family, was regulated by all three miRNAs. Therefore, we chose to explore TLR4 as a possible downstream target gene in subsequent experiments. A dual-luciferase reporter assay was performed to determine direct binding between circ-G042080/TLR4 and miR-4268 based on their complementary sequences (Fig. 4B). The luciferase activity results showed that hsa-miR-4268 mimics decreased the luciferase activity in the circ-G042080 WT group (P < 0.05) and TLR4 WT group (P < 0.05) but had no effect on the luciferase activity in the circ-G042080 MUT group (P > 0.05) and TLR4 MUT group (P > 0.05) (Fig. 4C, D). The above results indicate that hsa-miR-4268 might target circ-G042080 and TLR4 expression, suggesting that circ-G042080 can activate the Toll-like receptor signaling pathway to induce MM-related myocardial damage by regulating downstream miRNAs and the TLR4 axis.

Clinical relevance of the circ-G042080 level in MM-related myocardial damage
To further determine the clinical significance of circ-G042080 in MM-related myocardial damage, we analyzed the relationship between circRNA expression and the clinical characteristics of 20 MM patients. The results showed that circ-G042080 expression was positively correlated with TNT and proBNP levels (Fig. 5A). High circ-G042080 expression corresponded to a decrease in ventricular ejection fraction (EF) and systolic blood pressure (SPB) (Fig. 5B), which are markers of myocardial damage. Moreover, circ-G042080 expression was closely related to the clinical characteristics of MM-related myocardial damage, suggesting that circ-G042080 might be a reliable new biomarker of MM-related myocardial damage and a therapeutic target. In addition, the circ-G042080 level may also be an independent prognostic indicator of MM-related heart damage. The ROC curve showed that circ-G042080 levels had high sensitivity and specificity for predicting the prognosis of patients with MM-related myocardial damage, which suggests that circ-G042080 could be used as an effective clinical prognostic indicator. The area under the curve (AUC) of circ-G042080 was 0.909 (P = 0.05), corresponding to an optimal circ-G042080 cutoff value of 2.18 (Fig. 5C). Next, we further explored the relationship between clinical characteristics and survival rate using univariate and multivariate Cox regression analyses. Univariate Cox regression analysis showed that hemoglobin (HGB) < 100 g/L, D-S stage, NYHA stage, EF < 50%, NTproBNP ≥ 1800 and circ-G042080 ≥ 2.18 had significant negative effects on survival rate (P < 0.05). Multivariate Cox regression analysis showed that circ-G042080 ≥ 2.18 was an independent prognostic factor of death in patients with myeloma and cardiac failure ( Table 2). Through PCR, we further explored the exosomal circRNA levels in patients with MM-related myocardial damage. The results showed that the circ-G042080 level in myeloma patients with NYHA grades 3-4 was significantly higher than that in patients with NYHA grades 1-2 (Fig. 5D). In addition, the circ-G042080 level in patients with BNP ≥ 1800 was significantly higher than that in patients with BNP < 1800 (Fig. 5E). These results indicate that the circ-G042080 level could be used as a reliable new biomarker of damage in myeloma and a prognostic factor. Fig. 4 Construction of a circRNA-miRNA-mRNA interaction network and verification with a luciferase assay. A The circRNA-miRNA-mRNA network predicted by miRanda and TargetScan presented as a map constructed with Cytoscape software. B The binding sites between TLR4, hsa-miR-4268 and circ-G042080. C The luciferase assay results showed that hsa-miR-4268 mimics significantly reduced the luciferase activity in the circ-G042080 WT group compared with that in the HC group (P < 0.05) but had no effect in the circ-G042080 MUT group. D Luciferase activity results showed that hsa-miR-4268 mimics reduced the luciferase activity in the TLR4 WT group (P < 0.05) but had no effect on the luciferase activity in the TLR4 MUT group

Effects of exosomes from MM patients on H9C2 cells
We designed a cell experiment to further explore the role of circ-G042080 overexpression in MM-related myocardial damage. H9C2 cells are often used to explore cardiomyocyte apoptosis, autophagy and toxic damage. In our experiment, exosomes from U266 cells were used to treat H9C2 cells to study the effect of circ-G042080 overexpression on cardiomyocytes. PCR results showed that the levels of circ-G042080 and TLR4 in the exosome group of H9C2 cells were significantly increased compared with levels in the control group (P < 0.01), while the hsa-miR-4268 level was significantly decreased (P < 0.01). However, after blocking exosomes with GW4869, the circ-G042080, hsa-miR-4268 and TLR4 levels were restored in H9C2 cells treated with exosomes from U266 cells (Fig. 6A-C). These results indicate that the circ-G042080/hsa-miR-4268/TLR4 ceRNA axis could be present in H9C2 cells incubated with exosomes. We further explored the effect of exosomes from U266 cells on the function of H9C2 cells. CCK-8 assay results indicated that the proliferation rate of H9C2 cells in the exosome group was significantly reduced compared with that in the control group and that exosomes overexpressing circ-G042080 had a significant antiproliferative effect (P < 0.01). After GW4869 was used to block exosomes, the proliferation rate of H9C2 cells affected by the exosomes of MM patients was restored (Fig. 6D). PCR experiments confirmed that LC3 levels in the exosome group were significantly increased compared with those in the control group and that this effect was reversed after the exosomes were blocked with GW4869 (P < 0.01) (Fig. 6E). WB results showed increased LC3 and Beclin1 levels and a decreased P62 level in the exosome group (Fig. 6F). The LC3 expression level was observed via immunofluorescence, and the results showed that the expression level in the exosome group was significantly higher than that in the control group (Fig. 6G, H). Additional File (1, 2) are the amplification curve and dissolution curve of TLR4 respectively.
We further observed autophagic vesicles through electron microscopy, and the results showed that the number of autophagic vesicles in the exosome group was significantly greater than that in the control group (Fig. 6 I, J). The above results suggest that circ-G042080 in exosomes activated TLR4 in H9C2 cells through the ceRNA mechanism, thereby inducing autophagic death in cardiomyocytes.

Discussion
Exosomes are membrane-like structures that carry signaling molecules and have received extensive attention as important mediators of intercellular communication.
Previous studies have confirmed that by delivering abundant biologically active molecules (including miRNAs, growth factors, and signaling molecules) [46] exosomes can participate in the growth and drug resistance of several tumors, such as gynecologic cancers [47,48] and malignant glioma [49].In addition, exosomes have been verified to mediate bidirectional transfer of proteins, lipids and nucleic acids between the bone marrow microenvironment and MM and promote MM development by promoting angiogenesis, osteolysis and drug resistance [50]. Exosomes have broad prospects as biomarkers for early prediction of MM disease progression and drug resistance [51].
Amyloidosis is a serious complication of MM and is primarily caused by structural changes and extracellular deposition of monoclonal immunoglobulins [52]. In the clinic, cardiac involvement is the second most common manifestation after kidney involvement. Heart involvement usually has no characteristic clinical symptoms at the beginning. As the disease progresses, restrictive cardiomyopathy, heart failure, and abnormal rhythms will appear and then develop into heart failure [53], which seriously affects patient survival and quality of life. At present, pathological biopsy is the only way to diagnose myocardial amyloidosis in MM. Therefore, we urgently need a new early target index to diagnose and After blocking the exosomes with GW4869, the level of circ-G042080 decreased. B The level of hsa-miR-4268 was significantly increased in the exosome group (P < 0.01) and then decreased after blocking exosomes with GW4869. C, A The level of TLR4 was significantly increased in the exosome group (P < 0.01) and then decreased after blocking exosomes with GW4869. D CCK-8 assay results showed that the proliferation rate of H9C2 cells in the exosome group was significantly reduced (P < 0.01) and that the cell proliferation rate increased after the exosomes were blocked with GW4869. E The level of LC3 was significantly increased in the exosome group (P < 0.01) and then decreased after blocking exosomes with GW4869. F WB results showed increased LC3-II/I and Beclin1 levels and a decreased P62 level in the exosome group compared with the control group. G Immunofluorescence results showing the LC3 expression level in H9C2 cells. H Compared with that in the control group, the LC3 expression level in the exosome group was significantly increased (P < 0.01). I Observation of the number of autophagic vesicles in H9C2 cells via electron microscopy. The red marks from left to right indicate mitochondria, autophagic vesicles, autophagic lysosomes, mitochondria, and mitochondria. J The results showed easily visible autophagic vacuoles in the exosome group, and the number was obviously greater than that in the control group (See figure on next page.) treat myocardial amyloidosis and related heart damage in MM. circRNAs are noncoding RNAs that are widely present in human body fluids. Previously, circRNAs were though to result from random errors that occurred during transcription. As evidence accumulated, circRNAs were shown to serve as potentially important prognostic and diagnostic factors in cancer due to their high stability and specific expression [54,55]. A large amount of evidence has also shown that circRNAs are key regulators of MM [56]. In addition, circRNAs are closely related to heart failure and can predict heart failure development and aid in risk stratification [57]. A previous study by our team found that exo-circRNAs have excellent diagnostic and therapeutic potential for MM-related peripheral neuropathy (PN) [58]. Unfortunately, research on the relationship between MM-related myocardial damage and exosomes is still relatively rare.
To further explore the mechanism of action of MMderived exosomes in MM and MM-related myocardial damage, we extracted exosomes from the serum of 6 MM patients and 5 healthy individuals and studied the expression levels of exo-circRNAs ( Fig. 1A-C). The data showed that 2052 circRNAs were differentially expressed between MM patients and HCs (Fig. 1D, E). We focused on the characteristics of exo-circRNAs that were significantly upregulated in MM. GO and KEGG analyses indicated the role of these circRNAs in the pathogenesis of MM progression and myocardial damage. Among the enriched pathways, the "PI3K-Akt signaling pathway" plays an important role in myeloma cell proliferation, migration and apoptosis [38,39] (Fig. 2). In addition, bioinformatics analysis predicted that circ-G042080 might regulate a downstream miRNA/TLR4 axis (Fig. 4). At present, TLR4 has been identified as a crucial regulator in cardiovascular disease. Studies have found that TLR4 activation in rats promotes cardiac inflammation, mitochondrial dysfunction, apoptosis and fibrosis [59,60]. In addition, the "glutamate receptor pathway" and "calcium ion pathway" were often identified in GO and KEGG analyses. Previous studies have confirmed that calcium ions play an important role in regulation of heart function [40], and glutamate receptors can promote the death of ischemic human cardiomyocytes through the p38/ MAPK pathway mediated by calcium influx [43] (Fig. 2).
These findings indicate that exo-circRNAs in the serum of MM patients are likely to participate in the development of MM-related myocardial damage through their effect on calcium channels and glutamate receptor channels. This suggests that circ-G042080 might regulate downstream miRNAs and TLR4 to induce heart damage through a molecular sponge mechanism.
We further explored the relationship between circ-G042080 and the clinical characteristics of MM patients.
The results showed that circ-G042080 was closely related to the clinical features of heart damage (Fig. 5A, B), which suggested that circ-G042080 is a promising marker of MM-related heart damage. The ROC curve suggested that circ-G042080 could be used as an effective clinical prognostic indicator (Fig. 5C). Univariate and multivariate Cox analyses further confirmed that circ-G042080 was an independent prognostic factor for the death of patients with myeloma and heart failure ( Table 2). Subsequently, we treated H9C2 cells with exosomes derived from MM patients and found that the levels of circ-G042080 and TLR4 in these cells increased and that the level of hsa-miR-4268 decreased. These effects were reversed after exosomes were blocked with GW4869 ( Fig. 6A-C). The sphingomyelinase inhibitor GW4869 is a recognized exosome inhibitor and is thought to inhibit exosome secretion. We further explored the effect of exosomes on the function of H9C2 cells. CCK-8 assays revealed that the proliferation rate of H9C2 cells treated with exosomes derived from MM patients was significantly reduced and that the proliferation rate increased after the exosomes were blocked with GW4869. LC3, Beclin1, and P62 are all autophagy-related markers. LC3 is a commonly used autophagy marker. LC3-I is converted into LC3-II when autophagy is initiated. LC3-II, which is the structural protein of the autophagosome, can be adsorbed on the autophagosome membrane. The ratio of LC3-II/I can reflect the autophagy level. In addition, the autophagy-specific gene Beclin1 can participate in the occurrence of a variety of tumors by regulating the strength of autophagy [61,62]. Moreover, P62 can interact with LC3 to target it for degradation in autophagic lysosomes [63].
Autophagy is a highly conserved catabolic process used to remove abnormal cytoplasmic components, including protein aggregates and damaged organelles [64]. Inflammatory factors participate in the inflammatory response mainly through TLRs. There are many subtypes of TLRs, of which TLR4 is the most highly expressed in the heart [65]. In recent years, studies have found that there is a close relationship between autophagy and inflammation, and autophagy has been shown to be an important component of the innate immune response [66]. Studies have confirmed that TLR4 can be bound by endogenous ligands to recognize cell surface antigens and then activate autophagy [67]. Our experiments demonstrated that the autophagy level was significantly increased in H9C2 cells cocultured with MM patient exosomes and that GW4869 blocked this effect (Fig. 6E-J). The above results suggested that circ-G042080 in exosomes could activate H9C2 cells to overexpress TLR4 through the ceRNA mechanism and then induce autophagic death of cardiomyocytes. Overall, our research showed that exo-circRNAs have broad application potential in early diagnosis of MM myocardial amyloidosis and in targeted therapy for MM-related myocardial damage. However, in further explorations of the usability of exo-circRNAs in MM-related myocardial damage, the patient sample size must be increased. In addition, we only employed in vitro experiments to explore circR-NAs and did not use in vitro models to conduct further studies, which is the plan for our future work.

Conclusion
Our study found that exo-circRNAs are abundant in the serum of MM patients. Bioinformatics analysis showed that upregulation of circ-G042080 expression in exosomes might interfere with MM-related myocardial damage by regulating miRNA and the TLR4 axis, which can influence autophagy in the myocardium. Moreover, the circ-G042080 expression level was positively correlated with the clinical features of heart damage in MM patients and thus could be used as an independent prognostic indicator of myocardial damage. Cell experiments further confirmed that the circ-G042080/hsa-miR-4268/TLR4 ceRNA axis might exist in H9C2 cells incubated with exosomes and can cause an increase in the autophagy level in cardiomyocytes. In summary, our results indicate that serum exo-circRNAs have the potential to serve as new biomarkers of MM-related myocardial damage and might be therapeutic targets. However, the mechanism remains to be further explored.