Exosomal miR‐3131 derived from endothelial cells with KRAS mutation promotes EndMT by targeting PICK1 in brain arteriovenous malformations

Abstract Aims To explore the underlying mechanism by which low‐frequency KRAS mutations result in extensive EndMT occurrence. Methods Exosomes derived from primarily cultured brain arteriovenous malformation (bAVMs) and human umbilical vein endothelial cells (HUVECs) transfected with KRASG12D, KRASWT, or KRASNC lentiviruses were isolated, and their effects on HUVECs were identified by western blotting and immunofluorescence staining. The expression levels of exosomal microRNAs (miRNAs) were evaluated by miRNA microarray, followed by functional experiments on miR‐3131 and detection of its downstream target, and miR‐3131 inhibitor in reversing the EndMT process induced by KRASG12D‐transfected HUVECs and bAVM endothelial cells (ECs) were explored. Results Exosomes derived from KRASG12D bAVM ECs and KRASG12D‐transfected HUVECs promoted EndMT in HUVECs. MiR‐3131 levels were highest in the exosomes of KRASG12D‐transfected HUVECs, and HUVECs transfected with the miR‐3131 mimic acquired mesenchymal phenotypes. RNA‐seq and dual‐luciferase reporter assays revealed that PICK1 is the direct downstream target of miR‐3131. Exosomal miR‐3131 was highly expressed in KRASG12D bAVMexos compared with non‐KRAS‐mutant bAVMexos or HUVECexos. Finally, a miR‐3131 inhibitor reversed EndMT in HUVECs treated with exosomes or the supernatant of KRASG12D‐transfected HUVECs and KRASG12D bAVM ECs. Conclusion Exosomal miR‐3131 promotes EndMT in KRAS‐mutant bAVMs, and miR‐3131 might be a potential biomarker and therapeutic target in KRASG12D‐mutant bAVMs.


| INTRODUC TI ON
Brain arteriovenous malformations (bAVMs) are abnormal enlargements of vessels, wherein arterial blood flows directly into draining veins without the normal interposed capillary beds, while no brain parenchyma is contained within the nidus. 1 bAVMs usually occur in young people, presenting as seizures, neurological deficits, headaches, or devastating spontaneous intracranial hemorrhage. 2,3 Recently, low-frequency somatic KRAS mutations were detected in human sporadic bAVMs, [4][5][6][7][8] and in vivo experiments have confirmed that KRAS mutation can drive abnormal vascular morphology and arteriovenous malformation (AVM) development in mouse and zebrafish models. 9,10 Previous studies have found that somatic KRAS mutations also independently regulate endothelial-mesenchymal transition (EndMT) features. 11,12 EndMT is a process by which endothelial cells (ECs) can acquire a mesenchymal phenotype. 13 During EndMT, some EC markers can be lost or retained, while mesenchymal markers and morphology are acquired. [14][15][16] As EndMT contributes to a multitude of diseases, pharmacological modulation of the signaling pathways underlying EndMT may be effective as a therapeutic treatment. 17,18 A previous study reported that KRAS mutations occur in bAVM ECs with a low frequency ranging from 2.13% to 52.15%, 7

while
EndMT ECs are widespread in bAVMs. 11,12,19 Hao et al. 20 reported that during passaging, cultured bAVM ECs rapidly transition from a cobblestone morphology to a spindle-shaped morphology. This EndMT phenomenon in bAVMs can be manifested as only a decrease in endothelial markers or an increase in mesenchymal mark-

ers(ref). How low-frequency KRAS mutations result in widespread
EndMT occurrence among bAVM ECs remains unclear.
In this study, we utilized bAVM ECs with the KRAS G12D mutation and KRAS G12D -transfected human umbilical vein ECs (HUVECs) to explore the mechanism by which EC-EC communication induces EndMT in bAVMs. Furthermore, we explored the efficacy of a miR-3131 inhibitor in reversing this EndMT process.
The investigation of the underlying mechanism of the process will not only deepen the understanding of the pathogenesis of bAVMs but also contribute to the development of effective medical therapies for bAVMs.

| Culture and treatment of HUVECs
Commercially available HUVECs (#8000, ScienCell) were cultured in EC medium (ECM, #1001, ScienCell) according to the manufacturer's guidelines and were not used beyond passage 10

| Whole-exome sequencing (WES) and droplet digital PCR (ddPCR) validation of bAVM ECs
DNA extraction and whole-exome sequencing (WES) of primary cultured bAVM ECs were performed according to previous methods. 11 Briefly, commercially available kits (Gentra Puregene, QIAGEN) were used to isolate genomic DNA from bAVM ECs. We used two methods to examine the quality of DNA: DNA degradation and contamination were monitored on 1% agarose gels, and the concentration was measured with a Qubit® 2.0 Fluorometer (Invitrogen). We used a total of 0.6 μg of genomic DNA per case for library preparation. Sequencing libraries were generated, and index codes were added with an Agilent SureSelect Human All Exon V6 kit (Agilent Technologies) according to the manufacturer's recommendations. The index-coded samples were clustered on a cBot Cluster Generation System using a HiSeq PE Cluster Kit (Illumina). After cluster generation, the DNA libraries were sequenced on an Illumina HiSeq platform, and 150 bp paired-end reads were generated. The KRAS G12D mutation was further validated using droplet digital polymerase chain reaction (ddPCR). Ultimately, 6 out of 20 bAVM patients were found to have the KRAS G12D mutation in their ECs.

| Flow cytometry validation of EndMT in ECs
To assess the EndMT rate in bAVM ECs with KRAS mutation, we examined the typical EndMT markers CD31 and α-SMA in bAVM ECs using flow cytometry. Briefly, cell samples were collected using a standard method and incubated with an antibody against the cell surface protein CD31 (#303106, BioLegend) in cell staining buffer (#420201, BioLegend) for 15 min at room temperature in the dark.
Next, the cells were washed with DPBS three times and fixed using fixation buffer (#420801, BioLegend) for 40 min at 4°C. Then, the cells were washed with permeabilization wash buffer (#421002, BioLegend) three times and stained with an antibody against intracellular α-SMA (#197240, Abcam) at 4°C overnight. We used unstained cells as controls. With regard to gating, forward scatter (FSC) versus side scatter (SSC) was used to eliminate debris. In the following experiments, CD31 + αSMA-cells were recognized as normal ECs, whereas CD31α-SMA ± cells were recognized as EndMT cells.

| Isolation of exosomes from culture medium
KRAS G12D -mutant bAVM ECs, KRAS WT bAVM ECs, or HUVECs were cultured in exosome-depleted ECM for 48 h, and the culture supernatant was collected to isolate exosomes. To prevent contamination and interference of the transfected lentiviruses, the transfected HUVECs were washed three times with DPBS every passage and cultured for more than 7 days before exosomes were extracted as described in previous studies. Exosomes were isolated using differential ultracentrifugation as previously reported. 21 Briefly, the culture supernatant was centrifuged twice at 3000× g for 15 min at 4°C and then at 10,000× g for 30 min to remove cells and debris. The supernatant was further filtered with a 0.22μm filter (Millipore). The obtained medium was centrifuged at 110,000× g for 90 min at 4°C to pellet exosomes. The supernatant was discarded and finally resuspended again in PBS. We further confirmed the lack of eGFP expression in HUVECs cocultured with exosomes to verify that there was no lentivirus contamination.

| Identification of exosomes
The morphology of exosomes was observed by transmission electron microscopy (TEM, Tecnai G2 Spirit BioTWIN, FEI). Briefly, exosomes were fixed with 4% paraformaldehyde and spotted onto glow-discharged copper grids. The copper grids were dried for 5 min at room temperature. The samples were stained with 1% uranyl acetate for 1 min. Then, the samples were dried and observed at 80 kV.
The size distribution of exosomes was observed using nanoparticle tracking analysis (NTA; ZetaView S/N 17-310) and analyzed using OriginPro 8.5 (OriginLab). Western blot analysis was performed to detect the exosome markers CD63, CD81, and TSG101.

| RNA sequencing (RNA-seq)
The TRIzol method was used to prepare samples for RNA-seq.
RNA purity and concentration were measured using a Qubit® 2.0 Fluorometer (Life Technologies). An RNA Nano 6000 Assay Kit and a Bioanalyzer 2100 system (Agilent Technologies) were utilized to assess RNA integrity. RNA-seq was performed according to a previous method. 22 Each assay was repeated in three independent experiments.

| Statistical analysis
SPSS version 25.0, GraphPad Prism version 8.00, and R version 4.0.5 were used for statistical analyses. The K-S test for normality was used to assess data distribution. For normally distributed data, ttests or one-way ANOVA were used to assess differences between two groups or among three groups of quantitative variables. For data not normally distributed, nonparametric tests were used. A two-tailed probability value of 0.05 or less was considered to indicate statistical significance.

| The percentage of EndMT ECs in bAVMs is markedly greater than the percentage of ECs with KRAS mutation
To compare the percentage of EndMT ECs with the percentage of ECs with KRAS mutation in bAVMs, we simultaneously performed WES and flow cytometry analysis on primary bAVM ECs (n = 5) or HUVECs (n = 5). The allele frequency of KRAS G12D in bAVM ECs as determined by WES and confirmed by ddPCR is shown in Table S1.
The percentage of ECs with KRAS mutation was calculated according to a twofold allele frequency in WES. The corresponding EndMT percentage was obtained by flow cytometry ( Figure 1A, Figure S1A).
We found that the percentage of EndMT ECs was significantly higher than the percentage of ECs with KRAS mutation in the bAVMs, as shown in Figure 1B (paired t test, p = 0.010). Furthermore, we found only a small proportion of EndMT in normal ECs, and EndMT increased significantly in KRAS G12D bAVMs compared to normal ECs (p = 0.0079, Figure 1C). Our results support the idea that KRAS G12D may play a role in the EndMT phenotype transmitted to WT cells.

| Exosomes derived from bAVM ECs with KRAS G12D mutation promote EndMT in HUVECs
Previous studies have suggested that exosomes play a vital role in communication between cells. For example, in studies on the tumor microenvironment, exosomes secreted from KRAS-mutant prostate cancer cell lines have been found to contain small RNAs and to induce aggressive tumors in secondary recipients. 23 Therefore, we hypothesized that bAVM ECs with KRAS G12D mutations might secrete exosomes to promote EndMT in neighboring KRAS wt ECs. To test this hypothesis, exosomes were extracted from conditioned media of bAVM ECs with KRAS G12D mutation using differential ultracentrifugation. For identification of exosomes in bAVM samples (2, 5, and 6), we used TEM and NTA to determine that the exosomes had a characteristic cup-shaped appearance and a mean size of 50-150 nm ( Figure 2A). Additionally, the specific expression of the typical exosomal marker proteins CD63, CD81, and TSG101 was identified using western blot assay ( Figure 2B).
Then, in vitro experiments were performed to evaluate the effect of exosomes secreted by primary ECs derived from KRAS G12D bAVM ECs, and exosomes of normal HUVECs were used as control exosomes. As EndMT is widespread in bAVMs, 11,12  bAVM ECs were not used as controls. After coculture with exosomes from primary bAVM ECs for 3 days, the protein levels of the mesenchymal markers calponin, vimentin, fibronectin, and α-SMA were all significantly increased (p < 0.05, Figure 2C). Immunofluorescence staining also showed an increased level of α-SMA in HUVECs treated with bAVM exos derived from bAVM2 (p < 0.05, Figure 2D). These results revealed that bAVM ECs with the KRAS G12D mutation were able to secrete exosomes to promote EndMT in HUVECs.

| Exosomes derived from KRAS G12Dtransfected HUVECs promote EndMT in HUVECs
To further confirm the mediating effect of exosomes from KRAS G12Dmutant ECs on EndMT, HUVECs were transfected with KRAS G12D , KRAS WT , or KRAS NC lentiviruses. Then, exosomes were extracted and identified using differential ultracentrifugation. TEM and NTA identified 50-150 nm cup-shaped extracellular vesicles, and the typical markers CD63, CD81, and TSG101 were highly expressed in these exosomes ( Figure 3A,B).
Furthermore, the function of exosomes derived from KRAS G12Dmutant HUVECs was studied. After coculture with exosomes secreted by KRAS G12D ECs, western blotting showed increased protein levels of the mesenchymal markers calponin, vimentin, fibronectin, and α-SMA compared with HUVECs cocultured with exosomes secreted by KRAS WT or KRAS NC ECs (p < 0.05, Figure 3C, Figure S1A).
The mean intensity of α-SMA protein in the immunofluorescence staining assay also increased after coculture with KRAS G12D ECsecreted exosomes (p < 0.05, Figure 3D). These results revealed that exosomes secreted from KRAS G12D ECs promoted the EndMT process in HUVECs.
To further investigate the biological function of miR-3131 in the EndMT process, HUVECs were transfected with miR-3131 mimic, and western blotting was used to determine the protein expression level. After transfection with miR-3131, the protein levels of calponin, vimentin, fibronectin, and α-SMA were significantly increased (p < 0.05), and the miR-3131 inhibitor reversed this process (p < 0.05, Figure 4C, Figure S1A). The level of α-SMA protein was significantly increased after transfection with the miR-3131 mimic (p < 0.05), and this effect was reversed by the miR-3131 inhibitor, as identified using an immunofluorescence staining assay (p < 0.05, Figure 4D). These data demonstrated that miR-3131 in exosomes from KRAS G12D -mutant HUVECs could promote EndMT in HUVECs.

| PICK1 is the downstream target of miR-3131 in EndMT regulation
To further determine the downstream target of miR-3131, four databases (miRanda, RNA22, RNAhybrid, and TargetScan) were utilized to predict the potential downstream targets of miR-3131, and RNA-seq was used to assess the mRNA changes in HUVECs after transfection with the miR-3131 mimic. A total of 536 potential genes were predicted to be direct downstream targets of miR-3131 in all four databases, while 1622 genes were significantly downregulated in the RNA-seq expression profile of miR-3131 mimic-transfected HUVECs compared with NC mimic-transfected HUVECs (fold change > 1.2, p < 0.05, Figure 5A). Among them, 41 genes were identified as overlapping genes, and PICK1 has been reported to be a negative regulator of the TGFβ pathway, which  plays a vital role in EndMT by promoting caveolin-dependent degradation of the TGFβ type I receptor. [29][30][31] To confirm the mRNA level of PICK1 in miR-3131-transfected HUVECs, RT-PCR was used and identified significantly lower mRNA levels of PICK1 in miR-3131 mimic-transfected HUVECs (fold change = 0.77, p = 0.028, Figure 5B). To substantiate the site-specific repression of PICK1 by miR-3131, we constructed a mutated PICK1 3′ UTR luciferase reporter. The dual-luciferase reporter assay showed that the luciferase activity of PICK1 with the WT 3'UTR was significantly suppressed with an efficiency of 39% in miR-3131-expressing HEK 293FT cells (p < 0.05). In contrast, the luciferase activity of PICK1 with the 3′UTR mutation was not changed ( Figure 5C). These results indicate that PICK1 is the downstream target of miR-3131.
To further investigate the function of PICK1 in ECs, we knocked down PICK1 in HUVECs using siRNA. After transfection with PICK1 siRNA, the protein levels of the mesenchymal markers calponin, vimentin, fibronectin, and α-SMA were all increased in PICK1knockdown HUVECs (p < 0.05, Figure 5D, Figure S1A). These findings indicated that PICK1 was the direct downstream target of miR-3131, and that the knockdown of PICK1 promoted EndMT in HUVECs.

| DISCUSS ION
In this study, we found that bAVM ECs with KRAS G12D mutations are able to secrete exosomes to promote EndMT in neighboring KRAS WT ECs. Furthermore, we found that exosomal miR-3131 derived from ECs with KRAS mutations exerts this effect by targeting PICK1. Our results imply that miR-3131 is a potential biomarker, and that miR-3131 inhibitors are promising candidate treatment agents for KRASmutated bAVMs.
Intercellular communication through either cell-to-cell contact or paracrine effects is believed to be a key process in vascular remodeling. 34  MiRNAs have been observed in secreted exosomes, and many cells can secrete miRNAs via exosomes to exert their regulatory effects on recipient cells. 21 miRNAs are small single-stranded noncoding RNA molecules that bind to mRNA, promoting cleavage and subsequent degradation of the mRNA. 40 The exosomal transfer of miRNAs is a novel mechanism for intercellular communication.
Modification of the function of miRNAs to target multiple mRNAs and regulate gene expression is a new therapeutic approach for cancer. 41 Previous studies have indicated that overexpression of miR-3131 promotes proliferation and inhibits apoptosis in hepatocellular carcinoma (HCC) cell lines and that miR-3131 may act as a proto-oncogene in HCC. 42 The miR-3131 downstream target gene PICK1 can regulate the TGFβ signaling pathway by promoting caveolin-dependent degradation of the TGF-beta type I receptor. 29 Decreased expression of PICK1 regulates the metastasis of cancer cells and is associated with EndMT. 31 Our study indicates that exosomal miR-3131 in the microenvironment plays a crucial role in the EndMT process, miR-3131 is a potential biomarker, and miR-3131 inhibitors might be promising candidates for the medical treatment of KRAS-mutated bAVMs. However, further research is needed to determine how KRAS mutation induces ECs to secrete miR-3131 in exosomes, which are themselves potential targets for bAVM treatment.

CO N FLI C T O F I NTER E S T S TATEM ENT
None.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.