Patients and sample collection

Plasma was collected from 21 RMS patients at diagnosis and 13 healthy children (HC) at Pediatric Haematology/Oncology and Cell and Gene Therapy Department, Bambino Gesù Children’s Hospital. Among the RMS patients 9 were diagnosed with fusion positive ARMS and 12 with ERMS, 8 were females and 13 males; their median age was 49 months (range: 5–188 months). Among HC, 8 were males and 5 females, their median age was 74,9 months (range: 6–207 months). Patients’ clinical information is shown in Supplementary Table 1. Written informed consent was signed by all parents and the study was approved by our Institutional Ethics Committee (protocol number 1189_OPBG_2016).

Whole blood was collected in EDTA tubes (BD Vacutainer, Reading, UK) and processed within 2 h. The samples were first centrifugated at 500 × g for 10 min, and then supernatants were collected and centrifuged at 3000 × g and then at 12,000 × g for 20 min each. All the centrifugation steps were performed at 4 degrees. The plasma was collected and stored at -80◦C until EVs isolation.

Isolation of extracellular vesicles from plasma

EVs isolation from plasma was performed using the commercial kit miRCURY™ Exosome isolation kit-serum and plasma (Qiagen) according to the manufacturer’s protocol. Briefly, 3 UI of Thrombin was added to 0.6 ml of plasma and incubated for 5 min at room temperature (RT) and centrifuged for 5 min at 10,000 × g. An amount of 0.5 ml of supernatant was collected, 200 μl of precipitation buffer A was added, resuspended by vortexing for 5 s to mix and incubated for 60 min at 4 °C. After incubation, samples were centrifuged for 5 min at 500 g at RT and the supernatants were removed and discarded. Pellets were re-suspended by vortex in 270 μl resuspension buffer. The isolated EVs were characterized following the recommendations of “Minimal Information for Studies of Extracellular Vesicles” (MISEV) 2023, (Supplementary Methods) [25]. Transmission Electron Microscopy (TEM) confirmed the presence of EVs with homogeneous morphology, occasionally clustered, with a size ranging from 30 to 200 nm (Supplementary Fig. 1A). NanoSight analysis showed a mean vesicle diameter ranging from 105 to 146 nm (Supplementary Fig. 1B). Western Blot revealed an enrichment of the EVs-specific protein Tumor Susceptibility Gene 101 (TSG101), CD9, and CD63 in nanovesicles samples compared to Hela cells lysate; furthermore, the absence of endoplasmic reticulum protein Calnexin demonstrate that no cell debris were present in our preparation’s lysate (Supplementary Fig. 1C). The purified EVs samples were then processed for RNA extraction.

RNA isolation from plasma extracellular vesicles

RNA from plasma EVs was isolated using miRCURY RNA isolation kit-biofluids (Qiagen) according to the manufacturer’s protocol. Briefly, 300 μl of resuspended EVs were mixed with 90 μl Lysis solution biofluids (BF), vortexed for 5 s and incubated for 10 min at RT. 1 μl of RNA spike-in template mixture (miRCURY LNA™ Universal RT microRNA PCR, RNA spike-in kit) was added to each sample for downstream PCR analysis. Then, 30 μl Protein precipitation solution BF was added to samples and vortexed, incubated for 1 min at RT and centrifuged for 3 min at 11,000 g. The supernatants, after addition of 400 μl isopropanol, were vortexed for 5 s and then loaded in miRNA mini spin column BF. Columns were incubated for 2 min at RT, centrifuged for 30 s at 10,000 g, washed with Wash solution 1 BF and twice with Wash solution BF 2. Columns were centrifuged for 2 min at 11,000 g to dry membranes and RNA was eluted adding 30 μl RNase free H2O directly onto the membrane of the spin columns BF. Columns were incubated for 1 min at RT and then centrifuged for 1 min at 11,000 g. The purified RNA samples were stored at -80 °C.

qPCR assessment of extracellular vesicles miRNAs

Total RNA extracted from plasma exosomes was mixed with two artificial RNAs (RNA spike-ins as RT controls) and the final mixture (10 μl) was reverse transcribed at 42 °C for 60 min using the miRCURY LNA™ Universal RT cDNA Synthesis Kit (Qiagen) following the manufacturer’s instruction. The expression level of plasma EVs-miRNAs was evaluated by Serum/Plasma Focus microRNA PCR panels (Qiagen). The amplification curves were filtered (Ct < 36), imported into the GenEx software (ver.5, Qiagen) and normalized by global mean. The expression level (fold change [FC]) was calculated by taking the mean of individual Cq values for each group (HC, ERMS and ARMS patients). To validate the significant EVs-miRNAs in the plasma of an independent cohort of patients, we individually assayed mature miR-486-5p (cat.no. 339306-YP00204001), miR-17-5p (cat.no. 339306-YP02119304), miR-197-3p (cat.no. 339306-YP00204380), miR-483-5p (cat.no. 339306-YP00205693), miR-766-3p (cat.no. 339306-YP00204499), hsa-miR-132-3p (cat.no. 339306-YP00206035), hsa-miR-454-3p (cat.no. 339306-YP00205663), hsa-miR-484 (cat.no. 339306-YP00205636) and hsa-miR-335-5p (cat.no. 339306-YP02119293), by employing two endogenous miRNAs, namely miR-23a-3p (cat.no. 339306-YP00204772) and miR-320a (cat.no. 339306-YP00206042) that were selected by running Genorm and NormFinder analysis tools. QuantStudio 12 K Flex Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA) was employed for all the qPCR quantifications and the fold change was calculated by the 2−ΔΔCt method [26]. At least two independent amplifications were performed for each probe on triplicate samples. The raw Cq values from amplification curves (Serum/Plasma plates) were normalized by global mean using the GenEx qPCR analysis software (Exiqon ver 5), individual assays were normalized by taking miR-23a-3p and miR-320a as endogenous controls. Statistically significant (p < 0.05) miRNAs were obtained by ANOVA test (ERMS patients and ARMS patients versus controls). MiRNAs with a FC lower than -2 (FC < -2) and greater than 2 (FC > 2) in RMS patients and with a p-value lower than 0.05 compared to controls were considered highly dysregulated and retained for further bioinformatics analysis.

Bioinformatics analysis of Gene Expression Omnibus (GEO) dataset

A survey on the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/gds) repository was made to identify the publicly available miRNA expression datasets associated with RMS patients. The miRNA microarray-based expression data matrix from a cohort of 49 RMS Formalin-fixed paraffin-embedded tissues (primary not pre-treated tumors) collected at Fondazione IRCCS Istituto Nazionale dei Tumori (Milan, Italy) was retrieved for further analysis (ID: GSE135518). In this dataset, miRNA profile was performed using a SurePrint G3 Human miRNA r21 microarrays (Agilent) designed on miRBase 21.0 (miRNA). GSE135518 includes 27 pediatric RMS (0–14 years) and 22 AYA RMS (15- + 30 years)] as well as 13 normal tissue counterparts (CTRL) [27]. Primary data were collected using Agilent’s Feature Extraction software v10.7 (Agilent Technologies), background corrected, and quantile normalized using Bioconductor limma implementation in R. Differentially expressed miRNAs in the tissue sample between ARMS, ERMS and CTRL were identified imposing log2|FC|> 2 and adjusted p < 0.05 by using the GEO2R bioinformatics tool. Expression of miR-335-5p was retrieved from the data matrix and median expression was used to stratify patients for Kaplan–Meier analysis having overall survival as clinical endpoint. Patients’ clinical information’s are shown in Supplementary Table 2.

Bioinformatics target prediction of miR-335-5p and gene ontology

Target prediction of the miR-335-5p, was carried out by already reported procedure [28] that integrates the predictions of three different algorithms (i.e., TargetScan, MiRanda, and PITA). The list of target genes predicted in all databases were used for bioinformatics analysis. Gene Ontology (GO), KEGG pathway enrichment analysis, and annotation by DAVID bioinformatics tool [29] were performed, to determine the biological processes and signaling pathways in which the predicted targets of miR-335-5p were involved.

In situ hybridization (ISH)

Tumor sample paraffin embedded from 16 patients (10 ERMS and 6 ARMS) were obtained from the archives of Operative Unit of Pathology at Bambino Gesù Children’s Hospital. Formalin-fixed paraffin-embedded tumor samples were cut in RNAse-free environment at 5 µm thick, mounted on positive-charged slides. MiRNA in situ hybridization was performed as previously described [15]. Slides were analyzed by light microscopy with (Eclipse E600, Nikon). Each slide was scored by 2 independent, qualified observers, blinded on patient’s clinical information’s. miR-335-5p expression was semi-quantitatively evaluated based on staining intensity and distribution using a total score as follows: intensity score × proportional score. The intensity score (IS) was predefined as follows: 0, negative; 1, weak; 2, moderate and 3, strong. The proportional score (PS) was defined as follows: 0, negative; 1, < 10%; 2, 10%–50%; 3, > 50% positive cells [30].