Exploration of miR-326 delivery via biocompatible polymeric nanoparticles in medulloblastoma: A preliminary study Authors: Elena Splendiani, Gabriele Maiorano, Zaira Spinello, Davide Vecchiotti, Giorgia Gugliuzza, Marta Moretti, Angela Asquino, Mariapaola Izzo, Agnese Po, Nicole Feverati, Tiziana Raia, Zein Mersini Besharat, Daniela Verzella, Stefania Mardente, Alessandra Zicari, Giuseppe Ciccarella, Anna Citarella, Francesca Zazzeroni, Enrico De Smaele, Mary Anna Venneri, Angela Mastronuzzi, Giuseppe Gigli, Adriano Angelucci, Giuseppina Catanzaro, Ilaria Elena Palamà, Elisabetta Ferretti Published: 2026 ABSTRACT MicroRNAs play a crucial role in tumorigenesis, functioning as oncogenes or tumor-suppressors while orchestrating the differentiation states of cancer cells. In medulloblastoma, the most common malignant pediatric brain tumor, the downregulation of miR-326 is associated with a loss of tumor suppression that leads to uncontrolled proliferation and survival or de-differentiation. In this study, we explore an innovative approach for microRNA delivery using polycaprolactone/chitosan-based nanoparticles. We evaluate the biocompatibility of the nanocarrier in medulloblastoma cell lines and report results on in vivo nanoparticles administration, demonstrating minimal toxicity in the most exposed organs. In vitro findings reveal that nanoparticles-mediated miR-326 delivery significantly reduces stemness markers (Nanog and Sox2) and promotes neuronal differentiation, as evidenced by increased β3-tubulin expression. These results highlight polycaprolactone-based nanoparticles as a promising platform for miR-326 delivery, establishing a foundation for future investigations on therapeutic strategies in medulloblastoma. Introduction Non-coding RNAs (ncRNAs) dysregulation, including microRNAs (miRNAs), influences crucial cellular processes and has been implicated in the pathogenesis of tumors.1,2 Indeed, loss of miRNAs expression is a hallmark of tumors and sustains an undifferentiated cancer cell phenotype.3 Low levels of miRNAs favor pro-tumorigenic signaling pathways, which drive tumor progression and maintain the stem cell pool within the tumor bulk.4 Thus, re-expression of downregulated miRNAs can be a useful strategy to reshape cancer cells and make them responsive to targeted treatments.5,6,7 Loss of miR-326 expression is a molecular feature of medulloblastoma (MB),8 and promotes activation of oncogenic signaling pathways and uncontrolled proliferation.5,9,10 MB treatments are multimodal, achieving a 70% survival rate according to international protocols; however, they are associated with significant long-term effects that impact the quality of life of survivors. At the same time, 30% of children who experience relapse lack curative treatment options.8,11,12,13 Therefore, there is an urgent medical need for innovative and more effective personalized therapeutic approaches to improve outcomes and minimize therapy-related adverse effects. Restoration of miR-326 could provide new opportunities for intervention in MB therapy. Several strategies have been developed to modulate miRNA function for therapeutic purposes through miRNA re-expression, by taking advantage of “mimics”. However, naked mimics are highly susceptible to degradation in the bloodstream, necessitating the use of a carrier for effective delivery. Nanotechnology-based approaches have emerged as promising in various disease applications.14 Among nanoparticles, polymeric nanoparticles (NPs) offer several advantages for nucleic acid delivery.15,16 The polycation chitosan, when complexed with nucleic acid ensures transduction efficiencies for RNA equal to or higher than the state-of-the-art chemical methods. Moreover, these complexes are easy to prepare, show significant RNA-loading capacity, and show low immunogenicity.17 Additionally, aliphatic polyesters, such as poly(ε-caprolactone) (PCL), are biocompatible materials extensively employed in biomedical applications that are revealed to be suitable for payload delivery due to their capability for controlled and sustained drug release.18 This controlled release mechanism is particularly beneficial for targeting tissues and tumors, where NPs can exploit the enhanced permeability and retention effect due to leaky vasculature.19 In this study, we propose the delivery of miR-326 through biocompatible nanoparticles (miR-326-NPs) to MB cells. These nanoparticles are formed by miR/chitosan complexes, stabilized with polyvinyl alcohol (PVA), a nonionic surfactant that provides steric stabilization to maintain colloidal dispersion.20 The nanometer-scale structure of these complexes is achieved using the aliphatic polyester PCL. The ratio of negatively charged miR and positively charged chitosan was carefully optimized to produce nanoparticles with a surface charge near to neutrality. This approach allowed us to employ the lowest amount of chitosan for promoting the intracellular release of miR in the target cells. However, in order to enhance the cellular uptake of this nano formulation, a small amount of polyethylenimine (PEI) was then added as the final step of nanofabrication, thus achieving positively charged NPs.21,22 The employment of a small amount of PEI in the nanoparticle fabrication could reveal an useful strategy to decrease PEI toxicity, with consequent improvement of NP biocompatibility.23 We demonstrate that these NPs are effectively delivered into cells without toxicity and can release their functional payload, triggering differentiation along with stemness and apoptosis impairment to fulfill their tumor suppressor function. Results NPs preparation NPs were prepared by utilizing the properties of chitosan to efficiently complex negatively charged nucleic acids, due to its cationic nature. To attribute any specific biological effect to the miR326 sequence itself, three formulations were prepared: (1) miR-326 loaded NPs (miR-326-NPs), (2) scramble control miRNA loaded nanoparticles (miR-NT-NPs), and (3) NPs without miRNA in their structure, named empty NPs. The loading ability of our NPs was 100%, corresponding to a miRNA: NP weight ratio of 1:1,453. These rates were optimized to guarantee reliable dosing, reproducibility, and efficient miRNA delivery within batches. As reported in Figure 1, NPs were in the nanoscale range, with low poly-dispersion indexes (PDI) ranging from 0.09 to 0.22, indicating a narrow distribution range of NP diameters almost tending to mono-dispersions shown by dynamic light scattering (DLS) analysis (Figure 1C). Hydrodynamic diameter, assessed after incubation in tissue culture medium, ranged from 400 to 550 nm.Figure 1NPs preparation(A) DLS analyses and (B) ELS analyses of miR-326-NPs, miR-NT-NPs, and empty NPs; (C) table reporting the poly-dispersion indexes (PDI), mean hydrodynamic diameter and the surface charge expressed as zeta-potential values for the investigated nanoformulations. (D–F) Representative SEM images of miR-326-NPs, miR-NT-NPs, and empty NPs, respectively. Scale bars: 1 μm. NPs preparation (A) DLS analyses and (B) ELS analyses of miR-326-NPs, miR-NT-NPs, and empty NPs; (C) table reporting the poly-dispersion indexes (PDI), mean hydrodynamic diameter and the surface charge expressed as zeta-potential values for the investigated nanoformulations. (D–F) Representative SEM images of miR-326-NPs, miR-NT-NPs, and empty NPs, respectively. Scale bars: 1 μm. This size range arises from the specific requirements of our cationic formulation: strong electrostatic complexation between chitosan and the miRNA leads to the formation of polycation-polyanion complexes and a larger hydrodynamic radius, consistent with previous reports on chitosan-based systems.24,25 This charge density is necessary for stable condensation and protection of miRNAs in biological media. Although many gene-delivery systems target NP sizes below 200 nm, multiple studies have shown that NPs in the 300–600 nm range can still achieve efficient cellular internalization via macropinocytosis, adsorptive uptake, or receptor-mediated mechanism pathways particularly active for cationic carriers, such as PEI and chitosan derivatives.26,27 Moreover, several reports indicate that larger cationic NPs can cross endothelial barriers, including the blood-brain barrier (BBB), through charge-mediated or ligand-facilitated transcytosis.28,29 Our NPs demonstrated robust cellular uptake and miRNA-delivery performance, suggesting that their hydrodynamic diameter did not inhibit functional activity within our system. The highly positive surface charge of these NPs, resulting from the use of chitosan as a complexing agent and PEI as a coating agent that shields miRNA and protects it from degradation and interaction with other molecules, yields zeta-potential values between +30 and +36 mV (Figures 1A and 1B). These positive values correlate with efficient cellular uptake. Scanning electron microscopy (SEM) images reported in Figures 1D–1F show that NP dispersions have irregular circular shapes. As shown in Figure 2A, after the reconstitution of NPs from lyophilized powder the mean hydrodynamic diameters in the three samples are quite similar to the fresh ones (shown in Figure 2B), indicating that the structure of NPs is retained during the preservation process. To assess encapsulation efficiency, a gel retardation assay was conducted. Figure 2C demonstrates that both freshly prepared and reconstituted NPs effectively inhibited miRNA migration due to strong complexation, which also blocked ethidium bromide intercalation. Conversely, control free miRNA migrated freely through the gel. Gel retardation assays further demonstrated the stability and payload integrity of these NPs in cell culture medium. Absorbance measurements of centrifuged miR-loaded NPs confirmed the results, showing no free miRNA in the supernatants and 100% of miRNA molecules were efficiently loaded (data not shown). The aforementioned quality control assays confirmed that both miR-326-NPs and miR-NT-NPs efficiently encapsulated miRNA and maintained their cargo after lyophilization and subsequent dispersion in cell-culture medium.Figure 2NPs miRNA encapsulation and stability(A) DLS analyses of miR-326-NPs, miR-NT-NPs, and empty NPs reconstituted form lyophilized powder. (B) DLS analyses of miR-326-NPs, miR-NT-NPs, and empty NPs incubated in cell culture medium for 1 h. (C) Agarose retardation assay of miRNA loaded NPs. Lane 2: free miR-326; lane 4: free miR-NT. Lanes 3a, 5a, and 6a: miR-326-NPs, miR-NT-NPs, and empty NPs freshly prepared. Lanes 3b, 5b, and 6b: miR-326-NPs, miR-NT-NPs, and empty NPs reconstituted from lyophilized powder. Lanes 3c, 5c. and 6c: miR-326-NPs, miR-NT-NPs, and empty NPs after incubation with fully supplemented cell culture medium for 30 min. Lane 1: ssRNA ladder (1,000; 500; 300; 150; 80; and 50 bases). NPs miRNA encapsulation and stability (A) DLS analyses of miR-326-NPs, miR-NT-NPs, and empty NPs reconstituted form lyophilized powder. (B) DLS analyses of miR-326-NPs, miR-NT-NPs, and empty NPs incubated in cell culture medium for 1 h. (C) Agarose retardation assay of miRNA loaded NPs. Lane 2: free miR-326; lane 4: free miR-NT. Lanes 3a, 5a, and 6a: miR-326-NPs, miR-NT-NPs, and empty NPs freshly prepared. Lanes 3b, 5b, and 6b: miR-326-NPs, miR-NT-NPs, and empty NPs reconstituted from lyophilized powder. Lanes 3c, 5c. and 6c: miR-326-NPs, miR-NT-NPs, and empty NPs after incubation with fully supplemented cell culture medium for 30 min. Lane 1: ssRNA ladder (1,000; 500; 300; 150; 80; and 50 bases). NPs fluorescent-labeled empty uptake in MB cellular models To test the intracellular uptake of the NPs preparations, three MB cell lines were used to model different MB subgroups.30 The use of these cell lines allowed us to account for variation in culture conditions, including suspension (CHLA), semi-suspension (D283), and adherent cells (DAOY). Lipophilic carbocyanine fluorescent dye 3,3′-dioctadecyloxacarbocyanineperchlorate (DiO)-labeled empty NPs were used to study the kinetics of uptake at different time points. In previous treatments, NPs localized inside cells, as assessed by confocal laser scanning microscopy and subsequent z stack view (Figures 3A and 3B). Flow cytometry was used to quantify the percentage of DiO positive cells, showing NPs uptake starting from 3 h with a peak at 24 h. Over 50% of cells remained DiO positive for up to 48 h. Intriguingly, DAOY cells showed higher internalization capability compared to CHLA and D283. We attributed this result to the adherent nature of these cells, which provides a larger and more stable surface for nanoparticle interaction compared to semi-adherent or suspension cells.Figure 3NPs uptake evaluation in MB cell lines(A) Immunofluorescence analysis of DiO-labeled empty NPs (green) internalized by CHLA, DAOY, and D283 cell lines after 3, 6, and 24 h of treatment. Cells nuclei were counterstained with DAPI (blue) and cell membranes were counterstained with Phalloidin (Ph, red) (representative experiments). Bar plot on percentage of DiO-positive cells (%). (B) Cytofluorimetric analysis of DiO-labeled empty NPs after 3, 6, 24, and 48 h. Percentage of DiO-positive cells (% FITC) is expressed as mean ± standard deviation (left). Representative image of gate strategy used to discriminate DiO positive cells (right). Multiple t test FDR correction, ∗∗∗∗p < 0.00001, ∗∗∗p < 0.0001, ∗∗p < 0.005, ∗p < 0.05, error bars indicate standard deviation. NPs uptake evaluation in MB cell lines (A) Immunofluorescence analysis of DiO-labeled empty NPs (green) internalized by CHLA, DAOY, and D283 cell lines after 3, 6, and 24 h of treatment. Cells nuclei were counterstained with DAPI (blue) and cell membranes were counterstained with Phalloidin (Ph, red) (representative experiments). Bar plot on percentage of DiO-positive cells (%). (B) Cytofluorimetric analysis of DiO-labeled empty NPs after 3, 6, 24, and 48 h. Percentage of DiO-positive cells (% FITC) is expressed as mean ± standard deviation (left). Representative image of gate strategy used to discriminate DiO positive cells (right). Multiple t test FDR correction, ∗∗∗∗p < 0.00001, ∗∗∗p < 0.0001, ∗∗p < 0.005, ∗p < 0.05, error bars indicate standard deviation. Overall, these findings demonstrate that NPs are efficiently internalized and retained in all cell lines tested. NPs do not affect cell viability in vitro To assess the biocompatibility of miR-loaded NPs (miR-NT or miR-326-NPs), cell viability was evaluated at different time points following treatments. Measurements were also performed by comparing cell viability with a standard transfection method, using a positive control condition (HiPerfect, QIAGEN). Two different concentrations of NPs and Hiperfect prepared with scramble miRNAs (miR-NT) or miR-326 (5 and 10 nM) were employed over a period of 6–48 h. Negative controls with miR-NT-NPs demonstrated that although the NPs are internalized by cells of all cell lines, they do not significantly affect cell viability compared to the same cells transfected with Hiperfect (Figure 4).Figure 4Cell viability after NPs treatment on MB cell lines(A) Percentage of viable CHLA, DAOY, and D283 cells at 6 , 24 , and 48 h. Cells were treated with Scramble (miR-NT) delivered by NPs or HiPerFect and miR-326 embedded NPs (5 nM). (B) Percentage of viable cells after 10 nM treatments for 6, 24, and 48 h. Multiple t test were performed on non-treated control (100% cell viability) , error bars indicate standard deviation. Cell viability after NPs treatment on MB cell lines (A) Percentage of viable CHLA, DAOY, and D283 cells at 6 , 24 , and 48 h. Cells were treated with Scramble (miR-NT) delivered by NPs or HiPerFect and miR-326 embedded NPs (5 nM). (B) Percentage of viable cells after 10 nM treatments for 6, 24, and 48 h. Multiple t test were performed on non-treated control (100% cell viability) , error bars indicate standard deviation. Multiple t test revealed no significant differences on cell viability between the NPs and HiPerFect-treated cells at both 5 and 10 nM. Therefore, for all subsequent experiments, the lowest concentration (5 nM) of NPs was selected to reduce miRNA-nonspecific effect due to non-physiological overload. MiR-326-NPs efficiently increase miR-326 levels in MB cell lines Protection of miRNA from extracellular milieu degradation and therefore successful delivery of miRNA mimics into cells, still represent one of the major challenges for NPs platforms. To evaluate the efficacy of miR-326-NPs as a delivery system, the miR-326 expression level in all MB cell lines was evaluated after 24- and 48-h treatment. As shown in Figure 5A, delivery of miR-326 mimic via NPs resulted in a significant increase in miR-326 expression in all tested cell lines, both at the peak of NPs uptake (24 h, shown in Figure 3) and up to 48 h after treatment meaning that NPs are more efficient as miRNA delivery tool than the standard method used as positive control (Figure 5B).Figure 5miR-326 expression levels after delivery with miR-326-NPs and Hyperfect(A) Relative expression of miR-326 after 24 and 48 h treatment with miR-326-NPs in all cell lines. (B) Comparison of transfection methods. Relative expression of miR-326 after 24 h of miR-326 delivery with NPs or HiPerFect (Hip) in all cell lines. Expression was normalized to U6 snRNA and compared to controls (miR-NT embedded NPs/HiPerFect). Unpaired t test (parametric), ∗p < 0.05, ∗∗p < 0.005, error bars indicate standard deviation. miR-326 expression levels after delivery with miR-326-NPs and Hyperfect (A) Relative expression of miR-326 after 24 and 48 h treatment with miR-326-NPs in all cell lines. (B) Comparison of transfection methods. Relative expression of miR-326 after 24 h of miR-326 delivery with NPs or HiPerFect (Hip) in all cell lines. Expression was normalized to U6 snRNA and compared to controls (miR-NT embedded NPs/HiPerFect). Unpaired t test (parametric), ∗p < 0.05, ∗∗p < 0.005, error bars indicate standard deviation. MiR-326-NPs induce neuronal differentiation, apoptosis, and impair stemness MiR-326 has been widely recognized as a tumor suppressor in MB, particularly due to its role in regulating the stemness potential of MB cancer cells.9 Cancer stem cells represent the most aggressive and therapy-resistant cell population within the tumor. Given this, we investigated the biological effects of delivering miR-326 via NPs, focusing on its impact on differentiation, stemness, and apoptotic markers. To verify the transition to a more differentiated phenotype, we assessed mRNA expression of the differentiation neural marker tubulin-beta 3 (Tubb3) and stemness markers, homeobox Nanog and Sox2.10 A significant decrease in Nanog mRNA was observed across all MB cell lines after miR326-NP delivery, accompanied by a decreasing trend in Sox2 expression. Differentiation, assessed by evaluating Tubb 3 mRNA expression was found significantly upregulated in D283, after 24 h treatment with miR326-NPs and protein levels showed an increasing trend, in all tested cell lines (Figures 6A and 6B). Moreover, after miR-326-NPs treatment an increase of PARP and the cleaved form of PARP (c-PARP), an indicator of apoptosis activation was observed in all tested cell lines. Although c-PARP displayed a relatively faint signal in D283 and DAOY cells, this increase was consistently detected across replicates, supporting the pro-apoptotic role of miR-326 delivery (Figure 6B). Thus, restoration of miR-326 levels via NPs drives cell fate toward a more differentiated phenotype while promoting apoptosis, recapitulating the results obtained in previous studies.Figure 6Evaluation of the effect of miR-326 (miR-326-NPs) or non-targeting miRNA (miR-NT-NPs) after 24 h incubation on differentiation and stemness markers(A) mRNA expression (RT-qPCR) of TUBB3, Nanog and SOX2 in the different cell lines treated with miR-326-NPs compared to miR-NT-NPs (below dotted line). (B) Western blot analysis of tubulin beta III (TUBB3), PARP, and cleaved PARP (C-PARP). Representative experiment (n = 3). Histograms show protein levels normalized on GAPDH expression. Parametric unpaired t test, ∗p < 0.05, ∗∗p < 0.005, ∗∗∗p < 0.0001, error bars indicate standard deviation. Evaluation of the effect of miR-326 (miR-326-NPs) or non-targeting miRNA (miR-NT-NPs) after 24 h incubation on differentiation and stemness markers (A) mRNA expression (RT-qPCR) of TUBB3, Nanog and SOX2 in the different cell lines treated with miR-326-NPs compared to miR-NT-NPs (below dotted line). (B) Western blot analysis of tubulin beta III (TUBB3), PARP, and cleaved PARP (C-PARP). Representative experiment (n = 3). Histograms show protein levels normalized on GAPDH expression. Parametric unpaired t test, ∗p < 0.05, ∗∗p < 0.005, ∗∗∗p < 0.0001, error bars indicate standard deviation. MiR-NT-NPs did not induce toxicity in mice Having observed promising in vitro effects with the miR-326-NPs system and to evaluate the potential in vivo toxicity of the nanoparticles, we injected the miR-NT-NPs, in a group of C57BL/J6 mice for 5 days, focusing on the key organs involved in drug filtration (Figure 7A). The treatment administered during the study did not alter the mice’s eating, drinking, or sleeping patterns, and their growth remained unaffected. Body weight gain in the treated mice was comparable to controls. Given that nanoparticles are known to accumulate primarily in the liver,31,32 we assessed liver function (ALT and AST) and kidney function (urea and creatinine levels). While we observed an increase in ALT levels (Figure 7B), this change did not show a clear dose-dependent relationship. Importantly, serum levels of AST, urea, and creatinine showed no significant alterations (Figure 7B), indicating the absence of significant liver or kidney toxicity. Furthermore, histopathological analysis revealed no abnormal gross lesions in the treated groups (representative images in Figure 7C). Overall, these findings suggest that the nanoparticle formulation carrying miR-NT does not cause any apparent in vivo toxicity and provides a basis for further investigation in more complex models.Figure 7In vivo evaluation of miR-NT-NPs(A and B) (A) In vivo miR-NT-NPs treatment workflow design. Created in https://BioRender.com (B) Hematological tests of aspartate (AST) and alanine aminotransferase (ALT), creatinine, and urea measured in plasma. (C) Histopathological evaluation of the organs kidney, liver and spleen in control (Ctrl) and NPs-treated (0.5 mg, 3 mg) mice (representative figure). Parametric unpaired t test, ∗p < 0.05, ∗∗p < 0.005, error bars indicate standard deviation. In vivo evaluation of miR-NT-NPs (A and B) (A) In vivo miR-NT-NPs treatment workflow design. Created in https://BioRender.com (B) Hematological tests of aspartate (AST) and alanine aminotransferase (ALT), creatinine, and urea measured in plasma. (C) Histopathological evaluation of the organs kidney, liver and spleen in control (Ctrl) and NPs-treated (0.5 mg, 3 mg) mice (representative figure). Parametric unpaired t test, ∗p < 0.05, ∗∗p < 0.005, error bars indicate standard deviation. MiR-326-NPs antitumoral effect in vivo In order to recapitulate the ability of miR-326-NPs to reduce the progression of MB cells in vivo, we set up a pilot experimental protocol using Balb/c nude mice inoculated with D283 cells and sorted into two groups of seven mice receiving I.V. biweekly for 4 weeks, miR-NT-NPs and miR-326-NPs, respectively. For the last administrations, fluorescent NPs were used to track the distribution of NPs. We confirmed also in this animal model the safety profile of NPs, in fact, the serial I.V. administration of NPs did not determine any behavioral change or reduced body weight gain in mice, and the experiment was ended only due to the dimension of tumor masses. Results show a reduction in the dimensions of subcutaneous tumor in the miR-326-NPs group with respect to the miR-NT-NPs group with increasing differences between mean volumes on days 18, 22, and 25 after the cells inoculation (55.6, 111.4, and 285.3 mm3, respectively) (Figure 8A). On day 28 after inoculation, the difference between mean volumes went down to 185.0 mm3, although the results are not supported by statistical significance. Fluorescent NPs were administered and their distribution was analyzed after 24 h, revealing a permanence of fluorescent signal well distributed in the mouse body (Figure 8B). This distribution pattern likely reflects accumulation in organs with high vascularization and phagocytic activity (e.g., liver, spleen), which is commonly observed for systemically administered NPs. Importantly, despite this preferential abdominal accumulation, the NPs retained detectable presence in other several tissues, consistent with the potential for broader biodistribution and functional delivery.Figure 8In vivo antitumoral efficacy of miR-326-NPs(A) Growth curves of D283 cell line xenografts expressed as mean tumor mass volumes (mm3) from the control group (n = 7, miR-NT-NPs) and from the targeted group (n = 7, miR-326-NPs) on days 18, 22, 25 and 28 after the tumor cells grafting. (B) Mice were I.V. injected with ATTO700 dye-conjugated NPs, and after 24 h imaging analysis was performed. The localization of tumor xenograft was indicated by the oval striped line. The color scale represents a visual indicator for radiant efficiency (p/sec/cm2/sr)/mW/cm2)) with a min value = 2.99 × 107 and a max value = 1.07 × 109, error bars indicate standard deviation. In vivo antitumoral efficacy of miR-326-NPs (A) Growth curves of D283 cell line xenografts expressed as mean tumor mass volumes (mm3) from the control group (n = 7, miR-NT-NPs) and from the targeted group (n = 7, miR-326-NPs) on days 18, 22, 25 and 28 after the tumor cells grafting. (B) Mice were I.V. injected with ATTO700 dye-conjugated NPs, and after 24 h imaging analysis was performed. The localization of tumor xenograft was indicated by the oval striped line. The color scale represents a visual indicator for radiant efficiency (p/sec/cm2/sr)/mW/cm2)) with a min value = 2.99 × 107 and a max value = 1.07 × 109, error bars indicate standard deviation. Discussion Despite current therapeutic schedules that consider the molecular features of MB subtypes there are no significant improvements in survival rates or alleviation of long-term treatment sequelae.11 This highlights the need for innovative therapeutic approaches. In this context, the clinical application of nucleic acids stands out, due to their cost-effectiveness and ease of synthesis. MiRNAs offer a unique advantage by simultaneously disrupting multiple critical tumorigenic pathways. However, translating miRNA therapies into clinical practice faces challenges, including degradation susceptibility and stability issues.33 Our study explored the role of miRNA delivery via PCL/chitosan-based NPs in overcoming these challenges. Chitosan, a naturally abundant and biodegradable polymer, is well suited for nucleic acid delivery due to its cationic nature, which facilitates encapsulation of negatively charged miRNAs. Coating NPs with PEI, a highly cationic polymer with a high density of amine groups, significantly increases the positive surface charge of the nanoparticles. This leads to improved colloidal stability through increased electrostatic repulsion between individual nanoparticles, preventing aggregation and increasing cellular uptake. Enhancements such as the addition of PCL, further improve encapsulation efficiency and release kinetics.34 Furthermore, PCL/chitosan NPs can be surface-modified to enhance stability, target efficiency, and controlled release, positioning them as a promising tool for delivering miRNA-based therapies.35,36 The size of chitosan NPs is a critical parameter33 that depends on the cell type and surface properties. In our case, NPs tested with a size ranging from 400 to 550 nm, did not seem to influence cytotoxicity in both in vitro and in vivo systems. We hypothesized that this is due to the fact that NPs >200 nm do not interfere with cellular organelles or with intracellular enzymatic pathways. In this regard. it must be considered that chitosan is a natural polymer that breaks down safely within the body, eliminating concerns about long-term accumulation. In addition, PCL polymer after degradation can be excreted through the metabolism of the tricarboxylic acid cycle. Its biocompatibility further enhances its appeal as a carrier for delicate biomolecules, ensuring minimal adverse reactions upon administration. This combination of biodegradability and biocompatibility makes chitosan-based NPs promising for targeted and personalized drug delivery systems aimed at tackling the unique characteristics of individual tumors.33,37 We selected miR-326 as a prototype in this study because of its known tumor-suppressing role and its epigenetic re-activation during neural differentiation. Our previous findings9 demonstrated that low miR-326 levels in MB correlated with pro-tumorigenic pathways such as Hedgehog (Hh) signaling. By using PCL/chitosan nanoparticles to deliver miR-326, we successfully restored its expression in MB cell lines, confirming previously reported biological effects. Following the optimization of our NP formulation, we demonstrated the safety of PCL/chitosan-miRNA NPs across various MB cell lines representing group 4, SHH, and group 3 subtypes. Notably, PCL/chitosan NPs exhibited reduced toxicity compared to traditional transfection methods such as HiPerFect, even at higher concentrations and with prolonged exposure. These NPs were rapidly internalized by MB cells, leading to a peak in miR-326 expression at 24 h that persisted for at least 48 h. Importantly, miR-326 expression levels were significantly higher in NP-treated cells compared to those treated with traditional methods, suggesting improved stability and cellular retention conferred by the nanocarrier. Functionally, NP-delivered miR-326 induced neuronal differentiation, decreased the expression of stemness markers (Nanog and Sox2), and promoted apoptosis. These biological effects of miR-326-NPs align with findings from previous studies4,9 where miR-326 was delivered via traditional transfection methods in MB models. The in vivo studies demonstrating the low toxicity of NPs in the most responsive organs, warrant further investigation into their potential applications. Our in vivo findings suggest the use of our PCL/chitosan-based platform for delivering therapeutic molecules directly to tumor sites. This platform offers a large surface area, ideal for binding nucleic acids and other biomolecules that could enhance the targeting of tumor cells. Based on the results obtained in vitro, we have performed pilot investigations in a mouse model to preliminarily evaluate safety and biological activity in vitro. The safety profile was supported by behavioral and biochemical assessment in both immunocompetent and immunodeficient mouse strains. Then, an in vivo tumor model was established by D283 xenograft, to investigate the potential tumor growth impairment of the vehiculation of miR-326-NPs. Although results demonstrated an appreciable reduction of mean tumor volume in the miR-326-targeted group with respect to the control group, the high variability in tumor response did not permit reaching a statistically significant value. This moderate effect observed can be explained by the biological function of miR-326, which functions primarily as a tumor suppressor by modulating oncogenic pathways and promoting differentiation programs rather than inducing strong cytotoxicity. This suggests that while miR-326-NPs reach the tumor site, their biological impact may be more subtle or time dependent than traditional cytotoxic drugs. Consequently, the observed trend toward tumor reduction warrants further investigation using larger cohorts or extended observation periods to overcome the inherent biological variability. Notably, a moderate effect as monotherapy does not preclude therapeutic relevance. On the contrary, the differentiation-inducing and pathway-modulating properties of miR-326 may increase tumor cell susceptibility to additional treatments, supporting the rationale for combination strategies with conventional chemotherapeutic agents or targeted therapies as a promising direction for future investigation. Additionally, comprehensive transcriptomic profiling in future studies will enable a deeper understanding of the molecular programs modulated by miR-326 re-expression and further elucidate the downstream pathways influenced by NP-mediated delivery. We recognize that without a comprehensive characterization of the miR-326 targetome in this specific NP-delivered context, the precise molecular drivers of the observed phenotypic changes remain to be fully elucidated. Future studies incorporating RNA-seq or proteomic analysis will be essential to map the gene networks regulated by our delivery platform. The preliminary analysis of NP distribution after I.V, injection, tracked by fluorescent dye detection in the whole mouse, revealed the permanence of a widespread signal, well evident 24 h after the administration, suggesting a good stability of NPs, and the ability to also reach the tumor xenograft. We acknowledge, as a limitation of this study, the lack of data on the distribution in the central nervous system (CNS), which will require dedicated investigation. The versatility of this platform and the lack of toxicity extend beyond miRNAs, allowing for co-delivery of chemotherapeutic agents or other bioactive molecules, enabling a multi-faceted delivery approach with controlled release. Nevertheless, our findings suggest that PCL/chitosan miRNA-NPs support the ability to reprogram MB cancer cells toward a more differentiated state. This phenotypic shift could enhance their responsiveness to existing therapeutic strategies or render them susceptible to new treatment agents. Our study provides a proof of concept for the use of PCL/chitosan-based nanoparticles as a promising experimental scaffold for future therapeutic development. The persistent challenges in MB treatment, characterized by stagnating survival rates and severe long-term sequelae despite advancements in molecular subtyping, underscore the urgent need for novel approaches. Our research lays the foundation for addressing this need by leveraging the cost-effectiveness and ease of synthesis of nucleic acids, specifically miRNAs, which offer a unique advantage by simultaneously disrupting multiple tumorigenic pathways. The favorable safety profile, combined with potential for systemic delivery in vivo and the versatility to co-deliver chemotherapeutic agents or other bioactive molecules, positions our PCL/chitosan miRNA-NPs as a versatile candidate therapeutic tool. This approach offers significant opportunities to enhance cancer cells responsiveness to existing therapies or render them susceptible to new treatment modalities, ultimately opening new avenues for improved treatment strategies for this devastating pediatric cancer. Materials and methods Chitosan purification Chitosan (Merck, cat. no. 448877; 190–310 kDa; 75%–85% deacetylated) was treated with 20 mL of 1 M NaOH at 5 mg/mL and stirred at 70°C for 12 h. After cooling, the undissolved chitosan was filtered, washed three times with ultrapure water, and then dissolved in 100 mL of 0.2 M acetic acid with vigorous stirring. This solution was filtered, and pH was adjusted to 8 with NaOH to precipitate the chitosan. The precipitate was collected by filtration, washed three times with ultrapure water, and freeze dried. For the in vivo study, chitosan was fluorescently labeled using Atto 700 NHS-ester (Merck, cat. no. 16986). Briefly, purified chitosan was dissolved in 0.2 M hydrochloric acid to a concentration of 1 mg/mL under stirring for 12 h. Afterward, the pH was carefully adjusted to 6.8 with 1 M NaOH under vigorous stirring to achieve a fine particulate due to the pH shift. The coupling reaction was initiated by adding Atto 700 NHS-ester at 1 mg/mL in DMSO. The reaction proceeded for 12 h with gentle stirring. Then, the pH was increased to 8.0 with the proper amount of NaOH 1 M, the precipitate was collected and then washed three times with ultrapure water, and freeze dried. NPs formulation Purified chitosan was dissolved in 0.1 M MES buffer (pH 5.6) at a concentration of 5 mg/mL. The pH was checked and adjusted to 5.6 when needed. PCL (Merck, cat. no. 440752, average MW 14 kDa) was dissolved in anhydrous THF at 10 mg/mL, either alone or with the lipophilic dye DiO, 5 μM, (ChemCruz, cat. no. SC-214168) to obtain fluorescently labeled NPs. PVA (Merck, cat. no. 363170; MW 13–23 kDa) was dissolved in water at 1 mg/mL, as was PEI (Merck, cat. no. 408727, average MW 25 kDa). All solutions were filtered with 0.22 μm filters (polyethersulfone for aqueous solutions and PTFE for PCL solutions). Fluorescent NPs for in vivo studies were prepared as described previously by employing Atto 700-labeled chitosan. miRNA encapsulation in NPs In sterile vials under a biological hood, either miRNA scramble (miR-NT) or miR-326 was used. The miR-NT is a synthetic, validated random sequence with no known effect on miRNA expression. MiRNAs (Eurofins Genomic) were dissolved in nuclease-free water to 5 μM. At the time of assembling nanomaterials, 10 μL of miRNA solution was added to 40 μL of chitosan (5 mg/mL in MES buffer, pH 5.6) under stirring and gently mixed for 30 min at room temperature. These cationic-anionic complexes were then added to 225 μL of PVA solution (1 mg/mL) under vigorous stirring, followed by 5 μL of PCL and 2 μL of PEI (1 mg/mL) after 30 s. Unloaded NPs were prepared without miRNA and referred to as “empty NPs”. For in vivo experiments, synthetic C. elegans cel-miR-39 (Thermo Fisher Scientific, TaqMan MicroRNA Assays, cat. no. 4427975, ID: 000200) was used as negative control since it is not present in mammalian cells and is commonly used as a spike-in.38 NPs processing NPs were left in a sterile biological hood to allow THF evaporation for 2 h. Then, they were centrifuged at 16,000 g for 20 min at 4°C. Supernatants were discarded, and pellets were resuspended in 250 μL of PBS (pH 6.2) supplemented with mannose 1% w/v (Merck, cat. no. 63579) and freeze dried. DLS and ELS A Zetasizer Nano-ZS90 (Malvern Instruments, UK) with a 633 nm laser was used for DLS and electrophoretic light scattering (ELS) analyses to determine the mean hydrodynamic diameter and zeta potential (ζ) of the NPs. Measurements were performed on freshly synthesized NPs and on lyophilized NPs after reconstitution in ultrapure water (200 μL) to assess post-lyophilization stability. To evaluate stability in cell-culture medium, NPs were suspended in DMEM with 10% heat-inactivated FBS and antibiotics, incubated at 37°C (5% CO2) for 1 h and then analyzed by DLS for hydrodynamic diameter. SEM NPs morphological analysis was carried out using a Sigma 300VP microscope (Zeiss, Germany). NP suspensions were applied to SiO2 wafers, dried overnight, and then imaged. miRNA loading efficiency To determine miRNA loading efficiency, NPs were centrifuged at 50,000 g for 30 min at 4°C (Optima MAX-XP, Beckman Coulter, USA). Supernatants were collected and analyzed by DLS to confirm absence of NPs. Free miRNA in supernatants was quantified spectrophotometrically using a Nanodrop ND-ONEC-W (Thermo Fisher Scientific, USA) to calculate encapsulation efficiency. This was done for both freshly prepared and reconstituted NPs. Gel retardation assay for loading efficiency and payload protection A gel retardation assay was used to assess miRNA loading efficiency. Freshly prepared or reconstituted NPs (10 μL) were mixed with loading buffer and loaded onto a 2% agarose gel with 0.5 μg/mL ethidium bromide in TAE buffer. The gel was run at 90 V for 45 min miR-326 and miR-NT controls, and a low range ssRNA ladder (New England BioLabs) were included. Gels were imaged using a UV transilluminator (Chemidoc, Bio-Rad). The gel retardation assay was also used to evaluate the NPs’ ability to protect their miRNA payload from serum components. NPs were incubated in DMEM with 10% FBS and antibiotics at 37°C (5% CO2) for 30 min and then analyzed by gel electrophoresis to verify the integrity of the NPs and their payload. Cell culture The MB cell lines CHLA (group 3–4), DAOY (SHH Group), and D283 (group 3) were used in this study. CHLA-01-MED, D283 Med, and DAOY are MB cell models with distinct phenotypes. CHLA cells, derived from a primary tumor, grow as poorly differentiated neurospheres. D283 cells, isolated from metastatic lesions, display an aggressive phenotype with mixed suspension/adhesion growth. DAOY cells, originating from a primary desmoplastic medulloblastoma, exhibit adherent growth. All cell lines were purchased and authenticated by ATCC (Manassas, VA, USA) and cultured at 37°C in a humidified 5% CO2 atmosphere. CHLA cells were cultured in DMEM-F12 (Gibco, cat. no. 11320033) supplemented with 1% penicillin-streptomycin (Gibco, cat. no. 15140122), 2% serum-free B-27 Supplement (50×, Gibco, cat. no. 17504044), and freshly added EGF and FGF (20 ng/mL final concentration). D283 and DAOY cells were cultured in MEM (Gibco, cat. no. 11095080) supplemented with 1% L-glutamine (200 mM, Gibco, cat. no. 25030081), 1% sodium pyruvate (100 mM, Gibco, cat. no. 11360070), 1% penicillin-streptomycin (Gibco, cat. no. 15140122), 1% MEM non-essential amino acids solution (100×, Gibco, cat. no. 11140050), and 10% FBS (Gibco). NPs treatment and cells transfection CHLA, DAOY, and D283 cells were treated at 60%–70% confluence. Lyophilized NPs were resuspended in 200 μL of ultrapure water to obtain a concentration of 250 nM. Cells were then treated with NPs at final concentration of 5 or 10 nM. For transient transfection, 5 nM of miRNA mimic was used (Dharmacon). The mimic negative controls (cat. ID CN-001000-01-05), also referred as Scramble, and the miRIDIAN mimic hsa-miR-326 (cat. ID: C-300699-03-0005) were transfected using Opti-MEM (cat. no. 31085-047, Gibco) and HiPerFect Transfection Reagent (cat no. 301705, QIAGEN). Cytofluorimetric assay and immunofluorescence Empty NPs fluorescently labeled with DiO (DiO-labeled empty NPs) were used as a control to assess cellular uptake in both cytofluorometric and immunofluorescent analyses. Cytofluorometric analysis After NP treatment, cells were dissociated using Cell Dissociation Solution Non-enzymatic 1× (C5914, Sigma-Aldrich) and counted with trypan blue. 100,000 cells were resuspended in 100 μL of PBS supplemented with 1% FBS, and the percentage of FITC-positive cells was evaluated using a CytoFLEX S Flow Cytometer (Beckman Coulter). Excitation was performed with a 488-nm laser, and the green fluorescence channel (FL1) was used for detection. For each sample, 10,000 events were measured, and data were analyzed using CytExpert Software v.2.4 (Beckman Coulter). Immunofluorescent analysis For immunofluorescence detection, cells cultivated in semi-suspension and adhesion were treated with NPs and cultured in μ-Slide 8 Well chamber slides coated with Poly-L-Lysine (ibidi, cat. no.:80824). Cells growing in suspension were treated in MW 24 dishes. After treatment, 1,000 cells (for suspension cultures) were spotted onto glass slides using a cytospin centrifuge at 1,000 rpm for 5 min. All cells were then washed with DPBS with Ca2+ and Mg2+ (Sigma-Aldrich, cat. no.:D1283), fixed in 4% paraformaldehyde for 10 min at RT, and washed three times with PBS. Cells were stained with Phalloidin 1:250 (Sigma-Aldrich, P2141) for 50 min at 4°C. Nuclei were counterstained with DAPI (1:1,000), and coverslips were mounted using fluorescence mounting medium (Prolong Gold, Thermo Fisher Scientific, MA, USA). Images were acquired using an FV1200 MPE laser scanning confocal microscope (Olympus) and processed using Imaris 8.1 software (Oxford Instruments, https://imaris.oxinst.com). Cells with colocalization of DAPI and DiO were considered positive; uptake efficiency was calculated as percentage of positive cells on total cell count, using multiple images from each condition (at least 3). Cytotoxicity assay Cell viability was evaluated by using CellTiter 96 Aqueous One Solution Cell Proliferation Assay (MTS) (Promega, Madison, G7570, USA). Reagents were mixed at a 1:1 ratio with supernatant from the treatment plate. The mixture was incubated for 10 min at RT on a shaker, and the signal was detected by luminescence measurement using a luminometer GloMax Microplate Reader& (Promega). RT-qPCR The total RNA was isolated using TRIzol Reagent (Invitrogen, cat. no.: 15596026) as previously described. For gene expression, cDNA was synthesized using TaqMan Reverse Transcription Reagents (cat. no. N8080234, Applied Biosystems, Thermo Fisher Scientific). The following TaqMan assays for mRNA expression (cat. no. 4331182) were used: Nanog (ID: Hs04260366_g1), Sox2 (ID: Hs01053049_s1), Tubb3 (ID: Hs00801390_s1), and Actb (ID: Hs01060665_g1). For miRNA expression, TaqMan MicroRNA Reverse Transcription Kit (cat. no. 4366596, Applied Biosystems, Thermo Fisher Scientific) was used to generate miRNA-specific cDNA. The following TaqMan MicroRNA Assays (cat. no. 4427975) were used: hsa-miR-326 (ID: 000542), cel-miR-39-3p (ID: 000200), and U6 snRNA (ID: 001973). Quantitative reverse-transcription PCR (RT-PCR) analysis was performed with the ViiA7 Sequence Detection System (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instructions. Each amplification reaction was performed in triplicate, miRNA values were normalized to U6 and mRNA levels were normalized to Actin expression. Western blot Cells were lysed in RIPA buffer and protease inhibitors. 30 μg of protein lysates were subjected to electrophoresis on 4%–12% NuPage Bis-Tris gels (Life Technologies, cat. no.1402BOX) following the manufacturers’ instructions. The gels were transferred to nitrocellulose membranes in Tris-glycine buffer. The membranes were blocked in PBS-0.1% Tween 20 solution containing 5% milk. Membranes were incubated overnight with the following specific antibodies: β-Tubulin III (Sigma-Aldrich, cat. no. TU-20), Cleaved PARP (Cell signaling, cat. no. 9542). GAPDH (Cell signaling, cat. no. #5174). Secondary antibodies (Santa Cruz Biotechnology, Dallas, TX, USA) were used in combination with LiteAblot TURBO (Euroclone). Images were acquired with ChemiDocTM MP Imaging System (Bio-Rad), and quantification was performed using ImageJ version 1.53 as described in the ImageJ documentation. Animals and miR-NT-NPs treatment C57BL/6 J mice (8 weeks old) were purchased from Charles River (Wilmington, MA). For in vivo assessment of NPs toxicity, mice were randomly assigned into three groups (n = 3 mice/group): group 1: controls, group 2:0.5 mg NPs administered at day 0 and at day 3, and group 3: 3 mg administered at day 0 and at day 3. Injections were administered into the tail vein; dosages were chosen according to previous studies.39,40,41 Mice were sacrificed 48 h after the last treatment. Tissues were dissected on ice: some were snap frozen for RNA (n = 3/group) and others were fixed in 4% PFA for histological evaluations (n = 3/group). The mice were sacrificed at the designated time periods for serum biochemical and histopathological analysis. Serum biochemical analysis Serum was obtained from total blood by centrifugation at 2,000 g for 10 min at RT, followed by a second centrifugation at 2,500 g for 10 min at 4°C. Liver and kidneys functions were assessed by measuring the levels of AST and ALT, urea, and creatinine. The kits were used following the manufacturer’s instructions: QuantiChromTM Urea Assay Kit, QuantiChromTM Creatinine Assay Kit, EnzyChromTM Aspartate Transaminase Assay Kit and EnzyChromTM Alanine Transaminase Assay Kit (BioAssay Systems). Histopathological analysis Livers, kidneys, and spleens were dissected and fixed in 4% PFA solution in PBS overnight at 4°C. Tissues were then embedded in paraffin and sectioned at 5 μm. Hematoxylin and Eosin staining (Sigma-Aldrich, cat. no. GHS232 and cat. no. HT110116) was performed following standard procedures. Dehydration was carried out with increasing ethanol concentrations. Sections were mounted with CV Ultra Mounting Media (Leica Biosystems) and images were acquired using a PreciPoint M8 Microscope and Slide Scanner (Nikon). MB mouse model 5-to 6-week-old BALB/c nude female mice (BALB/BALB/cOlaHsd-Foxn1nu; Envigo RMS srl) were subcutaneously injected with 1.0 × 106 human D283 Med cells in 200 L of sterile 1× PBS and Matrigel (1:1; Corning). After 6 days, mice were randomized into treatment groups and then treated intravenously with miR-NT-NPs (vehicle) or miR-326-NPs twice a week with the same dose until the experimental endpoint (7 mice each group, all female, for 28 days). Tumor growth was measured using a caliper, and tumor volumes were calculated using the formula: volume = (A × B2)/2 (where A is the larger diameter, and B is the smaller diameter of the tumor). For in vivo imaging analysis, mice intravenously injected via a lateral tail vein with miR-NT-NPs-ATTO700, were anesthetized with isoflurane/oxygen mix, and fluorescence was analyzed at 24 h using IVIS Spectrum (In Vivo Imaging System). Imaging data were analyzed using Living Image software (Revvity). Statistical analysis Statistical analyses were performed using t test on GraphPad Prism software Version 8.0 (GraphPad Software, San Diego, California US). Differences were considered significant at a confidence interval (CI) of 95% (p ≤ 0.05) and 99% (p ≤ 0.01). Results are expressed as means ± SD. For the ELISA assay, one-way ANOVA was used with data presented as ± SD. For the MTS assay, multiple t tests were performed using the Two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli, with Q = 1%. Each row was analyzed individually without assuming a consistent SD. For in vivo MB mouse model, statistical analysis of results was done using the two-way ANOVA method. Data and code availability The data supporting the findings of this study are available from the corresponding author upon reasonable request. Acknowledgments E.S. and G.M. contributed equally to this work and are listed as co-first authors. E.F. and I.E.P., contributed equally to this work and are listed as co-last authors. E.F. and G.C. are listed as co-corresponding authors. G.M., G.G., and I.E.P., acknowledge “Tecnopolo per la medicina di precisione” (TecnoMed Puglia) - Regione Puglia: DGR n. 2117 del 21/11/2018; Hub Life Science – Terapia Avanzata (LSH-TA); PNC-E3-2022-23683269 within 10.13039/100006826EU funding within the PNC Italian Health Ministry; Project “Biotecnologia, Bioinformatica e Sviluppo farmaceutico” per la creazione di un hub delle scienze della vita, piano operativo salute (Piano di Sviluppo e Coesione del Ministero della Salute 2014-2020), Traiettoria 4, azione 4.1, cod. T4-AN-01. G.M., G.G., I.E.P., E.S., E.F., Z.B.M, A.P., A.A., D.V., D.V, F.Z., G.C., EU funding within the 10.13039/501100021856MUR PNRR “National Center for Gene Therapy and Drugs based on RNA Technology” (project no. CN00000041 CN3 RNA). A.A., D.V., D.V, F.Z., Bando a Cascata CODBAN_000532, n. 0002 23/05/2024. G.C. (Giuseppe Ciccarella) Bando a Cascata CODBAN_000532, n. 0008 23/05/2024. All animals were housed in ventilated cages (no more than 4 mice/cage) and in a controlled environment. All animal studies were conducted in accordance with all relevant ethical regulations for animal testing and research including the Italian code for the care and use of animals for scientific purposes. Experiments were approved by the 10.13039/501100003196Italian Ministry of Health (Authorization n. 152/2021-PR;897/2025-PR). Author contributions E.F., G. Gigli, I.E.P., E.S., G. Catanzaro, G.M, and A.M., conceptualization; E.S., Z.S., G. Gugliuzza, A.A., G.M., I.E.P., and T.R., performed experiments; M.M., M.I., A.A. A.P., G.Catanzaro, N.F., D.V., and F.Z., performed in vivo experiments; E.S., G.M., I.E.P., and Z.S., data analysis; G. Gigli, E.F., and G. Ciccarella, funding acquisition; G. Gigli, E.F., G. Catanzaro, S.M., and A.Z., supervision; E.S., A.P., G. Catanzaro, Z.M.B., Z.S., E.F., G. Gigli, G.M., I.E.P, A.C. E.D, A.M., M.A.V., and A.A., wrote and revised the manuscript. Declaration of interests The authors declare no competing interests.