Anti-miR-195-engineered mesenchymal stem cells promote migration, vascularization, and ECM protein expression in vivo Authors: Sara Reis Moura, Mafalda Montenegro Cortez, Isabel Brandão, Inês Alencastre, Susana Gomes Santos, Carla Cunha, Maria Inês Almeida Published: 2026 ABSTRACT AbstractAdvanced therapy medicinal products (ATMPs) involving mesenchymal stromal/stem cells (MSCs) have emerged as a promising therapeutic approach for a wide spectrum of diseases, including several musculoskeletal disorders and bone fractures. Despite their potential, clinical translation has often led to inconsistent outcomes and the therapeutic efficacy of unmodified MSCs remains suboptimal. To overcome these challenges, strategies to enhance the functional properties of MSCs are needed. MicroRNAs (miRNAs) are master regulators of gene expression and have emerged as powerful tools for engineering MSCs to enhance their pro-regenerative capacity and potentially improve clinical outcomes. High-throughput RNA sequencing was performed to characterize transcriptional changes induced by anti-miR-195 modulation. Anti-miR-195-modulated MSCs were seeded onto collagen/hydroxyapatite scaffolds and their potential was evaluated in vivo using a heterotopic bone formation mouse model. Histological and molecular analyses were performed at 14 and 28 days post-implantation to assess cell migration, osteogenic marker expression, collagen deposition and vascularization. Anti-miR-195-modulated MSCs displayed transcriptional profiles associated with enhanced migration, invasion, and cell cycle progression, together with reduced senescence, cell death and apoptosis. Expression of the bone mineral density-associated transcript IBSP was significantly increased. In vivo, scaffolds seeded with anti-miR-195-treated MSCs showed increased cell migration at day 14 post-implantation. At day 28 post-implantation, enhanced collagen deposition and increased expression of osteogenic markers, including RUNX2 and SPP1, were observed. In addition, the scaffold area covered by blood vessels and the number of larger vessels were significantly higher in anti-miR-195-MSC scaffolds compared with controls. These findings demonstrate that anti-miR-195 promotes cell migration and vascularization in vivo, supporting miRNA-based engineering as a promising strategy to improve the efficacy of MSC-based ATMPs for musculoskeletal diseases and bone defects.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13287-026-05082-6.Keywords: Non-coding RNAs, miR-195-5p, Skeletal diseases, Bone, Molecular engineering, Cell delivery, Advanced therapy medicinal products Introduction The intrinsic regenerative capacity of mesenchymal stem/stromal cells (MSCs) arises from their ability to differentiate into multiple mesenchymal lineages, including osteoblasts, chondrocytes, adipocytes and their immunomodulatory secretome, which supports tissue regeneration and regulates local immune responses [1]. In skeletal diseases and bone fractures, MSCs contribute directly to the regenerative process by forming new bone tissue and indirectly by recruiting cells, including endothelial cells, via paracrine signalling [2]. Despite promising preclinical and clinical results, MSC-based therapies for skeletal diseases or for regeneration applications often yield inconsistent outcomes, highlighting the need to bolster MSC functionality. In this context, molecular engineering approaches have emerged as tools to fine tune MSC phenotypes and overcome these limitations. Among these, modulation of bone morphogenetic protein (BMP) signaling is a prominent strategy, but dose-related safety concerns and delivery challenges have been raised [3, 4]. We propose the use of non-protein-coding transcripts (non-coding RNAs; ncRNAs), as an alternative to protein-based approaches for modulating MSCs toward a pro-regenerative phenotype. Specifically, microRNAs (miRNAs) are versatile molecules due to their ability to simultaneously regulate multiple mechanisms [5–7]. Our previous study identified miR-195-5p as an anti-proliferative and anti-osteogenic miRNA that also inhibits angiogenesis through a paracrine effect [8]. Here, we evaluate the use of miR-195-5p inhibitors as a strategy to engineer MSCs in vivo. We combined anti-miR-195-modified MSCs with collagen/hydroxyapatite scaffolds and assessed their performance in a heterotopic bone formation model. Concomitantly, we have also analyzed the transcriptome of anti-miR-195-MSC to investigate the gene expression changes, signalling pathways and mechanisms regulated by miR-195-5p inhibition. Notably, this study was conducted without the use of exogenous osteogenic supplements or MSC preconditioning protocols, relying exclusively on anti-miR-195 modulation and scaffold support. By avoiding additional biochemical priming steps, our approach simplifies the MSC manufacturing workflow and may facilitate translation of MSC-based advanced therapy medicinal products (ATMPs) for bone repair and regeneration. Materials and methods Human primary bone marrow-derived MSCs isolation and culture Mesenchymal stem/stromal cells (MSCs) were isolated from human bone marrow, obtained either by aspiration or from discarded bone tissue, after informed consent and in accordance with the Declaration of Helsinki. The cells were isolated less than 24 h after collection of the tissues, as previously reported [5, 8, 9]. Further sample information is provided in Table SI. Briefly, mononucleated bone-marrow cells were collected following a lymphoprep (Cosmo Bio USA lnc) gradient density centrifugation and cultured in low-glucose Dulbecco’s Modified Eagle’s Medium (DMEM, Corning) supplemented with 10% (v/v) MSC-qualified heat inactivated fetal bovine serum (FBS, Gibco) and 1% (v/v) of penicillin/streptomycin (P/S, Invitrogen) for 72 h, at 37 °C and 5% (v/v) CO2. Then, the media was renewed and the adherent cells left to grow. MSCs were characterized in accordance with the minimal criteria for defining multipotent MSCs, including: (i) plastic adherence; (ii) expression of CD73 (1:10, ImmunoTools), CD90 (1:10, ImmunoTools) and CD105 (1:25, eBioscience), and lack of expression of HLA-DR (1:25, eBioscience), CD14 (1:10, ImmunoTools), CD19 (1:10, ImmunoTools), CD34 (1:10, ImmunoTools) and CD45 (1:10, ImmunoTools), as assessed by flow cytometry; and (iii) evaluation of their multilineage differentiation capacity [10]. Oligonucleotides transfection MSCs were transiently transfected with a miR-195-5p inhibitor (anti-miR-195; 100 nM; catalog number: 4464084; assay ID: MH10827) and the respective miRNA inhibitor negative control (scrambled - SCR; 100 nM; catalog number: 4464076), both from Ambion, using Lipofectamine 2000 reagent (Invitrogen), according to manufacturer’s protocol. hsa-miR-195-5p FASTA format sequence annotation was obtained from the miRbase database (http://www.mirbase.org/, Supplemental file (Table SII)). MSCs were incubated for 6 h with the oligonucleotide-liposome complexes in medium without P/S. To track MSCs, transfected cells (SCR or anti-miR-195) were stained with 1 µg/mL CellTracker™ CM-DiI dye (Thermo Fisher Scientific) for 30 min. RNA isolation, reverse transcription and real-time quantitative polymerase chain reaction (RT-qPCR) Total RNA was extracted using TRIzol Reagent (Invitrogen), according to the manufacturer’s instructions. RNA concentration was measured using NanoDrop Spectrophotometer ND-1000 (Thermo Fisher Scientific). RNA integrity was determined by gel electrophoresis or through the Agilent RNA 6000 Pico kit (Agilent). For analysis of miRNA expression levels, the RNA was converted into cDNA using the TaqMan MicroRNA Reverse Transcription Kit and gene specific stem-loop Reverse Transcription primers (Applied Biosystems), according to the manufacturer’s protocol and as previously reported [5, 6, 11]. qPCR reactions were carried out in a CFX96 Touch Real-Time PCR Detection System (Bio-Rad) with the SsoAdvanced™ Universal Probes Supermix (Bio-Rad) and TaqMan probes (TaqMan® MicroRNA Assays, Applied Biosystems). The small nuclear RNA U6 was used as a reference gene. To analyze the expression of coding transcripts, total RNA samples were used to synthesize cDNA using random hexamers and the SuperScript® III Reverse Transcriptase kit (Life Technologies). qPCR was performed using SYBR Green PCR Mastermix (Bio-Rad), nuclease-free water, and specific primers (Supplemental file (Table SIII)) in a CFX Real-Time PCR Detection System (Bio-Rad), as previously reported [5, 6, 9, 11]. GAPDH was used as the reference gene. All qPCR reactions were performed in duplicate, and data analyzed using Bio-Rad CFX Manager software (Bio-Rad). Expression levels were calculated using the quantification cycle (Ct) method in accordance with MIQE guidelines [12]. For the gene expression analysis performed on animal samples, specific exclusion criteria were applied. Animals were excluded if (i) any sample from an animal showed low RNA integrity or (ii) the target genes exhibited low expression, defined as Ct values > 35 in both conditions from the same animal. RNA sequencing and bioinformatic analysis Total RNA from three different donors (aged < 37 years old) was extracted from anti-miR-195-MSCs or the respective control (SCR-MSCs), following 48 h of culture in basal conditions (complete growth media). Illumina sequencing was carried out at Novogene (UK) Company Limited. For library construction, quality control and sequencing, the messenger RNA (mRNA) was purified from total RNA using poly-T oligo-attached magnetic beads. After fragmentation, the first strand cDNA was synthesized using random hexamer primers, followed by the second strand cDNA synthesis using dTTP. The library was checked with Qubit and real-time PCR for quantification and bioanalyzer for size distribution detection. Quantified libraries were pooled and sequenced on Illumina platforms, according to effective library concentration and data amount. The clustering of the index-coded samples was performed according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina platform and paired-end reads were generated. For data analysis, raw data of fastq format were firstly processed through in-house perl scripts. In this step, clean data (clean reads) were obtained by removing reads containing adapter, reads containing ploy-N and low quality reads from raw data. Reference genome and gene model annotation files were downloaded from genome website directly. Index of the reference genome was built using Hisat2 v2.0.5 and paired-end clean reads were aligned to the reference genome using Hisat2 v2.0.5. featureCounts v1.5.0-p3 was used to count the number of reads assigned to each gene. To remove lowly expressed genes, only genes expressed in all the donors were considered for analysis. Differential expression was calculated using the edgeR package [13]. Transcripts with a FDR < 0.3 were considered differentially expressed (DEGs), to avoid overlooking potentially relevant targets in this exploratory analysis. Selected candidates were subsequently validated by RT-qPCR using samples derived from distinct donors as biological replicates. The full list of DEGs is provided in Table SIV. Enrichment analysis of Gene Ontology (GO) terms was conducted using the Protein Analysis Through Evolutionary Relationships (PANTHER) classification system [14, 15]. The Database for Annotation, Visualization, and Integrated Discovery (DAVID) tool was used for functional annotation and enrichment analyzes of the DEGs list [16, 17]. In addition, Metascape (https://metascape.org) was used to perform integrative functional enrichment and pathway analysis of the DEGs [18]. To further interpret the expression profiles, QIAGEN Ingenuity Pathway Analysis (IPA; https://digitalinsights.qiagen.com/IPA) was used to predict their impact on relevant functions [19]. Heterotopic bone formation model Animal experimentation was carried out at the i3S - Instituto de Investigação e Inovação em Saúde animal facility and approved by the i3S Animal Welfare and Ethics Review Body and the Portuguese Competent Authority (DGAV) (NT-BROD, ref. 0421/000/000/2019), and conducted in accordance with the European legislation on Animal Experimentation through Directive 2010/63/UE. Animal experimentation was reported in line with the ARRIVE guidelines 2.0. MSCs transfected with either SCR or anti-miR-195 were seeded in RegenOss scaffolds (Fin-Ceramica Faenza SpA, cylindrical shape with diameter: 8 mm and height: 2 mm) at a cell density of 0.15 × 106 cells per scaffold in DMEM low glucose supplemented with 10% (v/v) FBS for 18 h. Next, the MSC-seeded scaffolds were subcutaneously implanted into the torso of nude mice (Strain: Nude CBA, aged 9–10 weeks; N = 16). Before surgery 0.1 mL/10 g of buprenorphine was given to the animals. During the surgical procedures, the animals were maintained under general isoflurane anaesthesia. Briefly, a midline incision at the dorsal skin was performed and two scaffolds were subcutaneously implanted (left side: scaffold + SCR-MSCs; and right side: scaffold + anti-miR-195-MSCs). After the surgery, the mice received paracetamol dissolved in water. After recovery, mice were caged in pairs and housed in a clean environment. Animals were randomly divided to be euthanized at day 14 (N = 8) and day 28 (N = 8) with carbon dioxide inhalation. Two animals from the day-28 group developed scaffold-associated infections during the final days of the experiment and were therefore excluded from subsequent analyses. Data analysis was performed blinded to the treatment. Histology and immunohistochemistry analysis At 14 and 28 days post-surgery, each scaffold was bisected: one half was fixed in 10% (w/v) buffered formalin for 24 h and processed for histology and immunostaining, while the other half was used for RNA isolation with Trizol, as reported above. Fixed samples were embedded in paraffin and sectioned with a microtome at 5 μm. During processing of the scaffolds at the 14-day post-implantation time point, one was partially compromised, resulting in incomplete tissue recovery. Consequently, analyses at this time point were performed using 7 of the 8 animals to avoid potential artifacts. Scaffold sections were stained with hematoxylin-eosin (H&E), Alizarin Red and Picrosirius Red (PSR) staining. Lymph nodes were also collected, fixed and stained with H&E. The H&E staining was used to assess cell migration and quantify the formation of blood vessels within the scaffolds after 14 and 28 days post-implantation. Cell migration in the scaffolds was measured by calculating the mean of five measurements taken from one border of the scaffold to the opposite border, from the top of the scaffold surface to the deepest point where cells were observed within the scaffold. Images were acquired using the Leica DM2000 LED microscope. To assess the impact on blood vessel formation, whole-slice images were acquired by high-resolution slide digitalization using Nanozoomer S60 Ddigital Slide Scanner (Hamamatsu Photonics; Hamamatsu, Japan). Blood vessels were manually annotated based on the presence of intraluminal red blood cells (RBCs) and morphological characteristics visible in the H&E-stained images, including the presence of a vessel lumen, using the open-source QuPath software (https://qupath.github.io/). Morphological features were then extracted by exporting measurement data for subsequent quantitative analysis [20]. For the PSR staining, the sections were deparaffinized, hydrated in decreasing ethanol gradient to de-ionized water, stained with Sirius Red for 1 h and washed with acidified water. Sections were imaged under polarized light using a light microscope (Leica DM2000 LED). Based on birefringence color (green or red), fibers were classified and quantified as the percentage of thin (type III) and thicker (type I) collagen fibers [21–23] using ImageJ software [24]. For the scaffolds collected at day 14, 6 pictures (3 from the top and 3 from the middle) were captured and quantified. For those scaffolds collected at day 28, 3 pictures (3 from the top of the scaffold) were acquired and analyzed, as the middle part of most scaffolds was not intact. For the alizarin red staining, the paraffin-embedded sections were deparaffinized and rehydrated to distilled water. Then, the slides were stained with freshly prepared 2% (w/v) Alizarin Red S solution (pH ~ 4.1–4.3) for 5 min. Excess stain was removed by washing in tap water, followed by dehydration through a graded ethanol series and clearing in xylene. The Zeiss Axiovert 200 m Inverted Fluorescence microscope was used to visualize and acquire the images. The percentage of the scaffolds stained with alizarin was quantified using the ImageJ software. Whole-slide scaffold sections were used for quantification. Immunofluorescence was performed on paraffin-embedded scaffold sections following deparaffinization and rehydration. Antigen retrieval was carried out by heating the sections in 10 mM citrate buffer (pH 6.0) at 95 °C for 20 min. Prior to blocking (5% (v/v) FBS + 1.5 (w/v) % BSA), the sections were permeabilized (0.25% (v/v) Triton X-100) and treated with 50 mM ammonium chloride (NH₄Cl). After blocking, the sections were incubated overnight at 4 °C with primary antibody against CD31/PECAM-1 (1:1000, M0823, Dako). Following three washing steps, the sections were incubated for 1 h with the respective secondary antibody (1:1000, IgG (H + L) Alexa Fluor ® 647, Thermo Fisher Scientific). Nuclei were counterstained with 1 µg/mL DAPI (Invitrogen) for 5 min. Stained sections were visualized and imaged using the Operetta CLS system. The percentage of scaffold covered by cells was calculated by dividing the DAPI-stained area by the total scaffold area. Statistical analysis All graphs were generated using GraphPad Prism software, version 9 (GraphPad Software). Data sets were assessed for normality using the Shapiro-Wilk test. If the data followed a normal distribution, the Student’s t-test was used for comparisons between two groups and one-way ANOVA for comparisons involving more than two groups. If the data displayed a non-gaussian distribution, the two-tailed Wilcoxon matched pairs test was used for comparisons between two groups and the Friedman test for comparisons involving more than two groups, followed by the uncorrected Dunn’s multiple comparison test. Statistical significance was defined as p < 0.05 (* p < 0.05; ** p < 0.01; and *** p < 0.001). Results Transcriptomic profiling reveals that anti-miR-195-modulated MSCs are associated with activation of migration, invasion and cell cycle progression To identify novel molecular targets and explore the key signalling pathways regulated by anti-miR-195 in MSC, RNA-sequencing was performed at 48 h post-transfection with either miR-195-5p inhibitor or scrambled control (SCR) in 3 independent human MSC donors (Fig. 1A). Hierarchical clustering map of the Differentially Expressed Genes (DEGs) demonstrated distinct grouping of samples based on treatment conditions (Fig. 1B), highlighting treatment-specific transcriptional differences. Still, principal component analysis (PCA) revealed that the samples were primarily clustered by donor (Supplemental Fig. 1), consistent with the expected inter-donor variability in primary human samples, and highlighting the need for a paired analysis, which was therefore applied to account for these individual differences. DEGs were defined using an FDR < 0.3 to accommodate the exploratory nature of the study, with selected candidates subsequently validated by RT-qPCR (Table SIV). Based on this criterion, from a total of 14,730 genes identified (expressed in all donors, in at least one of the conditions), 436 were classified as DEGs, with 229 genes upregulated and 207 downregulated across three independent experiments (each corresponding to a different donor) (Fig. 1C). Interestingly, intersecting the DEG list with experimentally validated miR-195-5p targets from miRTarBase (limited to those validated by reporter assays, western blot or qPCR), revealed four common targets, all of which were upregulated, as expected upon miR-195-5p inhibition (Table SV). Then, KEGG pathway enrichment analysis was performed using DAVID to identify biological pathways associated with the DEGs. Upon treatment with anti-miR-195, 11 pathways were predicted to be altered (p < 0.05), with the Cell cycle (12 DEGs) being ranked with the highest significance (Fig. 1D). Although the transcript of the proliferation marker Ki-67 (MKI67) is not annotated within the Cell cycle pathway, it was highly upregulated (FC = 2.02; p = 1.78 × 10− 4; FDR = 0.0971) following miR-195-5p inhibition. This analysis also revealed enrichment in additional pathways, including PI3K-Akt signaling pathway (17 DEGs), Cytokine-cytokine receptor interaction (15 DEGs), MAPK signaling pathway (15 DEGs), p53 signaling pathway (6 DEGs) and Calcium signaling pathway (11 DEGs) (Fig. 1D), which are fundamental when envisioning a cell-based therapeutic approach and suggesting that miR-195-5p inhibition might enhance cell integration and function within host tissues. Further corroborating the previous in vitro findings [8] and RNA sequencing analysis, the Disease and Functions analysis of IPA predicted that miR-195-5p inhibition enhances Cellular movement and Cell cycle progression (Fig. 1E). Conversely, the analysis suggested a predominant suppression of the Cell death and survival function, in which subcategories such as Cell death, Apoptosis and Necrosis are predicted to be strongly inhibited, while cell viability and survival are estimated to be activated (Fig. 1E). Also, within the Skeletal and muscular system development and function, the subcategories Proliferation, Migration, Function and Bone mineral density were predicted to be activated (Fig. 1E). Bulk RNA sequencing analysis reveals downstream targets and molecular pathways modulated by anti-miR-195 in MSC. A Scheme of the experimental setup. MSCs were transfected with either SCR or anti-miR-195 and bulk RNA sequencing performed after 48 h under growth culture conditions (N = 3). B Sample clustering based on the differently expressed genes (DEGs). C Venn diagram showing the overlap between DEGs (FDR < 0.3) and experimentally validated miR-195-5p target genes retrieved from miRTarBase. D Bubble plot graph illustrating the KEGG pathways predicted to be enriched when considering the DEGs using DAVID. E Tree maps of Disease and Functions based on the Ingenuity Pathway Analysis (IPA). The colour gradient represents the activation status (z-score) of each biological function (darker blue: inhibition; and darker orange: activation). The size of the rectangles reflects the p-value significance (-Log(p-value)). The most significantly enriched categories have the largest rectangles miR-195-5p inhibition affects gene signatures linked to Bone Density Gene ontology (GO) enrichment analysis, performed using the PANTHER tool, identified 7 DEGs associated with the Locomotion category (Fig. 2A). This is consistent with the predicted activation of this function by IPA (Fig. 1E). Moreover, several Cell adhesion proteins were differentially expressed, indicating potential modulation of cell–matrix interaction pathways that may influence MSC migratory behavior. In line with these findings, terms related to Bone density and Bone mineral density were enriched in our datasets. Specifically, 17 DEGs were associated with Bone mineral density (p = 0.029), based on GAD_Disease annotations in DAVID (Fig. 2B). Among these, bone sialoprotein (IBSP; FC = 2.49; p = 0.00354; FDR = 0.220), a hallmark of late-stage osteogenic differentiation and a key regulator of matrix mineralization, was upregulated. In addition, a significant representation of extracellular matrix–related components, both Core matrisome and Matrisome associated were enriched, indicating transcriptional modulation of structural ECM proteins and ECM-associated regulators (Fig. 2C and D). Within these groups, the expression of one osteogenic and one vascularization-related transcript was validated by RT-qPCR using an expanded cohort of donors (N = 6). Results show that IBSP expression levels were significantly increased (FC = 2.32; p < 0.001), in contrast to the expression levels of the anti-angiogenic gene THBS1, which were significantly downregulated (p < 0.001) (Fig. 2E). In addition, the osteogenic marker SPP1 and the pro-angiogenic factor VEGFA, detected by RNA sequencing and exhibiting consistent upregulation across all three donors tested, were also validated (Fig. 2E), corroborating our previous in vitro findings [8] and aligning with the enrichment of the terms Regulation of angiogenesis and Extracellular matrix organization (Supplemental Fig. 2), as well as with the increased deposition of calcium nodules observed at 28 days under basal conditions in anti-miR-195-treated MSCs compared with control cells (Supplemental Fig. 3). Functional terms enriched among differentially expressed genes. A Heatmap showing the DEGs associated with the Gene Ontology (GO) terms Locomotion and Cell adhesion. The colour gradient represents the differential expression profile of each donor sent to RNA sequencing analysis, when compared to the respective SCR (darker yellow: upregulation; and darker blue: downregulation). B Bar plot illustrating the Bone mineral density associated DEGs transcripts using DAVID. C Bubble plot graph illustrating canonical pathway clusters associated with the matrisome (Metascape) and D Table with the respective genes encoding matrisome linked proteins. E Validation of DEGs and genes associated with osteogenesis and vascularization, following 48 h of culture in basal conditions (complete growth media) (N = 6) In vivo heterotopic bone formation model to test effects of miR-195-5p transfected MSCs on bone regeneration Anti-miR-195-MSCs and the respective control, SCR-MSCs, were seeded onto RegenOss, which provide a supportive matrix that mimics the major components of the bone extracellular matrix, namely hydroxyapatite and collagen (Fig. 3A). Anti-miR-195 transfection efficiently reduced miR-195-5p expression levels across distinct donors, with inhibition levels of approximately 89.7%, 97.2% and 85.9% relative to SCR (control), confirming successful transfection by RT-qPCR (Fig. 3B). MSCs were efficiently labelled with CellTracker™ CM-DiI Dye in vitro (Fig. 3B; left) and their presence within the implanted scaffolds confirmed in vivo, even after 28 days, using fluorescence microscopy (Fig. 3B; right). Although the collagen/hydroxyapatite scaffolds exhibit an intrinsic autofluorescence, the CellTracker-labelled MSCs display a distinct and clearly discernible signal, allowing reliable identification of cells within the scaffold (Supplemental Fig. 4). At 14 and 28 days post-implantation, blood vessel formation was observed at the surface of the scaffolds (Fig. 3C), which may suggest that the MSC-scaffolds have the potential to support neovascularization. Animal welfare was continuously monitored throughout the experiment. Two animals developed infections and were excluded from the study (28 days post-implantation group). As anticipated, the animals exhibited significant weight gain by day 28, when compared with day 0 (implantation day), with no significant differences observed by day 14 (Fig. 3D). Moreover, the lymph nodes draining the subcutaneously implanted MSC-seeded scaffolds showed no detectable structural alterations (Fig. 3E), supporting a favourable biocompatibility profile and a minimal immune response to the implanted scaffolds, even in nude CBA mice, which retain components of the innate immune system despite their T-cell deficiency. In vivo implantation of anti-miR-195-MSC and SCR-MSC on scaffolds. A Experimental setup. MSCs were transfected with either SCR or anti-miR-195, seeded onto collagen and hydroxyapatite scaffolds, and implanted into nude mice for 14 and 28 days. B miR-195-5p expression measured by RT-qPCR 48 h after transfection (N = 3), with each dot representing an individual donor, and the corresponding fold-change of inhibition relative to the respective control (SCR); and fluorescence microscopy images showing MSCs stained with CellTracker CM-DiI dye in vitro (left bottom corner) and post-tissue processing 28 days after implantation in vivo (red: cell tracker and blue: DAPI). C Representative images of the animals at 14 and 28 days post-implantation with the scaffolds seeded with MSCs. D Quantification of animal body weight (N = 8, per time point). E Draining lymph nodes stained with Hematoxylin and Eosin (H&E) (scale bar: 25 μm) anti-miR-195-MSCs enhance cell migration and scaffold colonization Histological analysis of the scaffolds at day 14 post-implantation supports that anti-miR-195-MSCs increase overall cell migration within the scaffold (Fig. 4A). The results show a 40% increase in the migration capacity of cells (MSC and/or host infiltrating cells) within the anti-miR-195-MSC scaffolds compared to control scaffolds, with an average cell penetration depth of 591.0 μm in the anti-miR-195-MSC condition versus 377.3 μm in the control (Fig. 4A). By day 28, cells in both conditions have migrated/infiltrated throughout the full scaffold thickness, rendering migration depth analysis infeasible. Moreover, scaffold colonization was quantified by measuring the cell-covered area using DAPI staining, which demonstrated complete cellular coverage of the scaffold surface at day 28 post-implantation for anti-miR-195-MSC (Fig. 4B). Although the difference did not reach statistical significance, scaffolds seeded with anti-miR-195-MSCs exhibited a trend toward increased cell infiltration in 4 out of 6 animals at day 28 post-implantation (Fig. 4B). Overall, these results suggest an increased cell recruitment and enhanced chemoattractive potential and/or proliferative capacity in the anti-miR-195-MSCs loaded scaffolds, compared with the SCR-MSCs loaded scaffolds. Impact of anti-miR-195-MSCs on cellular migration. A Representative image of scaffolds collected 14 and 28 days post-implantation (scale bar: 200 μm). The orientation of the scaffold, was identified based on the surrounding tissue architecture, particularly the presence of skin, which indicates the implantation side. This also corresponds to the upper part of the scaffold, where the MSCs were initially seeded. Quantification of depth of cellular migration in anti-miR-195-MSC scaffolds and SCR-MSC scaffolds at day 14 (N = 7). Each point in the cell migration graph represents the mean of five measurements obtained across the middle section of each scaffold. B Representative images (scale bar: 2 mm; day 28) and percentage of scaffold area covered by cells (N = 7 for day 14 post-implantation; and N = 6 for day 28 post-implantation) anti-miR-195-MSCs promote collagen deposition and osteogenic markers in vivo Collagen production was assessed using Picrosirius Red (PSR) staining of scaffold sections (Fig. 5A). Although no statistical differences were observed at day 14 post-implementation, scaffolds seeded with anti-miR-195-MSCs exhibited a statistically significant increase in both type I and type III collagen deposition by day 28 (Fig. 5A), demonstrating that extracellular matrix deposition was increased in the anti-miR-195-MSC group. As expected, mineralization was minimal at day 14 post-implantation and increased by day 28, consistent with the temporal progression of calcium deposition (Fig. 5B). Moreover, while at day 14 no differences were observed between groups (p = 0.6875; Fig. 5B), at day 28 all six animals in the anti-miR-195-MSC group exhibited higher percentages of Alizarin Red-stained area than controls, although this difference did not reach statistical significance (p = 0.0634; Fig. 5B). This pattern is expected, as mineralization is minimal at 14 days post-implantation. Moreover, these in vivo results corroborate our previous in vitro findings demonstrating the pro-osteogenic effects mediated by anti-miR-195-regulated MSCs [8]. Guided by our bulk RNA sequencing data, we then extracted RNA from the heterotopic bone formation subcutaneous model and validated selected proliferative/cell cycle and osteogenic related transcripts. At day 14, anti-miR-195-MSC-loaded scaffolds show a 2.70 fold increase in the expression of CDK6 (p = 0.0092), key regulator of the G1 phase of the cell cycle, but no differences for CDK1, coordinating the G2 to M transition (p = 0.9289) (Fig. 5C). Also, at this time point, RUNX2 was significantly downregulated (FC = 0.53; p = 0.0027) in anti-miR-195-MSC compared with control, and SPP1 expression was not detected in any of the groups. On the other hand, at day 28 post-implantation, no differences in cell cycle progression markers CDK1 (p = 0.1207) and CDK6 (p = 0.1606) were observed, while a significant 1.564-fold increase in RUNX2 (p = 0.0144) expression and a 3.488-fold increase in SPP1 (p = 0.0079) expression were detected in anti-miR-195-MSC-scaffolds compared with control-MSC scaffolds (Fig. 5C). These observations are consistent with a pro-osteogenic profile in anti-miR-195-treated MSCs after 28 days of implantation. In vivo effect of anti-miR-195-MSCs scaffolds on collagen deposition, mineralization and gene expression. A Representative images (at day 28 post-implantation; acquired under polarized light microscopy to visualize collagen birefringence) (scale bar: 100 μm) and quantification of PSR-stained area (N = 7 for day 14 post-implantation; and N = 6 day 28 post-implantation). B Representative images of anti-miR-195-MSC scaffolds and SCR-MSC scaffolds at day 14 and 28 post-implantation (scale bar: 100 μm) and quantification of Alizarin Red-stained area (N = 7 for day 14 post-implantation; and N = 6 day 28 post-implantation). C Gene expression by RT-qPCR of CDK6, CDK1, RUNX2 and SPP1 in anti-miR-195-MSC scaffolds and control-MSC scaffolds after in vivo implantation (N = 6 for day 14 post-implantation; and N = 5 day 28 post-implantation) anti-miR-195-MSCs promote vascular infiltration in vivo Analysis of H&E-stained sections revealed a progressive increase in vascularization within the scaffold area from day 14 to day 28 post-implantation. Blood vessels from whole-slide images were manually annotated based on the presence of intraluminal red blood cells and morphological features, such as vessel area. Importantly, CD31 immunostaining confirmed that the vascular structures identified and quantified in the H&E-stained sections corresponded to endothelial-lined vessels (Supplemental Fig. 5). Although a tendency toward increased vessel coverage was noted at day 14, this difference did not reach statistical significance between groups (Fig. 6A). Meanwhile, the area of the scaffold covered by blood vessels showed a significant increase at day 28 in the anti-miR-195-MSCs group compared to SCR-MSCs group (Fig. 6B and C). Moreover, while no differences were detected at day 14 (Fig. 6A), by day 28 the average vessel area showed a trend toward larger vessels in the anti-miR-195-MSC group, with increases observed in 6 of the 7, in addition to the increased vessel coverage (Fig. 6B). Interestingly, at day 14, the number of vessels within the scaffolds was significantly higher in the anti-miR-195-treated condition compared to controls (Fig. 6D), whereas no differences were detected for day 28 (Fig. 6E). Results were further stratified by size, with large vessels defined as those having areas above the third quartile (Q3) of each animal’s vessel size distribution. No significant differences in the number of large vessels were observed at day 14 (Fig. 6D). In contrast, by day 28, the anti-miR-195-treated condition exhibited a significantly higher number of large vessels compared to controls (Fig. 6E). These results suggest that, compared to controls, anti-miR-195-MSCs promote vessel formation at day 14, indicating early angiogenic activation and a higher prevalence of larger vessels by day 28, consistent with ongoing vascular development. Quantification of vascularization within the implanted scaffolds. A Quantification of scaffold area covered by blood vessels (left), average vessel area (middle) and fold change of the average vessel area (right) at day 14 post-implantation (N = 7). B Quantification of scaffold area covered by blood vessels (left), average vessel area (middle) and fold change of the mean vessel area (right) at day 28 post-implantation (N = 6). C Representative images of whole-slide sections from day 28 (SCR, left; anti-miR-195, right). Green: outline of the scaffold; Blue: outline of the vessels within the scaffold (scale bar: 500 μm). D Fold change of the number of vessels normalized to total scaffold area (left) and number of vessels normalized to total scaffold area (middle) at day 14 post-implantation (N = 7). Number of large vessels, defined as those larger than the Q3 of vessel area for each animal (right) (N = 7). E Fold change of the number of vessels normalized to total scaffold area (left) and number of vessels normalized to total scaffold area (middle) at day 28 post-implantation (N = 6). Number of large vessels, defined as those larger than the Q3 of vessel area for each animal (right) (N = 6) Discussion MSCs possess a unique combination of biological properties, including multipotency, immunomodulatory effects, low immunogenicity and paracrine activity, that make them promising cellular candidates for therapeutic applications in regenerative medicine, tissue engineering and immunomodulation [25, 26]. These characteristics have led to extensive investigation of MSC clinical trials, targeting a wide spectrum of diseases, including on the musculoskeletal system, such as osteoporosis and osteoarthritis. However, the clinical efficacy of MSC-based therapies remains limited by challenges such as poor in vivo survival and limited engraftment [27, 28]. Currently, genetic and epigenetic engineering approaches, such as modulation with miRNAs, small interfering RNAs (siRNAs), or bioactive small molecules, are being actively explored as strategies to overcome these limitations [29–31] and enhance specific MSC functions, including proliferation, osteogenesis and angiogenesis. In our previous study, we identified miR-195-5p as a negative regulator of MSC function, demonstrating its anti-proliferative and anti-osteogenic effects, while also shifting the MSC secretome toward an anti-angiogenic profile in vitro [8]. Building on these findings, in the present study we tested the effect of anti-miR-195 modulation in vivo to assess the impact on MSC-mediated regenerative properties, while investigating the mechanisms and pathways regulated by anti-miR-195. Notably, our approach required no exogenous osteoinductive factors (e.g., dexamethasone, β-glycerophosphate and/or ascorbic acid). In contrast to several studies that induce osteogenic lineage commitment of MSCs, through in vitro pre-differentiation, prior to in vivo implantation, our approach relied solely on miRNA-based priming of the cells before in vivo delivery. In addition, a collagen/hydroxyapatite scaffold was employed as a bone substitute in the heterotopic bone formation model to provide a structural bone-like matrix. Moreover, the use of scaffold approaches may facilitate clinical translation, as direct MSC injection has demonstrated limited efficacy in large bone defects and non-unions [32, 33]. Another important feature of our experimental approach was the use of transient gene modulation, which minimizes risks of unintended gene disruption and of long-term or off-target effects [34]. Importantly, the transient and non-viral nature of the anti-miR-195 modulation strategy may represent an advantage from a regulatory perspective compared with permanent genetic engineering approaches, although additional product characterization, including assessment of transfection efficiency, persistence and potential off-target effects, would still be required during product development. Despite the transient nature of anti-miR-195 transfection, we still observed sustained regenerative effects at day 28, including increased collagen deposition and vascularization in anti-miR-195-MSC scaffolds. We hypothesize that these outcomes may result from: (i) activation, at the initial stages of anti-miR-195 delivery, of key regenerative pathways (e.g., cell cycle, migration and osteogenesis); and/or (ii) transient paracrine signaling from transfected MSCs, which initiates host tissue remodeling cascades. However, these mechanisms remain to be validated through assessment of miR-195-5p levels, target engagement and secretome dynamics at later time points. Such longitudinal analyses, in combination with single-cell profiling of the scaffold microenvironment, would further clarify these hypotheses. Importantly, our data demonstrated a stage-specific response pattern following anti-miR-195 modulation, characterized by distinct outcomes at day 14 and day 28 post-implantation. This temporal pattern likely reflects the combined influence of multiple, interdependent factors, such as: (i) a gradual decline in anti-miR-195 levels after transfection, leading to transient but early inhibition of miR-195-5p targets; (ii) progressive infiltration of host cells within the scaffold, which modify the local cytokine and growth factor milieu over time; and (iii) scaffold degradation kinetics, which gradually expose new surface areas, releasing bioactive fragments and altering mechanical cues. Moreover, this temporal regulation suggests that the effects of anti-miR-195 emerge progressively and may align with the tissue remodeling phase. Within the context of tissue repair, day 14 likely corresponds to the early angiogenic phase of bone regeneration, when pro-angiogenic signaling and matrix deposition are initiated [35, 36]. Although vessel formation was less pronounced than at day 28, the presence of vessels and the increased vessel number at day 14 in the anti-miR-195-MSC group suggest an enhanced early angiogenic response. This data is also consistent with our previous results on the chick chorioallantoic membrane assay, where we demonstrated the anti-angiogenic effect of miR-195-5p, mediated by the downregulation of VEGF [8]. By contrast, the day 28 time point reflects a transition toward the tissue remodeling and maturation phase, characterized by collagen matrix organization, vessel enlargement and mineral deposition [36, 37]. The higher percentage of mineralized area, greater vessel coverage and presence of larger vessels at day 28 are consistent with this phase, supporting the notion that transient miR-195-5p inhibition triggers early signaling that translates into sustained structural and functional improvements. These interpretations align with established temporal frameworks of bone healing, where early angiogenesis is followed by matrix organization and mineralization [35–37]. At the mechanistic level, the phenotypic effects observed following anti-miR-195 modulation are likely mediated by a combination of direct and indirect regulatory mechanisms. miRNAs typically act within regulatory networks in which direct repression of a limited set of target genes initiates downstream signaling cascades that subsequently shape broader transcriptional and paracrine responses [38–40]. Consistent with this model, RNA sequencing showed selective, rather than uniform, upregulation of all migration-, angiogenesis- or extracellular matrix–related transcripts, arguing against a purely direct regulatory effect. Instead, these findings support a hierarchical response driven by a subset of direct miR-195-5p targets acting as upstream effectors. In vivo, RUNX2 expression was reduced at day 14 in the anti-miR-195-MSC group. RUNX2 expression is known to be temporally regulated during osteoblast differentiation and bone repair, and can decrease as cells progress through early to intermediate stages of matrix deposition and remodeling, even when downstream osteogenic markers continue to increase [41]. Therefore, the decrease observed at day 14 may reflect a transient re-organization of osteogenic signaling rather than a loss of osteogenic potential. This interpretation is supported by the subsequent increase in osteogenic markers and mineral deposition observed at day 28. Nonetheless, alternative explanations, including cellular stress or an altered MSC phenotypic state [42, 43], cannot be excluded in the absence of direct proliferation, apoptosis or stress-response markers. A limitation of our study is the use of MSCs derived from young adult donors in vivo and for RNA sequencing, rather than aged individuals, which may not fully replicate the regenerative potential in elderly patients, who represent the primary target population for such therapies [44]. However, transfection efficiency is not affected by the MSC donor age [6, 8, 9], supporting the translational relevance of our findings. Moreover, although the heterotopic implantation model used in this study does not fully recapitulate all aspects of the native bone microenvironment, such as mechanical loading, marrow interactions and vascular perfusion architecture, it remains a well-established and widely used approach for evaluating the osteogenic and angiogenic potential of MSC-based therapies in vivo [45, 46]. Nevertheless, the use of orthotopic bone-defect models (e.g., calvarial or femoral defects) would provide a more physiologically relevant setting and should be considered in future studies to further substantiate the therapeutic potential of anti-miR-195-engineered MSCs in bone regeneration. From a translational perspective, this strategy may be particularly relevant for scaffold-assisted repair of critical-size bone defects and delayed/non-unions, where vascularization and osteogenesis are essential for regeneration. In these settings, the combination of anti-miR-195-engineered MSCs with biomaterial scaffolds may provide structural support while simultaneously enhancing the regenerative microenvironment. Although both autologous and allogeneic MSC-based approaches could potentially be compatible with this strategy, an allogeneic platform would likely be favored given the manufacturing timelines and large cell numbers required for this type of therapy. In addition, several complementary analyses, while beyond the scope of the present study, would further enhance the mechanistic and functional interpretation of these findings. These include evaluation of functional vascular perfusion to determine whether the increased vessel formation observed histologically translates into effective tissue perfusion and hemodynamic functionality, characterization of inflammatory responses within the scaffold, incorporation of µCT-based bone morphometry, quantitative assessment of mineralized tissue formation and biomechanical testing to enable a more comprehensive evaluation of functional bone regeneration, and an in vivo characterization of MSC fate and mechanism of action, as the study did not assess long-term cell survival, phenotypic maintenance, differentiation or viability. In addition, comprehensive safety characterization, including biodistribution, persistence and potential off-target effects of anti-miR-195-modified MSCs, will be necessary prior to clinical translation. Comparative evaluation with clinically established osteoinductive agents, such as BMP-2, may also be valuable in future studies to benchmark anti-miR-195-engineered MSC therapy against current standards of care. Finally, although the in vivo persistence and biodistribution of anti-miR-195 were not quantified and direct mechanistic causality was not established, future studies incorporating pharmacodynamic analyses, rescue experiments and target validation will be required to confirm the pathways regulated by anti-miR-195. In conclusion, this work provides a proof-of-concept that transient anti-miR-195 engineering enhances MSC-mediated vascularization, matrix deposition and osteogenic gene expression in vivo. Transcriptomic profiling further revealed that miR-195-5p inhibition is associated with pathways related to cell cycle progression, migration, and cellular movement, together with increased expression of osteogenic and pro-angiogenic genes including IBSP, SPP1 and VEGFA, supporting a broader role in regulating MSC functional activity. While the mechanistic basis underlying the sustained effects requires further investigation, these findings support miR-195-5p inhibition as a promising transient engineering strategy to enhance MSC-mediated bone regeneration, particularly in scaffold-assisted repair approaches for critical-size defects and non-unions. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1. Acknowledgements The authors would like to thank to Serviço de Ortopedia e Traumatologia of Centro Hospitalar Universitário de São João (CHUSJ) for kindly donating the bone-marrow samples. The authors alson acknowledge the Advanced Light Microscopy, BioSciences Screening and Histology and Electron Microscopy, all members of the PPBI (PPBI-POCI-01-0145-FEDER-022122) and Cell Culture and Genotyping (CCGEN), services/platforms at i3S for the technical support. We specially thank to Fin-Ceramica Faenza S.p.A., Italy for providing the Regenoss scaffolds as bone substitutes. Author contributions Conceptualization: SRM and MIA; methodology: SRM, CC, SGS and MIA; data collection: SRM, MMC, IB, IA, CC, SGS and MIA; statistical analysis: SRM and MIA; writing, original draft preparation: SRM and MIA; writing, review and editing: SRM, MMC, IB, IA, CC, SGS and MIA. All authors have read and approved the final version of the manuscript for publication. Funding The project was supported by AO CMF (AOCMFS-21–23 A) and by FCT – Fundação para a Ciência e a Tecnologia (2023.08681.CEECIND/CP2844/CT0009 and 10.54499/CEECIND/00184/2017/CP1392/CT0001). The project was also financially supported from FEDER and national funds (FCT) related to operation no. 16123, operation code in the Funding Portal COMPETE2030-FEDER-00727500, under the Call for Applications no. MPr-2023-12. Data availability The dataset supporting the conclusions of this article is included within the article and its additional files (Supplemental Table SIV). Additional details are available from the corresponding author on reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement The dataset supporting the conclusions of this article is included within the article and its additional files (Supplemental Table SIV). Additional details are available from the corresponding author on reasonable request.