Isolation and Characterization of Marrow-Isolated Adult Multilineage Inducible (MIAMI) Cell-Derived Extracellular Vesicles Demonstrate Multifunctional Therapeutic Potential in Tissue Regeneration and Anti-Inflammatory Immunomodulation

TL;DR

Marrow-isolated adult multilineage inducible (MIAMI) cell-derived extracellular vesicles (MIA-EVs) show strong regenerative and anti-inflammatory effects, making them promising for tissue repair and immune modulation in disease environments.

Contribution

MIA-EVs are shown to have a conserved miRNA core and robust therapeutic potential under inflammatory and stress conditions, outperforming conventional MSC-EVs.

Findings

MIA-EVs accelerate keratinocyte wound closure and suppress osteosarcoma cell proliferation in vitro.

MIA-EVs maintain a conserved miRNA backbone across different conditions, enhancing stress resilience and immune reprogramming.

MIA-EVs promote M2 macrophage polarization and show consistent regulatory identity compared to MSC-EVs.

Abstract

What are the main findings?
MIAMI cells produce highly pure, exosome-enriched extracellular vesicles (MIA-EVs) with strong regenerative and immunomodulatory molecular signatures.MIA-EVs accelerate keratinocyte wound closure and suppress osteosarcoma cell proliferation in vitro.

MIAMI cells produce highly pure, exosome-enriched extracellular vesicles (MIA-EVs) with strong regenerative and immunomodulatory molecular signatures.

MIA-EVs accelerate keratinocyte wound closure and suppress osteosarcoma cell proliferation in vitro.

What is the implication of the main finding?
Their conserved regulatory identity under inflammatory and irradiation priming supports robust translational use in hostile disease environments.Compared with MSC-EVs, MIA-EVs amplify a conserved mesenchymal miRNA core into an expanded regulatory network that enhances stress resilience, regeneration, and immune reprogramming.

Their conserved regulatory identity under inflammatory and irradiation priming supports robust translational use in hostile disease environments.

Compared with MSC-EVs, MIA-EVs amplify a conserved mesenchymal miRNA core into an expanded regulatory network that enhances stress resilience, regeneration, and immune reprogramming.

Marrow-isolated adult multilineage inducible (MIAMI) cells are a subpopulation of mesenchymal stem/stromal cells (MSC) with enhanced self-renewal, multilineage plasticity, and anti-inflammatory properties, suggesting that their extracellular vesicles (MIA-EVs) may confer advantages over conventional MSC-EVs. MIAMI cells were transcriptionally profiled and expressed regenerative markers, including PDGFRB, CDX2, and TERT. We report the first successful isolation and characterization of MIA-EVs. EVs were isolated by ultracentrifugation and characterized by nanoparticle tracking analysis, transmission electron microscopy, flow cytometry, and surface markers. Cargo analysis identified growth factors (IGFBP-1, HGF, VEGF-D) and 19 highly expressed miRNA targeting survival, regenerative, and immune regulatory pathways. MIA-EVs were efficiently internalized, enhanced keratinocyte wound closure and suppressed osteosarcoma proliferation in vitro. Conditioned MIA-EVs reshaped pathway weighting without altering core regulatory identity, as a conserved 15-miRNA backbone persisted across naïve, irradiated, and cytokine-primed states. In contrast, a 9-miRNA core shared with MSC-EVs defined a basal mesenchymal framework, while MIA-EVs expanded regenerative, survival, and immune network connectivity. Similar to embryonic stem cell (ESC)-EVs, both MIA- and cytokine-primed EVs promoted M2 macrophage polarization, selectively upregulating IL1R2 and PPARG/STAT1, respectively. Meanwhile, MSC-EVs induced heterogeneous responses. These findings establish MIA-EVs as a conditioning-resistant, systems-regulated, cell-free platform with regenerative, immunomodulatory, and cytoprotective potential under hostile microenvironments.

Figures

Full text

1. Introduction

Extracellular vesicles (EVs) are nano- or microscale membrane-bound vesicles released by cells that mediate intercellular communication by transporting proteins, lipids, and nucleic acids. Enclosed by a phospholipid bilayer resembling that of the parent cell, EVs protect their molecular cargo from enzymatic degradation, ensuring stability and prolonged bioactivity [1,2,3,4,5]. EVs cannot replicate on their own, are secreted by virtually all cell types under physiological and pathological conditions and are present in various tissues and bodily fluids [6]. They modulate recipient cell behavior, evade immune detection, and mediate both local and systemic signaling [7].

Advancements in EVs isolation and characterization have expanded the understanding of their subtypes and therapeutic potential. Because EVs can be administered systemically, they offer minimally invasive delivery and broad biodistribution, advantages over many cell-based therapies that require in situ delivery for engraftment, survival, and integration [8,9]. Unlike live-cell therapies, EVs do not pose a risk of immune rejection or tumor formation, making them a potentially safer alternative for clinical use [10,11,12]. Moreover, EVs can be engineered for targeted delivery, enhancing therapeutic efficacy and minimizing off-target effects [13]. These features make EVs a promising cell-free platform in regenerative medicine.

While stem cell-based therapies also hold regenerative potential, their clinical use is limited by challenges such as poor engraftment, ethical concerns, complex delivery requirements, and loss of viability during cryopreservation [14,15]. EVs-based therapies, by contrast, offer a scalable, non-immunogenic, and standardized alternative with potential for off-the-shelf manufacturing [8,16,17,18]. Though they lack self-regulatory properties and have a shorter biological half-life [19], advances in EVs engineering and delivery continue to improve their promising therapeutic value [20,21].

Mesenchymal stem/stromal cell-derived EVs (MSC-EVs) have demonstrated regenerative and immunomodulatory effects. Through the delivery of cytokines, growth factors, and regulatory RNAs such as miRNAs, they promote tissue repair, reduce inflammation, and support angiogenesis. MSC-EVs have shown efficacy in preclinical models of wound healing, osteoarthritis, ischemia–reperfusion, and neurodegeneration [22,23,24,25,26,27,28,29,30,31]. Their relatively low immunogenicity allows for safer and more controlled therapeutic use [8].

Marrow-isolated adult multilineage inducible (MIAMI) cells are a subpopulation of MSCs distinguished by developmental immaturity, multilineage potential, and sustained expression of embryonic stem cell markers [32]. Their secretome is enriched in pro-reparative and immunomodulatory factors, supporting angiogenesis and tissue repair. MIAMI cells are cultured under conditions mimicking the bone marrow niche, including low oxygen (3% pO_2_), low serum, and lipid supplementation. This environment preserves stemness and enhances function by upregulating self-renewal transcription factors under low oxygen tension [33].

Unlike human embryonic stem cells (hESCs), MIAMI cells are derived from adult bone marrow and are not subject to ethical or regulatory concerns. While hESCs display broader pluripotency, their use is limited by unintended consequences including tumorigenic risk. Proteomic analyses show MIAMI cells share 62% ± 3% of their proteome with human MSCs (hMSCs) and 53% ± 5% with hESCs, compared to only 32% ± 7% overlap between hMSCs and hESCs, reflecting an intermediate phenotype [34]. This profile suggests that MIAMI cells could combine the safety and accessibility of adult stem cells with expanded differentiation potential, supporting their clinical relevance.

In regenerative and oncologic settings, EVs-producing cells are frequently exposed to genotoxic stress and inflammatory cytokines, such as those present after irradiation or within chronically inflamed tissues. These stimuli are known to modulate EVs cargo composition and biological activity, often enhancing cytoprotective, immunoregulatory, or stress-adaptive signaling [35]. Evaluating how MIAMI-derived EVs respond to irradiation and cytokine priming is therefore critical to determine whether their regenerative and immunomodulatory functions are preserved, refined, or disrupted under clinically relevant hostile microenvironmental conditions.

Although MSC-EVs have been extensively studied, EVs derived from MIAMI cells (MIA-EVs) remain unexplored. Given the unique properties of MIAMI cells, their EVs may exhibit enhanced anabolic and immunomodulatory activities. Comprehensive characterization of MIA-EVs is therefore essential to determine their bioactivity, functional properties, and potential clinical relevance. In this study, we isolated and characterized MIA-EVs to assess their regenerative capacity and therapeutic potential.

2. Materials and Methods

2.1. Cell Lines and Media

Human bone marrow-derived MIAMI cells were cultured in DMEM (Dulbecco’s modified Eagle medium; Gibco, Waltham, MA, USA) supplemented with 1 g/L glucose, L-glutamine, 110 mg/L sodium pyruvate, 3% fetal bovine serum (FBS; Avantor), 1% penicillin–streptomycin (100 U/mL and 1000 U/mL, respectively; Gibco), 1% HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; Gibco), 50 µM ascorbic acid (Thermo Fisher Scientific, Waltham, MA, USA), and 0.05% of a mix of lipids solution. This formulation is referred to as MIAMI medium throughout the manuscript.

Human bone marrow-derived MSCs were cultured in DMEM supplemented with 1 g/L glucose, L-glutamine, 110 mg/L sodium pyruvate (Gibco), 10% FBS (Avantor, Radnor, PA, USA), 1% penicillin–streptomycin (Gibco), and 1% HEPES (Gibco), henceforth referred to as MSC medium.

HaCaT cells (human keratinocytes) were cultured in DMEM supplemented with 4.5 g/L glucose, L-glutamine, 110 mg/L sodium pyruvate (Gibco), 10% EVs-depleted FBS (Avantor), 1% penicillin-streptomycin (Gibco), and 1% HEPES (Gibco), hereafter referred to as HaCaT medium.

KHOS cells (human osteosarcoma) were cultured in RPMI 1640 with L-glutamine (Gibco), supplemented with 10% EVs-depleted FBS (Avantor), 1% penicillin-streptomycin (Gibco), and 1% HEPES (Gibco), referred to as KHOS medium.

2.2. Cell Culture and Expansion

2.2.1. MIAMI Cells

Whole human bone marrow was obtained from cadaveric donors; the University of Miami Institutional Review Board (IRB) determined that this activity does not constitute research involving human subjects as defined by DHHS and FDA regulations. Samples were collected within 2 h post mortem from three donors (a 12-year-old female, a 21-year-old male, and a 31-year-old female). Cells were directly plated at a density of 1 × 10^5^ cells/cm^2^ onto fibronectin-coated (10 ng/mL) dishes (Thermo Fisher Scientific, Waltham, MA, USA), without prior gradient centrifugation, immunoselection, or immunodepletion, as previously described [36]. Cell cultures were maintained at 37 °C in a tri-gas incubator, under 5% (v/v) CO_2_ and 3% (v/v) O_2_ (low oxygen) to mimic the native bone marrow microenvironment. After 14 days, non-adherent cells were removed, and single-cell-derived colonies were isolated and expanded at low density (≤50% confluency) on fibronectin-coated flasks in MIAMI medium. The resulting population, referred to as MIAMI cells [32], was used for subsequent experiments.

2.2.2. MSC

Whole bone marrow aspirates were obtained from de-identified adult donors; the University of Miami IRB determined that this activity does not constitute research involving human subjects as defined by DHHS and FDA regulations. MSC were subsequently isolated from the aspirates of three donors (a 33-year-old female, a 34-year-old male, and a 37-year-old male). Mononuclear cells obtained by Percoll gradient separation (Sigma, St. Louis, MO, USA) were seeded at a density of 2 × 10^5^ cells per 175 cm^2^ flask in MSC medium. Cells were cultured in a humidified incubator at 37 °C and 5% (v/v) CO_2_ until reaching approximately 80% confluency (passage 0; P0), then expanded at a 1:10 ratio until passage 3 (P3). For passaging, cells were detached using TrypLE™ Select Enzyme 1× (Gibco, Thermo Fisher Scientific) and viability was assessed using 0.4% (w/v) Trypan Blue (Invitrogen, Carlsbad, CA, USA).

2.2.3. HaCaT and KHOS Cell Lines

The HaCaT cell line was kindly provided by VIVEX Biologics (Miami, FL, USA). The KHOS cell line was generously donated by Dr. Zhenfeng Duan’s Sarcoma Biology Lab in the Department of Orthopaedic Surgery at the University of Miami [37,38,39].

2.3. MIAMI Cells Characterization

To study the gene expression profiling of MIAMI cells, two pre-designed 90-gene human mesenchymal stem cell (hMSC)-qPCR arrays (STEMCELL Technologies) were used to evaluate 172 unique target genes associated with stemness, differentiation potential, and regenerative function. Three biological replicates of MIAMI cell samples were selected for analysis, following the isolation and culture procedures as previously described.

Total RNA was extracted using the RNeasy Plus Mini Kit (QIAGEN) in accordance with the manufacturer’s protocol. RNA concentration and purity were assessed using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific), with absorbance measured at 260/280 and 260/230 nm monitored to ensure sample integrity.

For cDNA synthesis, 1 µg of total RNA per sample was reverse transcribed using the SuperScript™ VILO™ cDNA Synthesis Kit (Thermo Fisher Scientific), following the recommended protocol. The resulting cDNA was used at 1 µg in two separate qPCR arrays: the Human Pluripotent Stem Cell Naïve State qPCR Array and the Human Mesenchymal Stem Cell qPCR Array (STEMCELL Technologies, Vancouver, Canada), as per the manufacturer’s instructions.

qPCR reactions were carried out on a StepOnePlus™ Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA) using SYBR Green chemistry and standard cycling conditions, including a melt curve step to verify amplification specificity. Data were analyzed using the STEMCELL Technologies Online qPCR Data Analysis Tool “https://shiny.stemcell.com/ShinyApps/qpcr_tool/” (accessed on 3 October 2025). Expression levels were calculated using the 2^−ΔCt^ method and normalized to the optimal housekeeping gene (HKG), selected as the endogenous control gene that exhibited the lowest fold change and most stable expression across the three biological replicates.

2.4. Extracellular Vesicle Isolation by Ultracentrifugation

2.4.1. Extracellular Vesicles Derived from MIAMI Cells (MIA-EVs)

MIAMI cells were seeded at low density into ten 175 cm^2^ flasks (Thermo Fisher Scientific) and cultured under low oxygen conditions for approximately 21 days, with partial medium changes (50%) performed every 3 to 4 days. Upon reaching 60% confluency, the culture medium was completely removed, and the flasks were rinsed twice with 1× Dulbecco’s Phosphate-Buffered Saline (DPBS; Corning, Corning, NY, USA) to eliminate residual serum-derived vesicles and cellular debris.

Fresh MIAMI medium was prepared using EVs-depleted FBS. EVs depletion was achieved by ultracentrifuging FBS at 120,000× g for 4 h at 4 °C using a Beckman Coulter Optima XPN (Beckman Coulter, Brea, CA, USA) ultracentrifuge equipped with the SW 40 Ti swinging-bucket rotor, following established ultracentrifugation conditions [25]. Cells were then incubated in this EVs-depleted medium for 48 h to allow EVs accumulation in the conditioned medium.

At the end of the incubation period, the conditioned medium (secretome) from all flasks was pooled and supplemented with protease inhibitor (4 µL per 10 mL of secretome). The pooled secretome was first centrifuged at 910× g for 10 min at 4 °C using a Eppendorf 5810R Centrifuge (Eppendorf, Hamburg, Germany) to remove cellular debris and large vesicles. The supernatant was then ultracentrifuged using the established ultracentrifugation conditions. The EVs-containing pellet was resuspended in DPBS and passed through a 0.22 µm filter to ensure sterility.

Following secretome collection, adherent cells were detached using 0.25% trypsin-EDTA (Gibco) and resuspended in fresh medium. Cell viability and concentration were assessed manually using trypan blue exclusion (Gibco) and a hemocytometer. Cells were then cryopreserved at a density of 5 × 10^5^ cells per cryovial in 1 mL of freezing medium composed of 80% FBS, 10% complete MIAMI medium, and 10% DMSO (Corning), and stored at −80 °C for future use.

Differential ultracentrifugation was selected for MIA-EVs isolation due to its widespread adoption, scalability, and ability to yield structurally intact vesicles with acceptable purity [2]. While alternative methods such as size exclusion chromatography, immunoaffinity capture, polymer-based precipitation, and microfluidics offer specific advantages, they present limitations related to yield, throughput, or potential bias in EVs subpopulation recovery. Ultracentrifugation avoids chemical- or affinity-based selection pressures, ensuring broader vesicle representation and compatibility with downstream functional and molecular assays. Its methodological transparency and prevalence in the field also facilitate reproducibility and cross-study comparison, aligning with the study’s translational goals and commitment to generating biologically meaningful data.

2.4.2. Extracellular Vesicles Derived from TIC-Primed MIAMI Cells (MIA-TIC-EVs)

MIAMI cells were cultured in MIAMI medium until they reached approximately 60% confluency. Cells were then washed twice with DPBS to remove residual serum, and the medium was replaced with EVs-depleted MIAMI medium supplemented with a cytokine cocktail containing 15 ng/mL TNFα, 10 ng/mL IFNγ, 10 ng/mL CTGF, collectively referred to as TIC priming factors. Cells were incubated under these conditions for 72 h to mimic a sustained inflammatory environment. This priming strategy is adapted from [40], which demonstrated that cytokine preconditioning enhances the therapeutic properties of MSCs by increasing their immunomodulatory and anabolic potential.

Secretome was collected after priming and processed for EVs isolation using the same differential ultracentrifugation workflow described above. The resulting EVs population is referred to throughout the manuscript as MIA-TIC-EVs. These EVs were used for miRNA qPCR profiling (Section 2.5.6) and functional assays (Section 2.6.4).

2.4.3. Extracellular Vesicles Derived from Irradiated MIAMI Cells (MIA-IR-EVs)

MIAMI cells were expanded in MIAMI medium until reaching approximately 60% confluency. Cells were washed twice with DPBS to remove residual serum, and the medium was replaced with EVs-depleted MIAMI medium prior to irradiation. A single 5 Gy dose was delivered at room temperature using an Xstrahl CIX3 cabinet X-ray irradiator (Xstrahl, Suwanee, GA, USA) [41]. Immediately after exposure, cultures were returned to the tri-gas incubator and maintained for 48 h. Secretome was then collected and processed for EVs isolation using the same differential ultracentrifugation workflow described above.

The resulting EVs population is referred to throughout the manuscript as MIA-IR-EVs. These EVs were used exclusively for miRNA qPCR profiling (Section 2.5.6).

2.4.4. Extracellular Vesicles Derived from Mesenchymal Stem/Stromal Cells (MSC-EVs)

MSCs were seeded into 175 cm^2^ flasks (Falcon, Corning, NY, USA) and cultured at 37 °C and 5% (v/v) CO_2_ in a humidified incubator for 15 days, with partial medium changes (50%) performed every 3–4 days. Once cultures reached approximately 80% confluency, the medium was removed, and cells were rinsed twice with DPBS to eliminate residual serum-derived vesicles and debris. Fresh MSC medium prepared with EVs-depleted FBS was then added, and cells were incubated for 48 h.

Secretome was collected and processed for EVs isolation using the same differential ultracentrifugation workflow described above. The resulting EVs population is referred to throughout the manuscript as MSC-EVs. These EVs were used exclusively for miRNA qPCR profiling (Section 2.5.6).

2.5. EVs Characterization

2.5.1. Nanoparticle Tracking Analysis (NTA)

The size distribution and concentration of MIA-EVs were analyzed using a NanoSight NS300 system (Malvern Panalytical) operated with NTA software version 3.4. Samples were diluted 1:10 in DPBS to a final volume of at least 1 mL and equilibrated to room temperature prior to measurement. The system was flushed with 2–3 mL of Water for Injection (WFI) before each run to minimize background noise. Samples were manually loaded using a syringe until particles were visible in the field of view.

Camera settings were standardized across all measurements, with the level set to 11 and gain to 1.0. For each sample, five videos of 15 s each were recorded at an infusion rate of approximately 50 µL/min. Manual focus was adjusted to ensure particles appeared sharp and without halos. Data were analyzed using default parameters: detection threshold set to 5, 30 frames captured per video, and a measurement time of 30 s per replicate. Only red-traced particles were included in the analysis. Settings were kept constant for all samples to ensure measurement consistency.

Results were reported as particle size distribution metrics, including mean, mode, and D-values (D10, D50, D90), and particle concentration in particles/mL. Each result represents the mean of five individual measurements per sample.

2.5.2. Transmission Electron Microscopy Imaging

To visualize the morphology and structural integrity of MIA-EVs, transmission electron microscopy (TEM) was performed. Samples were adsorbed onto carbon-coated copper grids for 30 min, rinsed in phosphate buffer, fixed with 2% glutaraldehyde, and washed with double-distilled water. Grids were then stained with 2% aqueous uranyl acetate for 5 min, protected from light, and stored overnight. Imaging was performed at 80 kV using a JEOL JEM-1400 transmission electron microscope (JEOL, Tokyo, Japan), and images were acquired using an AMT BioSprint 12 digital camera (Advanced Microscopy Techniques Corp., Woburn, MA, USA).

2.5.3. Flow Cytometry for Exosomes Markers

In accordance with MISEV2023, the generic term ‘EV’ is preferred, and the use of biogenesis-based terms such as ‘exosome’ and ‘ectosome’ is discouraged [2]. However, surface marker profiles can provide indirect insights into EVs origin: enrichment in CD9, CD81, and some CD63 typically indicates a mixed population of small EVs; predominant CD63 expression suggests exosome enrichment; and predominant CD9/CD81 with low CD63 suggests ectosome enrichment [42]. In this study, anti-CD63 immunoaffinity capture was employed, a method that preferentially isolates CD63-positive small EVs, consistent with an exosomes-enriched fraction.

MIA-EVs were characterized by flow cytometry to validate the presence of exosome-enriched particles based on CD9 expression in CD63-captured vesicles. Samples were first incubated overnight at 4 °C on a rotator with human CD63 magnetic beads (Invitrogen, Carlsbad, CA, USA), using 200 µL of beads per 1 mL of EVs suspended in DPBS. Following incubation, samples were transferred to 15 mL conical tubes, brought to a total volume of 4 mL with DPBS, and placed on a magnetic rack for 2 min. Supernatants were carefully removed without disturbing the beads, which were then resuspended in 200 µL of flow buffer (DPBS supplemented with 0.5% BSA and 2 mM EDTA) to prevent aggregation.

Samples were split into two tubes for immunostaining: one for antibody staining and one as unstained control. 5 µL of anti-human CD9-FITC antibody (Invitrogen) were added to 200 µL of bead suspension. After gentle mixing, tubes were incubated for 15 min in the dark inside a biosafety cabinet. Samples were washed twice with 5 mL of flow buffer, using magnetic separation to retain the beads after each wash. Beads were finally resuspended in 200 µL of flow buffer, and 150 µL per sample was transferred into a 96-well plate for acquisition.

Flow cytometry data were acquired using a CytoFLEX S cytometer (Beckman Coulter), collecting 20,000 events per sample. Data were analyzed using Kaluza Analysis Software v.2.4 (Beckman Coulter). Only CD9+ events within the CD63-gated population were considered indicative of exosome-enriched EVs.

2.5.4. Expression Levels of Exosomes Surface Markers

EVs were isolated from the MIAMI cell secretome via ultracentrifugation, as described above. The expression levels of the canonical multi-pass transmembrane proteins (tetraspanins) CD9, CD63, and CD81 were analyzed using the Leprechaun™ Exosome Human Tetraspanin Kit (Unchained Labs, Pleasanton, CA, USA), following the manufacturer’s Leprechaun Exosome Assay Protocol.

For each sample, 40 µL of EVs suspension was mixed with 40 µL of incubation solution, and 50 µL of the resulting mixture was loaded per Luni chip lane. Luni chips are coated with anti-CD9, anti-CD63, and anti-CD81 antibodies, spotted in technical triplicates per chip, with additional isotype control spots for the assessment of non-specific binding. Each chip surface is coated with SiO_2_ to enhance signal contrast for nanoparticle detection. All incubations were performed at room temperature for 1 h. Automated washing was carried out using the Luni Washer (Unchained Labs).

Detection was performed using the Leprechaun platform, which utilizes dual fluorescence imaging scans (background and sample) to quantify surface marker expressions. Fluorescent signals were captured in the following channels: CD9 (CF^®^ 488A, blue), CD81 (CF^®^ 555, green), and CD63 (CF^®^ 647, red). Data acquisition and analysis were conducted using the Leprechaun Client Software v.2.0.1, which computes signal intensity based on replicate spot analysis and subtracts background to determine relative tetraspanin abundance.

2.5.5. Protein Cargo Profiling

The total protein content of MIA-EVs was quantified using the Pierce BCA Protein Assay Kit (Thermo Scientific) according to the microplate protocol. Briefly, bovine serum albumin (BSA) standards (0–2000 µg/mL) and EVs samples (25 µL) were mixed with 200 µL of working reagent and incubated at 37 °C for 30 min. Absorbance was measured at 562 nm using a SpectraMax ABS Plus microplate reader (Molecular Devices, San Jose, CA, USA). Protein concentrations were determined by interpolation from a second-order polynomial standard curve (R^2^ ≥ 0.995) using GraphPad Prism v10.5.0.

Growth factor profiling was performed using the Human Growth Factor C-Series Antibody Array (RayBiotech, Peachtree Corners, GA, USA). For each membrane, 300 µg of EVs protein was incubated overnight at 4 °C with gentle rocking. Following washing and blocking (1 h, room temperature), membranes were sequentially incubated with biotin-conjugated detection antibody cocktail (2 h) and HRP-streptavidin (1 h), according to the manufacturer’s instructions. Chemiluminescent signal was captured using the LI-COR Odyssey M imaging system (30 s exposure, high-gain setting).

TIFF images were converted to 8-bit grayscale and analyzed in Fiji (ImageJ v1.54) using the Protein Array Analyzer plug-in. Arrays were rotated, cropped, and processed to automatically define circular regions of interest (ROIs), which were manually verified. Mean pixel intensities were averaged across duplicates, background-subtracted using local membrane intensity, and normalized to internal positive controls. Signals below the mean background were considered non-detectable. Normalized values were log_2_-transformed for visualization. Three independent samples were analyzed on separate membranes, and data are reported as mean signal intensity.

2.5.6. miRNA Cargo Profiling

To investigate the molecular cargo of EVs, a curated microRNA (miRNA) qPCR array approach was employed. Compared to RNA sequencing, qPCR-based arrays offer greater sensitivity and reproducibility for low-input RNA samples, such as those typically obtained from EVs [43]. This method, widely adopted in MSC-EVs research to identify regenerative and immunomodulatory miRNAs, was selected here to characterize the profile of all four EVs species.

The EVs miRNAs profile was assessed using a 166-miRNA hMSC-qPCR array (GeneCopoeia, Rockville, MD, USA). EVs from three biological replicates were concentrated using 50 mL AMICON Ultra-30 centrifugal filter units. Samples were adjusted to 15 mL with PBS and centrifuged at 4000× g for 30 min at 20 °C using a Beckman Coulter Allegra X-15R (Beckman Coulter). The retained volume (approximately 2 mL) was collected for RNA isolation.

RNA was extracted using the Total Exosome RNA and Protein Isolation Kit (Thermo Fisher Scientific), following the manufacturer’s protocol. RNA concentration and purity were determined using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific), using 2 µL of RNA per reading.

For cDNA synthesis, 100 ng of total EVs RNA was reverse transcribed using the All-in-One miRNA First-Strand cDNA Synthesis Kit (GeneCopoeia), following the miProfile™ miRNA PCR Arrays user manual. qPCR was performed using pre-designed 166-human MSC exosome miRNA arrays (GeneCopoeia), with 100 ng of synthesized cDNA per sample. Reactions were run on a StepOnePlus Real-Time PCR System (Applied Biosystems) using SYBR Green detection chemistry and manufacturer-recommended cycling conditions. Each plate included technical controls and HKG.

Ct values were exported, and data analysis was performed using the GeneCopoeia online qPCR Data Analysis System. Ct values above 34 and undetermined were excluded. Relative expression was calculated using the 2^−ΔCt^ method and normalized to internal controls.

2.5.7. Network and Pathway Analysis

To evaluate the functional significance of selectively abundant miRNAs, network and pathway analyses were conducted using the miRNet 2.0 platform “https://www.mirnet.ca” (accessed on 3 October 2025), with parameters set for ‘Homo sapiens’ and ‘exosomes’ as the tissue source to simulate the biological context of EVs delivery. Experimentally validated miRNA–target interactions were retrieved from miRTarBase v9.0. Because functional annotation in miRNet is constrained by the availability of experimentally validated targets, only miRNAs with supported target interactions were retained for network and enrichment analyses, providing a conservative but high-confidence representation of regulatory activity across all EVs conditions and comparisons. The resulting interaction networks were refined using a betweenness centrality filter (cutoff ≥ 2.0) to prioritize high-connectivity regulatory nodes, as described by [44].

Functional enrichment analysis was performed using multiple annotation resources, including the Kyoto Encyclopedia of Genes and Genomes (KEGG), Reactome, and Gene Ontology:Biological Process (GO:BP). KEGG and Reactome provided curated pathway-level insights into signaling, metabolic, and survival mechanisms, while GO:BP was used to identify biological processes enriched among target genes. Statistical significance was assessed via hypergeometric testing, and p-values were adjusted for multiple comparisons using Benjamini–Hochberg false discovery rate (FDR) correction.

To streamline interpretation, enriched outputs were manually curated to remove redundancy and retain biologically relevant categories. GO:BP terms were subsequently grouped into seven high-level functional classes to summarize dominant cellular processes (e.g., stress response, cell cycle, transcriptional regulation).

All network and enrichment analyses were performed using the version of miRNet maintained by the Xia Laboratory at McGill University and last updated on 15 August 2025. Final network visualizations were rendered using force-directed layouts to emphasize pathway connectivity and regulatory hierarchy. This analytical workflow was applied uniformly across all EVs conditions and comparisons, including MIA-, MIA-IR-, MIA-TIC-, and MSC-derived EVs datasets.

2.6. Functional Assessment of MIA-EVs

2.6.1. MIA-EVs Uptake

MIA-EVs cellular uptake was evaluated using a modified PKH dye-labeling protocol. Recipient cell membranes were stained with PKH67, nuclei with DAPI, and EVs were labeled with PKH26, allowing tri-color fluorescence imaging. HaCaT and KHOS cells were first labeled with PKH67 green-fluorescent membrane staining kit (Fluorescent Cell Linker Kits, Sigma-Aldrich, St. Louis, MO, USA), according to the manufacturer’s instructions. Cells were seeded in two wells per cell line in a 12-well plate at a density of 200,000 cells per well in 2 mL of the appropriate culture medium.

To label EVs with the PKH26 green-fluorescent membrane staining kit (Fluorescent Cell Linker Kits, Sigma-Aldrich), which is typically used for cell membrane staining, EVs were first incubated overnight at 4 °C with CD63 magnetic beads (Invitrogen) to isolate the CD63-positive, exosomes-enriched subpopulation. Following isolation, the bead-bound EVs were stained with PKH26, according to the manufacturer’s protocol. The labeled EVs, 6 × 10^7^ particles per well, were then added to HaCaT and KHOS cells pre-labeled with PKH67 and incubated for 48 h at 37 °C in a humidified incubator with 5% (v/v) CO_2_.

Following the 48 h incubation period, cells were stained with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen) to visualize nuclei. Briefly, the medium was removed, and cells were washed twice with DPBS to remove residues. Cells were then fixed with 500 μL of 4% paraformaldehyde (PFA) per well for 10 min at room temperature. After fixation, the PFA was removed, and cells were washed twice with 1 mL of DPBS to remove any remaining fixative. DAPI staining was performed by adding 500 μL of a working solution containing 1 μL of DAPI dye diluted in DPBS (1:500 dilution) to each well, followed by incubation for 10 min at room temperature. The staining solution was then removed, cells were washed once with 1 mL of DPBS to eliminate excess dye, and 1 mL of fresh DPBS was added to each well.

Fluorescence imaging was performed using a Leica DMi8 inverted fluorescence microscope equipped with a DFC365 FX CCD camera (Nikon, Tokyo, Japan) to evaluate MIA-EVs uptake. Images were acquired at 20× magnification under standard fluorescence settings. The use of CD63 magnetic beads prior to PKH26 staining enabled specific visualization of the exosomes-enriched EVs subpopulation. Dual labeling with PKH26 (EVs) and PKH67 (recipient cell membranes) allowed assessment of EVs–cell interaction, while DAPI staining enabled nuclear visualization. This three-channel imaging approach facilitated the identification of internalized EVs relative to cell membranes and nuclei, supporting qualitative assessment of EVs uptake and intracellular localization, providing descriptive insights without quantitative measurement.

2.6.2. Cytotoxicity and Cell Proliferation

The cytotoxic and proliferative effects of MIA-EVs were evaluated using the Cell Counting Kit-8 (CCK-8; Sigma-Aldrich) assay. Two cell lines were tested: HaCaT and KHOS. Cells were seeded in Falcon 96-well flat, clear-bottom plates (Corning) in 200 µL of complete medium per well. All plates were incubated in a humidified environment under standard culture conditions: 37 °C and 5% (v/v) CO_2_. Preliminary optimization established that a seeding density of 2000 cells per well was optimal for a 96 h proliferation window.

The CCK-8 assay was selected over traditional MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) due to its enhanced sensitivity, speed, and non-destructive readout. Unlike MTT, which requires a solubilization step to quantify insoluble formazan crystals, CCK-8 produces a water-soluble formazan dye that can be directly measured in the culture medium, reducing procedural variability and handling time. These benefits allowed reliable viability assessments across cell types without compromising sample integrity [45,46].

For each cell line, five 96-well plates were prepared to assess cell viability over time. Plates 1 and 2 were designated for the 0 h and 24 h time points and included only control wells (no EVs). Plates 3, 4, and 5 were used for the 48 h, 72 h, and 96 h time points and included triplicate wells for control (no EVs), and MIA-EVs treatments were performed at particle-to-cell ratios of 5 × 10^2^, 5 × 10^3^, and 5 × 10^4^ EVs per cell and repeated in two biological replicates. Due to the minimal volume of EVs added, dilution effects were considered negligible.

At each time point, the culture medium was replaced with 100 µL of fresh medium (DMEM for HaCaT or RPMI for KHOS), followed by the addition of 10 µL of CCK-8 reagent per well, in accordance with the manufacturer’s protocol. Plates were incubated for 2 h at 37 °C, and absorbance was measured at 450 nm using the SpectraMax ABS Plus plate reader (Molecular Devices). Blank wells containing only medium and CCK-8 were included on each plate to correct for background absorbance.

Statistical analysis of the assay data was performed using GraphPad Prism v10.4.2 (GraphPad Software, San Diego, CA, USA). A two-way analysis of variance (ANOVA) was used to evaluate the effects of treatment (four EVs doses) and time (five-time points: 0, 24, 48, 72, and 96 h), as well as their interaction, on cell proliferation. Technical triplicates were included for each condition. The model included both main effects and interaction terms and was run without matching (no repeated measures). Tukey’s multiple comparisons post hoc test was applied to identify significant differences between treatment groups at each time point. A p-value smaller than 0.05 was considered statistically significant. Data are presented as mean ± standard deviation (SD).

2.6.3. Cell Migration

A scratch assay was conducted to evaluate the effect of MIA-EVs on HaCaT cell migration and their role in wound healing. Cells were seeded at a density of 5 × 10^5^ cells per well in 6-well plates, with 2 mL of HaCaT medium per well, and incubated at 37 °C in a humidified atmosphere with 5% (v/v) CO_2_ until 90% confluency was reached (approximately 3 days). A uniform vertical scratch was created at the center of each well using a sterile 1000 µL pipette tip, guided by the plate lid to ensure straight alignment. Wells were rinsed 2–3 times with DPBS to remove detached cells, and fresh medium containing EVs-depleted FBS was added to all wells.

Experimental conditions were tested in technical triplicates, including untreated control (no EVs), MIA-EVs at 1 × 10^7^ particles/well, and MIA-EVs at 1 × 10^8^ particles/well. Images were acquired immediately after scratch formation (0 h) and subsequently every 24 h. Imaging was performed using a Leica inverted microscope at 4× magnification and continued until complete scratch closure was observed. The same field of view was consistently imaged over time by marking the position on the plate bottom.

Gap closure was quantified using ImageJ (NIH) by measuring the scratch area at each time point. The percentage of closure was calculated relative to the 0 h area. Data were plotted in GraphPad Prism v10.4.2 to compare cell migration dynamics across conditions. Results represent the mean ± standard deviation of technical triplicates from two independent biological replicates.

2.6.4. Macrophage Polarization

Macrophages can be stimulated to adopt distinct functional phenotypes, including classically activated M1 macrophages, alternatively activated M2 macrophages, and tumor-associated macrophages (TAM). To evaluate the effects of MIA-EVs on macrophage polarization, an in vitro assay was performed based on the protocol described by [20]. THP-1 monocytes (ATCC) were cultured in suspension at a density of 4 × 10^6^ cells per T-75 flask in RPMI 1640 medium (Cytiva, Marlborough, MA, USA) supplemented with 10% FBS (Avantor). Cells were maintained under 3D culture conditions for 48 h at 37 °C in a humidified atmosphere with 5% (v/v) CO_2_ (Figure 1).

Cells were subsequently collected, counted using a hemocytometer with trypan blue exclusion, and seeded into 24-well plates at a density of 5 × 10^4^ cells per well in M1 Macrophage Generation Medium DXF (PromoCell, Heidelberg, Germany), marking Day 0 of differentiation into immature M1 macrophages. On Day 2, without changing the medium, polarization into mature M1 macrophages was induced by the addition of phorbol 12-myristate 13-acetate (PMA; 100 ng/mL) and ionomycin (IO; 1 µg/mL). After 48 h (Day 4), macrophages were treated with EVs.

For EVs treatment, each well received 1 mL of fresh medium containing either no EVs (control), MIA-EVs, or MIA-TIC-EVs, with EVs doses normalized to the yield of 1 × 10^6^ parental MIAMI cells. Macrophages were incubated with EVs for 48 h (Day 6), after which cells were harvested by trypsinization as previously described.

Following cell harvest, total RNA was extracted to assess changes in macrophage polarization-associated gene expression. Cells were lysed in RLT Plus Buffer (QIAGEN), and RNA was isolated using the RNeasy Plus Mini Kit (QIAGEN, Venlo, Netherlands) according to the manufacturer’s instructions. RNA concentration and purity were determined using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific). Complementary DNA (cDNA) was synthesized using the SuperScript™ VILO™ cDNA Synthesis Kit (Thermo Fisher Scientific) following the manufacturer’s protocol in a C1000 Touch Thermal Cycler (Bio-Rad, Hercules, CA, USA).

Gene expression profiling was performed using the GeneQuery™ Human Macrophage Polarization Markers qPCR Array Kit (ScienCell, Carlsbad, CA, USA), which targets 40 established markers associated with macrophage polarization, including M1, M2, TAMs, activation, and resting state markers. Reactions were conducted in a 96-well format on a StepOnePlus™ Real-Time PCR System (Applied Biosystems) using 20 ng of input RNA per reaction and manufacturer-recommended thermal cycling parameters. Target gene expression was normalized to the HKG serving as an internal control, and a cycle threshold (Ct) cutoff of 34 was applied. Two independent MIA-EVs preparations were analyzed as biological replicates. An M1 macrophage group that did not receive EVs treatment served as external control for comparative analysis.

Data acquisition and analysis were performed using StepOne Software v2.3 (Applied Biosystems). Relative gene expression was calculated using the 2^−ΔΔCt^ method. Statistical comparisons between groups were conducted using GraphPad Prism v10.4.2.

2.7. Statistical Analysis

The data in this study were evaluated using two-way analysis of variance (ANOVA) or unpaired t-tests to identify group differences, with the analyses performed using GraphPad Prism v10.4.2. Statistical significance was defined as a p-value less than 0.05. All results in this paper are expressed as means ± standard deviations, based on a minimum of two independent experiments.

3. Results

3.1. Integrated Gene Expression Profiling Confirms the Primitive, Regenerative, and Immunomodulatory Identity of MIAMI Cells

MIAMI cells have been extensively characterized in previous studies, particularly for their secretory and functional properties [32,47,48,49,50,51]. Proteomic analyses have shown that MIAMI cells secrete high levels of cytokines and growth factors involved in tissue repair, angiogenesis, and immunomodulation [34]. Complementary transcriptomic studies under low oxygen culture conditions (3% O_2_ vs. 21% O_2_) further revealed upregulation of self-renewal and pluripotency-associated transcription factors, including OCT4, NANOG, and SOX2, underscoring their developmentally primitive state and regenerative capacity [33]. Despite this substantial body of work, a systematic, targeted gene expression profiling of MIAMI cells has not been previously reported.

The distribution of gene expression in MIAMI cells was highly asymmetric, with a small subset of genes showing markedly elevated transcript abundance (Figure 2). PDGFRB, CDX2, and TERT were among the most highly expressed genes, reflecting robust activation of mesenchymal, developmental, and telomerase-related pathways. Other top-expressed genes included MMP13, PROM1, IFNG, and MIXL1, further indicating an active transcriptional landscape associated with remodeling, stem/progenitor identity, and immune modulation. For visualization purposes, only genes with expression values above 10 (2^−ΔCt^) were displayed in the figure to improve clarity, while all quantified genes were included in the analysis.

Among the most highly expressed genes, PDGFRB, a transmembrane tyrosine kinase receptor, is a key regulator of mesenchymal proliferation, angiogenesis, and tissue repair, mediating fibroblast expansion and pericyte recruitment during wound healing [52]. CDX2 and MIXL1 are homeobox transcription factors that regulate early mesendodermal patterning and early lineage specification, contributing to developmental plasticity [53]. TERT, the catalytic subunit of telomerase, maintains telomere length and supports extended replication potential, a feature associated with stem cell self-renewal capacity [54]. MMP13 (collagenase 3) encodes a matrix metalloproteinase that mediates extracellular matrix degradation and remodeling and is upregulated during tissue repair and bone development [55]. PROM1 (CD133), a pentaspan transmembrane glycoprotein recognized as a stemness cell marker, reinforces the undifferentiated phenotype of MIAMI cells [56]. The cytokine IFNG (interferon gamma) suggests active or primed immunomodulatory signaling, as it plays a central role in macrophage activation, antigen presentation, and immune regulation [57].

A gradual decline in expression was observed across the remaining gene set, with a long tail of moderately-to-lowly expressed transcripts. Several canonical pluripotency genes, such as POU5F1 (OCT4), SOX2, and NANOG, were either absent or expressed at very low levels, consistent with the non-embryonic, adult stem cell phenotype of MIAMI cells [33]. However, expressions of LIN28A, DPPA5, GDF3, and ESRRB indicate partial retention of early developmental potential [58,59].

The presence of both mesenchymal (e.g., VIM, COL1A1, THY1) and immunomodulatory (e.g., LIF, IL6, TGFB1) genes highlights the dual regenerative and immunoregulatory phenotype of MIAMI cells [60,61,62,63]. In addition, expression of angiogenic (VEGFA) and osteogenic (RUNX2) markers further support their broad differentiation potential [64,65].

These results are consistent with previous reports of MIAMI cells exhibiting a transcriptional signature that overlaps with both hMSCs and hESCs [34]. The observed gene expression landscape confirms that MIAMI cells retain a unique molecular identity compatible with regenerative therapeutic applications.

3.2. Multimodal Structural and Molecular Characterization of MIA-EVs Reveals a Pro-Regenerative and Immunoregulatory Signature

The identity, purity, and structural integrity of MIA-EVs were evaluated using a multimodal characterization approach, including nanoparticle tracking analysis (NTA), transmission electron microscopy (TEM), flow cytometry, and quantitative surface marker profiling.

3.2.1. Nanoparticle Tracking Analysis Confirms Size Distribution and Particle Yield of MIA-EVs

NTA confirmed a homogeneous population of EVs with a mean size of 125.9 nm ± 28.7 nm and a mode of 108.4 nm (Figure 3A). The size distribution was consistent with small EVs, with D10, D50, and D90 values of 100.0 nm, 118.9 nm, and 156.8 nm, respectively. Particle concentration across three independent MIA-EVs preparations was 3.11 × 10^8^ particles per mL on average.

3.2.2. Transmission Electron Microscopy Validates Morphological Integrity and Structural Preservation

TEM confirmed the presence of spherical, membrane-bound EVs with diameters consistent with the expected size range (30–200 nm). The vesicles displayed intact lipid bilayer structures and minimal aggregation, indicating successful isolation and preservation (Figure 3B). The absence of cellular debris or contamination further supports the purity of EVs preparation.

3.2.3. Flow Cytometric Validation of EVs Identity and Purity

Bead-based flow cytometry confirmed the identity and purity of the isolated MIA-EVs. A high percentage of vesicles, 88% of gated events, were positive for the exosomes markers CD9 and CD63, indicating that the EVs preparations were highly enriched in exosome-like vesicles (Figure 3C). This surface marker expression validates the isolation protocol and confirms the molecular integrity of MIA-EVs, supporting their suitability for downstream functional assays.

3.2.4. Quantitative Profiling of Canonical Tetraspanin Markers

Quantitative analysis of exosomes surface tetraspanins (CD9, CD63, and CD81) was performed to confirm the vesicular identity and molecular consistency of the MIA-EVs preparations. All three markers were detected with high signal intensity across three independent biological replicates, demonstrating reproducible expression of canonical exosome markers. The mean concentrations of CD9-, CD63-, and CD81-positive EVs were 5.17 × 10^8^ ± 2.46 × 10^7^, 7.79 × 10^8^ ± 3.19 × 10^7^, and 3.37 × 10^8^ ± 7.45 × 10^7^ particles/mL, respectively (Figure 3D). These findings provide molecular-level validation of the EVs identity and complement the purity assessment by flow cytometry, together supporting the robustness of the isolation protocol and the reliability of MIA-EVs for functional applications.

3.2.5. Protein Cargo Profiling Reveals Enrichment of Regenerative and Angiogenic Growth Factors

Multiplex antibody array analysis of MIA-EVs detected 40 growth factors with distinct signal intensities (Figure 3E). The most abundant proteins were IGFBP-1 (1910 AU) and HGF (1869 AU), followed by IGF-II (989 AU), IGF-I (914 AU), EGF (805 AU), and VEGF-D (803 AU). These factors are involved in signaling pathways associated with cell proliferation, survival, and vascular development, indicating that MIA-EVs are enriched in proteins relevant to tissue repair and remodeling.

Additional factors with relatively elevated signals included M-CSF (634 AU) and its receptor M-CSF R (830 AU), suggesting potential relevance to macrophage regulation and myeloid lineage dynamics [66]. GDNF (803 AU) and the neurotrophins NT-3 and NT-4 were also detected, indicating the presence of trophic signaling components [67]. Moderate signal intensities were observed for PDGF-AA, PDGF-AB, SCF, and SCF R, which are implicated in stromal cell activation and hematopoietic support [68].

In contrast, TGF-α, TGF-β1, TGF-β2, TGF-β3, PDGF-Rb, and VEGF-A showed minimal signals (<20 AU), suggesting limited incorporation of ligands commonly associated with inhibitory or fibrotic responses. Overall, the protein cargo of MIA-EVs reflects selective enrichment of mitogenic and pro-survival factors, with reduced representation of profibrotic signaling molecules. This profile closely parallels the previously reported secretory phenotype of MIAMI cells, supporting the retention of specific parental cell signaling components within the EVs cargo [34].

3.2.6. miRNA Cargo Profiling Identifies Selective Enrichment of Regulatory Sequences

qPCR profiling identified 138 miRNAs in MIA-EVs, among which 19 were highly expressed, representing a targeted molecular cargo with potential regulatory significance (Figure 4A). The top three miRNAs, hsa-miR-4466 (384.14), hsa-miR-7975 (128.14), and hsa-miR-4454 (91.37), exhibited elevated expression relative to the endogenous control. These miRNAs were followed by miR-4516 (30.85), miR-6089 (25.00), and miR-4286 (16.69), suggesting non-random enrichment patterns. Additional moderately expressed miRNAs included miR-3665 (15.64), miR-301a-3p (8.31), miR-21-5p (5.85), and miR-4488 (5.24), several of which have been previously implicated in inflammation, tissue remodeling, and mesenchymal cell signaling. Lower abundance miRNAs such as miR-107 (2.87), miR-27b-3p (2.59), miR-99a-5p (2.49), and miR-4792 (1.96) were also consistently detected across replicates.

The overall miRNA expression profile reflects a non-random and selective cargo-loading mechanism in MIA-EVs, enriched with regulatory sequences known to modulate key biological processes. The presence of both well-characterized and less-studied miRNAs associated with inflammation, tissue remodeling, and cellular stress responses underscores the potential of MIA-EVs to influence recipient cell behavior. These findings provided the molecular basis for the subsequent network and pathway analysis, aimed at elucidating the broader functional implications of the identified miRNA repertoire.

3.2.7. Network and Pathway Enrichment Analysis Demonstrates Coordinated Survival, Regenerative, and Immunomodulatory Signaling

To evaluate the biological significance of the MIA-EVs miRNAs cargo, the 19 most abundant miRNAs were analyzed using miRNet 2.0, with target gene interactions retrieved from miRTarBase v9.0. Network topology analysis (Figure A1) revealed a well-organized interactome, with highly connected hub nodes corresponding to regulatory genes implicated in multiple cellular functions. This architecture suggests a coordinated regulatory potential of MIA-EVs miRNAs across diverse biological pathways.

Functional enrichment analysis of target genes (Figure 4B) revealed statistically significant associations with signaling pathways involved in cell survival, regeneration, immune modulation, and inflammatory resolution. The most highly represented category was RTK–PI3K–AKT–mTOR growth and survival signaling (393 targets), indicating strong convergence on pro-survival, pro-growth, and stress-adaptive pathways. This was closely followed by enrichment in growth factors and angiogenic signaling (388 targets), supporting a robust regenerative and pro-vascular profile. Additional high-impact survival-associated clusters included cell cycle and DNA damage response (349 targets), transcriptional, epigenetic, and RNA regulation (230 targets), and metabolic and mitochondrial homeostasis (183 targets), together indicating coordinated regulation of proliferation, stress tolerance, and bioenergetic stability.

Within the regenerative domain, target genes were highly enriched in developmental morphogen and differentiation signaling (235 targets), neurodevelopment, axon guidance, and neural plasticity (212 targets), extracellular matrix remodeling and tissue architecture (166 targets), and RTK-driven growth and proliferation signaling (164 targets). These findings support a multi-level regenerative program encompassing tissue patterning, structural remodeling, and coordinated growth signaling. Additional enrichment in signal integration and network coordination further suggests that MIA-EVs miRNAs are positioned to orchestrate complex regenerative signaling networks rather than acting through isolated pathways.

Immune-related enrichment revealed strong representation of TLR-mediated innate receptor signaling (108 targets), adaptive immune activation and lymphocyte signaling (68 targets), innate immune danger sensing (65 targets), immune signal integration (59 targets), and antigen processing, presentation, and phagocytosis (56 targets). These profiles indicate that MIA-EVs miRNAs target both innate and adaptive immune surveillance nodes. Additional enrichment in cytokine, JAK–STAT, and stress-activated MAPK signaling (39 targets), chemokine signaling and leukocyte trafficking (19 targets), interferon-mediated signaling (13 targets), and Fc receptor-mediated antibody-dependent immunity (9 targets) highlight a broad immunoregulatory capacity spanning immune activation, migration, and effector function.

Anti-inflammatory regulatory enrichment was primarily reflected through TGF-Beta/SMAD signaling (62 targets), innate immune dampening (17 targets), PPAR signaling (10 targets), and metabolic inflammation control (7 targets), consistent with coordinated suppression of excessive inflammatory signaling and restoration of tissue homeostasis. Importantly, the survival-associated cluster cellular stress, hypoxia, and senescence (155 targets) further support a cytoprotective role under hostile microenvironmental conditions, while enrichment of programmed cell death and regulated necrosis (86 targets) suggests simultaneous fine-tuning of apoptotic thresholds rather than indiscriminate survival signaling.

To further assess the functional impact of MIA-EVs miRNAs, GO:BP enrichment analysis was performed, and enriched terms were grouped into seven biological classes: Apoptosis and Cell Death Regulation, Cell Cycle and Proliferation, Differentiation and Development, Metabolism and Biosynthetic Process Regulation, Signal Transduction and Receptor Signaling, Stress and Damage Response, and Transcription and Gene Expression (Figure 4C).

The most enriched category was Metabolism and Biosynthetic Process Regulation (5587 genes), reflecting extensive modulation of metabolic, lipid, nucleotide, protein, and energy-generating pathways. This was followed by Signal Transduction and Receptor Signaling (4161 genes), indicating broad engagement of intracellular signaling networks governing immune responses, kinase cascades, second-messenger systems, and receptor-mediated communication.

Substantial enrichment was also observed in Differentiation and Development (2727 genes) and Stress and Damage Response (2186 genes), consistent with the involvement of MIA-EVs miRNAs in tissue remodeling, cellular maturation, and adaptive responses to environmental and inflammatory stress. The Transcription and Gene Expression category (1887 genes) further suggested widespread regulation of RNA processing, chromatin organization, and transcriptional control.

The Cell Cycle and Proliferation group (1779 genes) encompassed key regulators of DNA replication, checkpoint control, mitotic progression, and growth factor-responsive signaling, indicating that MIA-EV-associated miRNAs actively engage core mechanisms governing cellular expansion, self-renewal, and proliferative capacity.

Despite representing the smallest gene set, the Apoptosis and Cell Death Regulation category (552 genes) captured central components of the apoptotic machinery, including caspase activation, survival signaling, and mitochondrial death control, indicating a focused regulatory influence of MIA-EVs miRNAs on the balance between cell survival and programmed cell death.

Together, these integrated analyses demonstrate that the MIA-EVs miRNA cargo engages interconnected regulatory networks involved in stress adaptation, cell cycle control, lineage commitment, immune signaling, and apoptotic regulation. The coordinated distribution of enriched biological processes across these domains supports a unified regulatory landscape in which cell survival, regenerative capacity, immune modulation, and inflammatory control are functionally linked under the experimental conditions evaluated.

3.3. Functional Evaluation of MIA-EVs Demonstrates Selective Regenerative and Immunomodulatory Activity

3.3.1. Efficient Cellular Internalization of MIA-EVs in Non-Malignant and Malignant Cells

Fluorescence microscopy revealed robust internalization of MIA-EVs by both non-malignant HaCaT and malignant KHOS cells. Abundant red puncta (PKH26-labeled EVs) were observed within the cytoplasm of green-labeled cells, indicating successful endocytic uptake rather than surface binding. Strong red–green colocalization, along with DAPI-stained nuclei, confirmed intracellular localization and preserved cell integrity. These findings demonstrate the efficient uptake of MIA-EVs across diverse cell types, supporting their potential for intracellular delivery of therapeutic cargo. Representative images are shown in Figure 5A.

3.3.2. Selective Modulation of Cell Proliferation: Enhanced Keratinocyte Growth and Suppressed Osteosarcoma Expansion

Cell proliferation assays using the CCK-8 method revealed that MIA-EVs exhibited no cytotoxic effects on human keratinocytes (HaCaT) across all tested doses (5.0 × 10^2^, 5.0 × 10^3^, and 5.0 × 10^4^ particles per cell). HaCaT cell proliferation increased in a dose- and time-dependent manner over 96 h, indicating the biocompatibility and potential pro-regenerative effect of MIA-EVs on healthy epithelial cells (Figure 5B). In contrast, MIA-EVs exerted a dose-dependent inhibitory effect on the proliferation of human osteosarcoma (KHOS) cells. Lower doses (5.0 × 10^2^ and 5.0 × 10^3^ particles per cell) induced only modest growth reduction, whereas the highest dose (5.0 × 10^4^ particles per cell) significantly suppressed cell proliferation (Figure 5C). The lack of divergence between groups at early time points (0–24 h) likely reflects the time required for EVs internalization and downstream functional effects to influence cell proliferation. These results indicate that MIA-EVs selectively impair osteosarcoma cell viability while supporting healthy cell growth, underscoring their potential as a targeted, cell-free therapeutic approach in EVs-based cancer treatment. Data is presented as mean ± SD from three technical replicates across three independent biological experiments.

These observations were supported by two-way ANOVA, which revealed a significant interaction between treatment and time in KHOS cells (p < 0.0001), indicating that the inhibitory effects of MIA-EVs on proliferation evolved and were dose-dependent. EVs treatment was the dominant source of variation (77.4%), with both treatment and time contributing significantly to overall growth dynamics. In contrast, HaCaT cells showed no significant interaction between treatment and time (p = 0.9907), and minimal variation was attributed to time (0.21%), confirming that cell proliferation remained stable over the 96 h period; although the main effect of treatment was statistically significant (p = 0.0392), accounting for 21.68% of the total variance. This indicates that differences in proliferation were primarily driven by treatment conditions, with minimal influence from temporal factors. Together, these findings demonstrate that MIA-EVs selectively inhibit cancer cell proliferation while supporting non-malignant cell growth, with treatment effects highlighting their dose-dependent biological activity.

3.3.3. MIA-EVs Accelerate Keratinocyte Migration and Wound Closure

Treatment with MIA-EVs significantly enhanced wound closure in HaCaT cells compared to the untreated control. As shown in Figure 5D, the wound area progressively decreased over time in both groups; however, EVs-treated cells exhibited a significantly faster healing trajectory.

At 48 h, a critical time point associated with maximal EVs uptake, the wound area in the MIA-EVs group was reduced by approximately 79% from the baseline (0.174 mm^2^), compared to only 54% in the control group (0.397 mm^2^). By 96 h, wounds in the EVs-treated group had nearly closed completely (0.010 mm^2^), while control wells retained a visible gap (0.139 mm^2^).

These findings are supported by representative images showing the scratch contour at each time point, with enhanced wound closure clearly visible in the EVs-treated group beginning at 48 h. The observed effect highlights the regenerative potential of MIA-EVs and their capacity to stimulate keratinocyte migration and proliferation.

3.3.4. Targeted Immunomodulation: Selective IL1R2 Upregulation Without Induction of a TAM-Like Phenotype

Gene expression profiling revealed that treatment with MIA-EVs induced a selective transcriptional shift from M1- to M2-polarized macrophages. Among 37 polarization-related genes analyzed, IL1R2 (a key M2-associated immunosuppressive marker) was the only gene showing a strong and consistent upregulation, with an average 9.55-fold increase (ΔΔCt method) across two biological replicates compared to untreated M1 controls (Figure 6).

IL1R2 encodes the decoy receptor for IL-1 [69], and its upregulation suggests a shift toward an anti-inflammatory phenotype. In contrast, the expression of canonical pro-inflammatory M1 genes such as IL1B (0.017-fold), TNF (0.041-fold), and NOS2 (0.048-fold) was markedly suppressed relative to controls. Similarly, other M2 markers, including CD163 (0.051-fold), IL10 (0.057-fold), and MRC1 (0.030-fold), remained low, underscoring the specificity of IL1R2 induction.

These findings demonstrate that MIA-EVs do not broadly induce M2 polarization but instead selectively modulate macrophage behavior by upregulating IL1R2 expression within a stable M1 transcriptional background. This targeted immunomodulatory effect suggests a precise mechanism by which MIA-EVs may reprogram inflammatory macrophages toward a pro-resolving phenotype, offering therapeutic potential for regenerative medicine and inflammation-driven tissue repair.

To determine whether MIA-EVs induced a tumor-associated macrophage (TAM)-like phenotype, the expression of established TAM markers, including TGFB1, CD68, CD163, NOS2, HIF1A, CCL2, and PECAM1, was evaluated. These markers reflect the immunosuppressive and pro-angiogenic features of TAMs commonly observed in the tumor microenvironment and chronic tissue remodeling contexts. All genes exhibited negligible fold changes, with average 2^−ΔΔCt^ values below 0.1, indicating minimal induction of TAM-associated transcriptional factors, suggesting that MIA-EVs do not promote a pro-tumoral or immunosuppressive TAM-like state under these conditions. Gene expression values are presented as 2^−ΔΔCt^ fold changes relative to untreated M1-polarized macrophages. Data represents the average of two biological replicates. Values indicate relative fold changes compared to the MIA-EVs-untreated M1 control group.

3.4. Stimulus-Dependent Remodeling of MIA-EVs Reveals a Conserved Regulatory Core and Context-Specific Functional Modulation

3.4.1. Inflammatory Priming and Irradiation Induce Distinct Yet Overlapping miRNA Cargo Signatures

qPCR profiling of MIA-EVs derived from primed MIAMI cells revealed distinct miRNA signatures shaped by inflammatory priming or irradiation (Figure 7A,B). MIA-IR-EVs contained 113 detectable miRNAs, of which 25 were highly expressed, whereas MIA-TIC-EVs exhibited 99 detectable miRNAs, including 35 which were highly expressed. Compared with MIA-EVs, both primed conditions displayed unique enrichment as well as a partially overlapping profile, indicating condition-specific modulation of EVs cargo loading.

In MIA-IR-EVs, the most abundant miRNA was hsa-miR-136-5p, exhibiting markedly elevated expression (54,624.03), followed by strong enrichment of miR-4466 (188.97), miR-3665 (71.65), miR-7975 (63.35), miR-6089 (52.96), and miR-4454 (51.60). Additional moderately expressed miRNAs included miR-154-5p (34.13), miR-4516 (13.29), miR-4286 (10.15), miR-301a-3p (7.82), miR-4488 (7.26), and miR-107 (3.07), among others. Several of these miRNAs have established roles in DNA-damage responses, oxidative stress, and cellular repair, consistent with expected irradiation-induced transcriptional reprogramming [70,71,72,73,74].

In contrast, MIA-TIC-EVs displayed a priming-specific miRNA profile dominated by hsa-miR-140-5p (1397.19), a regulator associated with tissue remodeling and inflammatory signaling. Other highly expressed miRNAs included miR-301a-3p (87.37), miR-107 (77.82), miR-4454 (63.95), miR-4466 (44.66), miR-7975 (39.91), miR-125a (38.73), and miR-378g (23.16). Additional enriched miRNAs such as miR-27b-3p (14.91), miR-182-3p (9.88), miR-193b-5p (8.01), let-7e-5p (7.94), miR-4286 (7.88), and miR-98-5p (5.36) were also consistently detected. These miRNAs have been implicated in immunomodulation, cytokine responsiveness, and macrophage polarization, aligning with the biological intent of TIC priming [75,76,77].

Collectively, the primed-EVs groups showed both unique and overlapping miRNA cargo, suggesting that MIAMI cells can selectively modify EVs cargo in response to external stimuli. Such stimulus-dependent loading likely reflects adaptive transcriptional responses triggered by either inflammatory cytokines or irradiation. These results prompted further comparative analyses, including Venn mapping of shared and exclusive miRNAs across the three MIAMI-derived EVs conditions and downstream pathway enrichment to determine the biological functions represented by these miRNA profiles.

Comparison of miRNA profiles across baseline and primed MIAMI-derived EVs revealed both broad overlap and condition-specific signatures (Figure 7C). The majority of detected miRNAs—86 species (57.7%)—were shared among MIA-EVs, MIA-IR-EVs, and MIA-TIC-EVs, representing a conserved vesicular cargo. These shared miRNAs include regulatory sequences involved in stemness, stress responses, metabolic regulation, and immunomodulation, suggesting a core MIAMI-derived EVs identity [78,79].

Beyond this conserved backbone, each condition displayed unique enrichment patterns. MIA-EVs contained 25 miRNAs (16.8%) not detected in either primed EVs group, indicating a baseline regulatory profile that is partially diminished under inflammatory or irradiation stress. MIA-IR-EVs exhibited eight unique miRNAs (5.4%), several of which (e.g., miR-136-5p) are linked to DNA-damage responses and apoptotic regulation, consistent with irradiation-induced cellular reprogramming. In contrast, MIA-TIC-EVs showed a single uniquely detected miRNA (0.7%), indicating a highly selective but biologically meaningful cytokine-driven shift in EVs cargo [80].

Pairwise overlaps further highlighted condition-specific divergence. MIA-EVs and MIA-IR-EVs shared 17 miRNAs (11.4%), whereas MIA-EVs and MIA-TIC-EVs shared 10 miRNAs (6.7%). Only two miRNAs (1.3%) were shared exclusively between MIA-IR-EVs and MIA-TIC-EVs, underscoring that inflammatory priming and irradiation elicit distinct transcriptional responses despite both representing cellular stress conditions.

Together, these findings demonstrate that MIAMI cells modulate their EVs miRNA cargo in a stimulus-dependent manner, generating both a stable core signature and context-specific enrichment patterns. This distribution of shared and exclusive miRNAs informed the subsequent pathway analysis, aimed at elucidating their functional implications.

3.4.2. Conditioning-Resistant 15-miRNA Core Preserves Regenerative, Survival, and Immunomodulatory Network Architecture

To define the conserved regulatory architecture underlying the miRNA cargo of MIA-, MIA-IR-, and MIA-TIC-EVs, the analysis was restricted to the most biologically meaningful fraction of the dataset. Although 86 miRNAs were shared across all three EVs populations, only 15 fulfilled the dual criteria of (i) presence in all conditions and (ii) inclusion within the respective high-abundance fractions. This conservative core—miR-4466, miR-7975, miR-4454, miR-4516, miR-6089, miR-4286, miR-3665, miR-301a-3p, miR-21-5p, miR-4488, miR-107, miR-27b-3p, miR-3065-5p, miR-378g, and miR-98-5p—represents a stable, conditioning-resistant regulatory fingerprint intrinsic to MIAMI-derived EVs. Importantly, this core persists despite exposure to irradiation-induced stress or inflammatory cytokine priming, indicating that these miRNAs reflect foundational regulatory programs rather than adaptive- or transient stress responses.

Mapping the conserved 15-miRNA module onto experimentally validated targets reveals convergence onto a defined set of regulatory nodes controlling survival, stress adaptation, immune sensing, and tissue remodeling. Rather than dispersing across unrelated processes, these targets cluster into interconnected control points that coordinate growth factor responsiveness, checkpoint regulation, inflammatory sensing, and metabolic adaptation. This indicates that the shared MIA-EVs miRNA core is structured for coordinated regulation of cellular state transitions, specifically the balance between quiescence, damage response, immune activation, and regenerative re-entry into the cell cycle. The preservation of these regulatory axes across naïve, irradiated, and cytokine-primed EVs further indicates that MIAMI conditioning modulates signal intensity rather than network identity, maintaining a stable regulatory backbone that ensures functional continuity under environmental stress.

When examined across all three EVs states (Figure 7D), the miRNA regulatory pathway remains functionally preserved despite irradiation and inflammatory priming. Regenerative signaling is prominently represented through FGFR, VEGF, Wnt, Notch, neurotrophin, BMP, PDGF, Tie2, adhesion/Rho GTPase, axon guidance, developmental biology, ECM remodeling, and MAPK pathways, closely mirroring the broader MIA-EVs enrichment in growth factor and angiogenic signaling, developmental morphogen and differentiation programs, neurodevelopment/axon guidance, and extracellular matrix remodeling.

Survival and stress-adaptive regulation are likewise preserved. Enrichment of growth factor survival signaling, PI3K/AKT, mTOR, p53 checkpoint control, cellular stress and hypoxia/HIF signaling, apoptosis, senescence, and regulated necrosis in conditioned EVs recapitulates the dominant MIA-EVs survival modules centered on RTK–PI3K–AKT–mTOR signaling, cell cycle and DNA damage control, cellular stress, hypoxia and senescence, and programmed cell death. This indicates stable regulation of checkpoint control, metabolic stress tolerance, and apoptotic thresholding across conditions.

Immune and inflammatory regulatory programs remain consistently targeted across EVs states. The shared miRNA core regulates JAK–STAT/interleukin, TLR, NLR/inflammasome, BCR, DAP12, Fc receptor, chemokine signaling, antigen processing and presentation, leukocyte migration, NK cytotoxic, TCR, C-type lectin signaling, and phagosome pathways, corresponding directly to the MIA-EVs enrichment in TLR-mediated innate signaling, adaptive lymphocyte activation, antigen processing and phagocytosis, cytokine/JAK–STAT signaling, chemokine-driven leukocyte trafficking, Fc receptor-mediated immunity, and immune signal integration.

Finally, the persistent representation of TGF-Beta/SMAD, cellular stress and hypoxia/HIF, innate immune dampening, and PPAR signaling aligns with the anti-inflammatory and metabolic inflammation control signatures observed in MIA-EVs. Together, these patterns indicate that conditioning predominantly modulates the relative weighting of pre-existing regulatory pathways rather than introducing new signaling axes.

GO:BP enrichment analysis of the combined MIA-, MIA-IR-, and MIA-TIC-EVs miRNA targets (Figure 7E) revealed a functional hierarchy dominated by Metabolism and Biosynthetic Process Regulation (6006 genes), indicating that metabolic and biosynthetic control represents the primary regulatory axis across basal, irradiated, and inflammatory priming conditions. The increased gene magnitude within this category relative to MIA-EVs alone suggests an expansion of pathways governing lipid remodeling, energy production, and macromolecular biosynthesis under stress and cytokine conditioning, consistent with enhanced bioenergetic and anabolic demands during cellular adaptation.

In this combined dataset, Differentiation and Development (1685 genes) ranked second, followed by Stress and Damage Response (1424 genes) and Cell Cycle and Proliferation (1343 genes). This reorganization of functional dominance reflects a relative shift toward pathways regulating tissue remodeling, stress resilience, and controlled proliferative capacity in response to irradiation and inflammatory priming.

By contrast, Signal Transduction and Receptor Signaling displayed reduced relative representation in the combined analysis (996 genes), and Transcription and Gene Expression was further condensed (755 genes), suggesting that priming narrows upstream signaling and transcriptional breadth in favor of downstream effector and metabolic programs. Apoptosis and Cell Death Regulation, although the smallest class (375 genes), remained selectively enriched for core survival and apoptotic control pathways, indicating preserved regulation of cell fate under primed conditions. Collectively, these shifts define a priming-dependent remodeling of the MIA-EVs regulatory landscape toward metabolic reinforcement, stress adaptation, survival control, and regeneration-oriented programs.

Overall, this multi-layered network and pathway analysis demonstrates that MIA-derived EVs are governed by a conditioning-resistant miRNA regulatory backbone optimized for coordinated control of cell survival, immune modulation, and regenerative readiness. Rather than encoding condition-specific signaling programs, this architecture preserves a stable regulatory framework that enables recipient cells to tolerate injury, resolve inflammatory stress, and re-enter productive repair programs without destabilizing cell fate. This systems-level organization provides a direct mechanistic basis for the functional phenotypes observed herein, including enhanced wound closure, sustained proliferative competence under irradiation, and controlled macrophage inflammatory modulation. Together, these findings position MIA-EVs not as passive carriers of trophic signals, but as active regulators of tissue-state transitions in hostile microenvironments, supporting their translational relevance for regenerative and post-injury therapeutic applications.

3.4.3. Enhanced Immunoregulatory Polarization by MIA-TIC-EVs Without Induction of a TAM-Like Phenotype

Macrophages treated with MIA-TIC-EVs exhibited increased expression of genes associated with M2 and M2-like phenotypes (Figure 8A). Among the M2-associated genes, PPARG displayed the highest relative expression (56.16, normalized to HKG). PPARG encodes a nuclear transcription factor involved in lipid metabolism, glucose homeostasis, and anti-inflammatory regulation, and has been implicated in driving M2 macrophage polarization through activation of tissue-repair programs [81]. Similarly, the M2-like transcription factor STAT1 was markedly upregulated (16,647.22), consistent with its role in immune modulation and remodeling in response to external cues [82]. These expression levels exceeded those observed in macrophages treated with MIA-EVs (Figure 8B), indicating enhanced transcriptional responses following exposure to TIC-derived vesicles.

While M2 and M2-like genes were predominant, a subset of M1-associated genes was also upregulated. CD86 (10.76), a co-stimulatory molecule involved in T cell activation and antigen presentation [83], along with IL23A (5.98) and IL1B (5.79), were among the most induced M1 transcripts. However, the majority of other classical M1 genes, including IL6, TNF, and NOS2, remained below 1-fold expression. This profile suggests that MIA-TIC-EVs preferentially activate anti-inflammatory macrophage polarization while minimally expressing pro-inflammatory genes.

Expression analysis of TAM markers (Figure 8B) demonstrated that both MIA-EVs and MIA-TIC-EVs induced low transcript levels of TGFB1, CD68, CD163, NOS2, HIF1A, and CCL2 in polarized macrophages. MIA-TIC-EVs resulted in a modest increase relative to MIA-EVs; however, overall expression remained negligible. The slight upregulation of TGFB1 and CCL2 may reflect early or context-specific modulation rather than commitment to a pro-remodeling or immunosuppressive phenotype. This expression profile underscores the restrained immunomodulatory impact of both EVs treatments, reinforcing the therapeutic safety of MIA-EVs and indicating that even under inflammatory priming, MIA-TIC-EVs do not induce TAM-associated gene programs. This low-level engagement is favorable in translational applications where maintaining immune homeostasis is critical, including tissue repair following oncologic resection or radiation injury.

Gene expression values are presented as 2^−ΔΔCt^ fold changes relative to untreated M1-polarized macrophages. Data represents the average of two biological replicates. Values indicate relative fold changes compared to the MIA-EVs-untreated M1 control group.

3.5. Comparative Molecular Profiling of MSC-EVs Reveals a Conserved Mesenchymal Core with Reduced Regulatory Amplitude Relative to MIA-EVs

3.5.1. miRNA Cargo Profiling Identifies a Shared Mesenchymal Signature with Lower Quantitative Enrichment

qPCR profiling of MSC-EVs identified 153 detectable miRNAs, including 12 highly expressed species (Figure 9A). The most abundant MSC-EVs miRNAs—miR-7975 (2^−ΔCt^ = 40.34), miR-4454 (25.17), and miR-4466 (6.96)—overlapped with the highest-ranking miRNAs identified in MIA-EVs, suggesting the presence of a shared mesenchymal transcriptional footprint across both vesicle types. Additional enriched MSC-EVs miRNAs, including miR-4516, miR-575, miR-6089, miR-3665, and miR-4286 were also among the highly expressed MIA-EVs subset. Despite this shared identity, the magnitude of enrichment differed: miRNAs that reached high abundance in MIA-EVs (e.g., miR-4466 at 384.14) were present at lower levels in MSC-EVs. This divergence in dynamic range suggests that the naïve transcriptional state of MIAMI cells enhances selective miRNA loading into EVs, thereby increasing the effective regulatory dose delivered to recipient cells. Functionally, this quantitative amplification is consistent with the enhanced wound closure, sustained proliferative competence under irradiation, and macrophage polarization observed following MIA-EVs treatment. While this study focuses on the properties of MIA-EVs as a distinct EVs source, the observed quantitative differences in shared miRNA abundance relative to MSC-EVs indicate that cell origin alone can substantially influence vesicular miRNA loading, even in the presence of overlapping regulatory species.

A direct comparison between the two EVs sources provided additional insight into how cellular identity shapes EVs cargo composition (Figure 9B). MSC- and MIA-EVs shared 136 miRNAs (87.7%), representing a robust mesenchymal core enriched in regulators of immunomodulation, metabolic adaptation, cytoskeletal remodeling, and stress signaling. However, MSC-EVs contained 17 unique miRNAs (11.0%), whereas MIA-EVs contained 2 exclusive miRNAs (1.3%). The broader MSC-specific miRNA subset is consistent with the more differentiated and lineage-committed transcriptional state of MSCs, whose EVs cargo often mirrors stem/stromal specialization and extracellular matrix-associated functions [84,85]. In contrast, MIAMI cells possess a more naïve and developmentally plastic phenotype, positioned between stem/stromal progenitors and pluripotent cells [32]. This difference is evident in the EVs miRNA landscape, in which MIA-EVs exhibit reduced reliance on lineage-restricted signatures and instead display enrichment in miRNAs associated with adaptability, reparative programming, and environmental responsiveness.

Importantly, the limited number of MIAMI-exclusive miRNAs does not imply a reduced regulatory repertoire; rather, the greater expression amplitude of several shared miRNAs suggests that MIAMI cells preferentially amplify a select subset of highly active regulators species involved in cellular resilience, stress tolerance, and immune crosstalk. This quantitative feature is relevant because EVs-mediated biological effects depend not only on the presence of specific miRNAs but also on their vesicular dosage and target engagement efficiency. Accordingly, the higher dynamic range observed in MIA-EVs may support a broader or more adaptable paracrine influence compared with MSC-EVs, a hypothesis that is directly tested and supported by the functional analyses presented in this study.

Overall, while MIAMI and MSCs secrete EVs with largely overlapping mesenchymal miRNA profiles, the primitive state of MIAMI cells confers distinct quantitative features to their EVs cargo, characterized by amplified enrichment of select regulatory miRNAs and reduced reliance on lineage-restricted signatures. This quantitative modulation of a shared mesenchymal core provides a mechanistic basis for the subsequent network and pathway analyses and supports the rationale for investigating MIA-EVs as a novel, mechanistically distinct EVs therapeutic platform.

3.5.2. Network and Pathway Analysis Defines a Minimal Mesenchymal Regulatory Core Compared to the Expanded MIA-EVs Architecture

To define the shared regulatory architecture underlying the miRNA cargo of MIA- and MSC-EVs, the analysis was restricted to the most biologically meaningful fraction of the dataset. Among the 136 miRNAs common to both EVs populations, only those meeting the dual criteria of (i) presence in both sources and (ii) inclusion within the respective high-abundance subsets (19 in MIA-EVs and 12 in MSC-EVs) were retained. This filtering identified a conservative core of nine shared miRNAs—miR-4466, miR-7975, miR-4454, miR-4516, miR-6089, miR-4286, miR-3665, miR-301a-3p, and miR-107. This subset defines a shared mesenchymal regulatory signature that is conserved across both EVs sources, indicating that these miRNAs encode core regulatory pathways rather than source-specific or lineage-restricted signals.

Network analysis of this shared 9-miRNA core segregates into three dominant functional axes—regeneration, cell survival, and immune modulation—(Figure 9C), while anti-inflammatory signaling does not emerge as a major independently structured module, in contrast to the MIA-EVs network (Figure 4C). Regenerative signaling represented the largest category, led by FGFR signaling (53 target genes), followed by MAPK signaling (32 genes), axon guidance and neuronal plasticity (26 genes), adhesion and Rho GTPase signaling (20 genes), the developmental regulators NOTCH signaling (12 genes), and Wnt signaling (12 genes). These pathways define a conserved mesenchymal program of proliferation, migration, and morphogenesis.

Cell survival pathways formed the second dominant axis, with enrichment in cell cycle and DNA damage control (35 genes), growth factor-mediated survival signaling (20 genes), cellular senescence (19 genes), apoptosis (14 genes), checkpoint surveillance through p53 checkpoint control (seven genes), and PI3K/AKT signaling (seven genes). These same modules constitute the dominant backbone of the full MIA-EVs network but are represented there at substantially higher expression levels and broader regulatory coverage, indicating that cytoprotection is a conserved mesenchymal EVs function that is more strongly expressed in MIA-EVs.

Immune regulation within the shared MIA–MSC-EVs network is dominated by innate immune activation centered on TLR signaling, with secondary contributions from DAP12, Fc receptor, antigen presentation, JAK–STAT, and chemokine signaling. In contrast, anti-inflammatory regulation is confined to a discrete module composed of innate immune dampening and TGF-Beta/SMAD signaling, indicating that immune suppression is present but subordinate to immune activation within the shared mesenchymal core. Relative to this conserved regulatory minimum, the full MIA-EVs network exhibits a broader and more complex organization across regenerative, survival, and immune axes, with expanded signal integration and pathway diversity, consistent with the heightened functional breadth observed for MIA-EVs.

Collectively, MIA- and MSC-EVs share a tightly conserved mesenchymal miRNA core governing regeneration, survival, and immune calibration, whereas MIA-EVs uniquely expand the depth, connectivity, and regulatory range of this architecture. This network amplification provides a mechanistic basis for the enhanced wound repair, radiation tolerance, and macrophage reprogramming observed following MIA-EVs treatment.

GO:BP enrichment analysis of the shared MIA–MSC-EVs miRNA target network (Figure 9D) reveals a functional hierarchy dominated by Metabolism and Biosynthetic Process Regulation (827 genes), indicating that metabolic and anabolic control represents the principal conserved biological axis across both mesenchymal EVs sources. Although this category is likewise dominant in MIA-EVs alone (5587 genes), its markedly reduced magnitude in the shared MIA–MSC-EVs dataset indicates that only a restricted metabolic core is conserved between sources, while MIA-EVs uniquely expand metabolic regulatory breadth.

Differentiation and Development emerged as the second most represented category in the shared network (453 genes), consistent with preservation of mesenchymal programs governing morphogenesis, migration, and lineage plasticity. However, this remains substantially contracted relative to MIA-EVs alone (2727 genes), indicating that MIA-EVs uniquely extend developmental and reparative signaling beyond the conserved mesenchymal baseline.

In contrast to the MIA-EVs-only dataset, where Signal Transduction and Receptor Signaling represents a dominant regulatory axis (4161 genes), this category is reduced to 313 genes in the shared MIA–MSC-EVs core. This contraction indicates that broad upstream signal integration is not a defining feature of the conserved mesenchymal EVs program but instead represents a distinguishing property of MIA-EVs. Similarly, Stress and Damage Response (253 genes) and Cell Cycle and Proliferation (194 genes) are preserved only at minimal basal levels relative to their extensive representation in MIA-EVs alone (2186 and 1779 genes, respectively), indicating that robust stress adaptation and proliferative control are selectively amplified in MIA-EVs rather than universally shared across mesenchymal EVs.

Transcription and Gene Expression (118 genes) remains modestly represented, consistent with a conserved but limited transcriptional regulatory backbone. Finally, Apoptosis and Cell Death Regulation, while the smallest category in the shared dataset (43 genes), remains consistently represented relative to the MIA-EVs baseline (552 genes), indicating that core apoptotic and survival checkpoint regulation is a conserved mesenchymal EVs function, though greatly expanded in MIA-EVs.

Together, this GO:BP analysis demonstrates that the shared MIA–MSC-EVs miRNA core preserves a minimal mesenchymal regulatory framework centered on basal metabolism, developmental competence, and survival maintenance, whereas MIA-EVs uniquely elaborate this framework into higher-order programs of stress resilience, proliferative control, signal integration, and regenerative specialization. This hierarchical expansion provides a functional basis for the enhanced plasticity, radio-resistance, and immune-reprogramming capacity observed for MIA-EVs relative to conventional MSC-EVs.

4. Discussion

This study provides the first comprehensive characterization of MIA-EVs, establishing their unique molecular identity and therapeutic potential in tissue regeneration, modulation of inflammatory responses, and oncology, where they demonstrate inhibitory effects on cancer cell proliferation. Designed not only to test their biological function but also to establish a robust and reproducible production and characterization pipeline, our strategy strictly adhered to the Minimal Information for Studies of Extracellular Vesicles (MISEV 2023) guidelines [2]. We employed orthogonal techniques (NTA, TEM, and surface marker profiling) to verify isolated vesicle identity and purity before linking structural integrity to functional outcomes.

Functional assays were chosen to model regenerative and immune-related processes. The cell migration assay enabled quantifiable analysis of keratinocyte migration, an essential step in wound healing. While not as physiologically complex as 3D or organotypic models, the 2D monolayer system enabled controlled, quantifiable comparisons across conditions. EVs uptake, confirmed via CD63 bead and PKH26 labeling, validated that observed cellular changes were directly mediated by MIA-EVs internalization. Macrophage polarization was evaluated using a targeted qPCR panel featuring 40 markers across M1, M2-like, M2, and TAM states, enabling a detailed transcriptomic analysis. The significant induction of IL1R2 and concurrent suppression of IL1B and TNF underscores MIA-EVs’ capacity to promote a pro-resolving M2 phenotype and modulate inflammatory environments, an insight beyond standard surface marker assays.

Effective uptake of MIA-EVs was consistently observed in all tested cell types including HaCaT, NP, HDFa, IMR-90, MSC, HUVEC, KHOS, U2OS, and 143B, after 48 h in culture, establishing this time point as a reliable window for functional engagement. This observation was confirmed through biological and technical replicates of multiple assays, including cell migration, EVs uptake, cytotoxicity, and proliferation, underscoring the reproducibility and relevance of the 48 h mark for downstream functional studies.

The gene expression profile of MIAMI cells reflects a transcriptionally active and multipotent state, consistent with and extending beyond the molecular signatures commonly described for MSCs. As previously shown [25], MSCs typically exhibit a high expression of mesenchymal markers such as VIM, THY1, COL1A1, and FGF9; in contrast, MIAMI cells express these genes at lower levels. Instead, they demonstrate strong expression of genes associated with developmental plasticity (CDX2, MIXL1, LIN28A) and stem/progenitor maintenance (TERT, PROM1, PDGFRB), as shown in Figure 2. The high expression of TERT is particularly interesting, suggesting the retention of a more naïve, pre-committed phenotype. This profile may be attributable to the isolation and well-defined culture conditions used for MIAMI cell expansion, which have been shown to enhance stemness and delay senescence in MSCs [48].

The origin and physiological state of the parental cell directly influence the cargo and biological activity of secreted EVs. Compared to heterogeneous MSC-EVs, MIA-EVs are produced from a defined cell line, offering greater standardization than bulk MSC populations obtained from diverse tissue sources such as bone marrow, umbilical cord blood, placenta, adipose tissue, or dental pulp. Although MIAMI cells are donor-derived, their stable identity and consistent expansion protocols help maintain stemness and modulate paracrine signaling, directly impacting the molecular composition and functional properties of the resulting EVs (Figure 4) [86,87]. To interrogate the robustness and adaptability of this EVs’ regulatory program under clinically relevant stressors, MIAMI cells were further exposed to irradiation or inflammatory cytokine priming prior to EVs collection.

miRNA profiling and network analysis indicate that MIA-EVs harbor a highly organized and non-random regulatory miRNA architecture that converges on pathways governing metabolic resilience, cell survival, regeneration, and immune regulation. This structured functional bias suggests that MIA-EVs are preconfigured to stabilize cellular states under stress while supporting controlled tissue repair. Conditioning by irradiation or inflammatory priming modulates the relative weighting of these regulatory programs without disrupting their core identity. The persistence of a conserved 15-miRNA module across MIA-, MIA-IR-, and MIA-TIC-EVs indicates that MIAMI-derived EVs retain a conditioning-resistant regulatory backbone, while priming selectively reinforces pathways related to metabolic control, stress adaptation, and survival thresholds.

In contrast, although MIA-EVs and MSC-EVs share a substantial mesenchymal miRNA core, this shared framework appears more limited in regulatory breadth. Relative to MSC-EVs, MIA-EVs display increased pathway connectivity across regenerative, survival, and immune-related processes, consistent with a more integrated regulatory profile. These features may contribute to the observed enhancement of wound repair, maintenance of proliferative capacity under irradiation, and restrained tumor-supportive immune polarization. Consistent with this interpretation, several miRNAs shared between MIA-EVs and MSC-EVs—including miR-4454, miR-4516, miR-575, and miR-7975—have been independently associated with cellular stress responses, cytoskeletal organization, immune signaling, and proliferative regulation [88,89,90,91,92,93,94,95]. Collectively, these pathways are central to tissue repair and regenerative processes.

Macrophage polarization assays, shown in Figure 6, revealed that MIA-EVs induce a selective and limited shift in the transcriptional profile of M1-polarized macrophages. Rather than broadly activating M2-related genes, MIA-EVs specifically upregulated IL1R2, a decoy receptor associated with IL-1 signaling inhibition and inflammation resolution. Other M2 markers (CD163, MRC1, IL10) remained unchanged or low, while classical M1 genes (IL1B, TNF, NOS2) were suppressed. This expression pattern indicates partial modulation of the inflammatory phenotype without promoting full alternative macrophage activation. This targeted shift may assist in resolving inflammation without inducing broad immunosuppression, supporting its relevance in chronic wounds, autoimmune diseases, or cancer, where specific rather than full cell reprogramming is therapeutically advantageous.

The polarization pattern induced by MIA-TIC-EVs diverged from that of MIA-EVs, likely reflecting alterations in EVs cargo following inflammatory priming. In MIA-TIC-EVs-treated macrophages, concurrent upregulation of PPARG and STAT1 suggests activation of transcriptional programs involved in immune regulation and tissue remodeling. A limited induction of M1-associated genes was also observed, but expression levels remained low, indicating partial engagement of pro-inflammatory pathways without a dominant M1 phenotype. The relatively modest M2-associated response observed in both conditions may be attributed to intrinsic features of the parental MIAMI cells and their defined culture conditions, which preserves an undifferentiated state and may limit the capacity to fully activate anti-inflammatory response.

This restrained immunomodulatory profile aligns with the broader paradigm that EVs derived from naïve or undifferentiated stem cells generally exhibit limited macrophage-polarizing activity unless preconditioned or primed. For example, ESC-EVs have been reported to selectively promote M2 macrophage polarization while concurrently suppressing M1-associated gene expression [96]. Like MIAMI cells, ESCs maintain a naïve, lineage-uncommitted state, which may contribute to the generation of EVs with more defined and controlled bioactivity. Such specific characteristics are particularly advantageous in clinical contexts requiring minimal immune response, due to post-surgical recovery and tissue regeneration.

In contrast, MSC-EVs exhibit a broader range of activity, capable of inducing both pro- and anti-inflammatory pathways, as previously shown [25,97,98]. This dual-modulatory capacity makes MSC-EVs particularly suited for conditions that require controlled immune responses, such as chronic inflammation, autoimmune diseases, and tumor-associated immune modulation. However, this versatility also introduces variability. For instance, dental pulp stem cells-derived EVs (DPSC-EVs) display only modest anti-inflammatory effects unless activated by external stimuli, such as bacterial lipopolysaccharide (LPS) [99].

The MIA-EVs consistency is especially relevant in oncologic contexts. Studies have shown that MSC-EVs can have opposing effects on tumor biology depending on their origin and conditioning. For example, Weng et al. [100] reported that MSC-EVs from human bone marrow and umbilical cords promoted proliferation, migration, and tumorigenesis in several cancers, including nasopharyngeal carcinoma, osteosarcoma, renal, lung, and breast cancers. Similarly, Chulpanova et al. [101] highlighted the dual role of MSC-EVs in cancer, demonstrating that they can both influence tumor progression, either promoting dormancy or inhibiting angiogenesis, and serve as delivery vehicles for therapeutic agents such as miRNAs and chemotherapeutic drugs, thereby enhancing targeted cancer treatment strategies. Further supporting this context-dependent functionality, Luo et al. [102] showed that bone marrow-derived MSC-EVs enriched with long non-coding RNAs enhanced tumor cell proliferation, invasiveness, and angiogenic potential in osteosarcoma and other cancer types. In a related study, Luo et al. [103] demonstrated that MSC-EVs can regulate tumor vascularization in both stimulatory and inhibitory directions, reinforcing their bidirectional role in cancer biology. In contrast, Zhuo et al. [104] found that EVs derived from human placental MSCs suppressed neovascularization in a murine breast cancer model, suggesting that their bioactive cargo may alter cancer cell signaling to reduce the release of pro-angiogenic factors and ultimately inhibit tumor-associated angiogenesis.

Conversely, MIA-EVs demonstrated a targeted functional profile (Figure 5). They selectively inhibited cancer cell proliferation while supporting non-malignant cell growth in a dose-dependent manner, without upregulating TAM-associated markers such as TGFB1, CD68, CD163, NOS2, HIF1A, and CCL2. This low-level TAM engagement is favorable in translational applications where immune balance and tumor control are critical, as shown in Figure 6.

Our promising in vitro findings foster a next logical step of in vivo model testing. Such investigations should address biodistribution, dosing, safety, and efficacy. Additional mechanistic studies will aim to validate the functional roles of highly expressed miRNAs identified in MIA-EVs through EVs fractionation and both loss-of-function and gain-of-function approaches. Specifically, miRNA mimics will be used to overexpress target miRNAs, while miRNA inhibitors (antagomirs) will be employed to suppress their activity. To facilitate clinical translation, the incorporation of MIA-EVs into delivery platforms such as hydrogels will also be explored to enable controlled, localized applications in regenerative and therapeutic contexts.

5. Conclusions

This study provides the first comprehensive molecular and functional characterization of extracellular vesicles derived from MIAMI cells (MIA-EVs), a developmentally primitive mesenchymal stem/stromal cell subpopulation with enhanced regenerative and anti-inflammatory properties. MIA-EVs were found to retain molecular features of their parental cells, including selective enrichment of growth factors and miRNAs governing tissue repair, immune modulation, metabolic adaptation, and cell survival.

Across naïve (MIA-EVs), irradiated (MIA-IR-EVs), and cytokine-primed (MIA-TIC-EVs) conditions, MIAMI-derived EVs preserved a conserved, conditioning-resistant miRNA regulatory backbone that maintained coordinated control of survival signaling, stress adaptation, and regenerative readiness. Conditioning reshaped the relative weighting of these programs without disrupting their core regulatory identity, indicating that irradiation and inflammatory priming refine, rather than reprogram, the intrinsic functional logic of MIA-EVs. In parallel, comparative analysis with MSC-EVs demonstrated that while both EVs populations share a mesenchymal miRNA core, MIA-EVs uniquely expand pathway depth and network connectivity across regenerative, survival, and immune regulatory axes.

Functionally, MIA-EVs were non-cytotoxic to healthy cells, suppressed osteosarcoma cells proliferation, accelerated wound closure, and selectively modulated macrophage polarization. Baseline MIA-EVs promoted a controlled, pro-resolving macrophage phenotype through IL1R2 upregulation, while MIA-TIC-EVs further shifted macrophage transcriptional programs toward immunoregulatory and tissue-remodeling states marked by PPARG and STAT1 induction.

While miRNA and gene expression profiling provided mechanistic insight into these effects, all major conclusions were supported by functional assays demonstrating biologically measurable outcomes across multiple cell systems.

Together, these findings establish MIA-EVs as a novel and multifunctional EVs platform with therapeutic potential in regenerative medicine and inflammation-driven pathologies. The preservation of stable regulatory architecture across naïve, irradiated, and inflammatory environments provides a mechanistic foundation for their consistent functional performance under stress. Further preclinical studies are warranted to evaluate their in vivo efficacy, optimize delivery strategies, and elucidate cargo-specific mechanisms of action.