Human adipose stem cell-derived exosomes modulate the transcriptome of D-galactose-Induced neuronal cells

Abstract

Brain aging, which influences neurological function across cellular and molecular domains, is a critical concern in the elderly population. Therapeutic strategies for mitigating age-related neurodegeneration should target molecular pathways that are primarily involved in neuroinflammation. Exosomes derived from human adipose tissue mesenchymal stem cells (hASCs) have demonstrated anti-inflammatory and rejuvenating properties, making them promising agents for neurochemical intervention. However, their transcriptomic impact on neuronal cells remains largely unexplored. To address this research question, we applied high-throughput mRNA sequencing and downstream bioinformatic analysis. As an in vitro model for aging and neurodegeneration, CNS mouse-derived CAD cells were exposed to D-galactose (DG) to trigger molecular responses and were used to evaluate the efficacy of the isolated exosomes. The hASC-exosomes were isolated via ultrafiltration and subsequently characterized via nanoparticle tracking analysis, cryo-EM microscopy, and immunoassays. The internalization of PHK26-tagged hASC exosomes in the cytosol of the neuronal cells was monitored. Illumina-based mRNA sequencing has allowed expression profiling of more than 27,000 genes. Comparative transcriptomic profiling revealed 3951 differentially expressed genes (DEGs) associated with DG-induced cells and 3091 DEGs modulated by hASC-exosome treatment. In DG-treated cells, many genes were upregulated in response to cellular stress. The DEGs whose expression was upregulated in response to DG play roles in the DNA damage response, cellular senescence, and apoptosis. In the presence of hASC-derived exosomes, many DEGs (1948) were downregulated, suggesting that the exosomes suppressed stress-induced gene expression. The functional pathway analysis indicated that hASC-exosomes significantly downregulated processes related largely to translation, neuroinflammation, cellular senescence, apoptosis, and other age-associated molecular pathways. A set of genes involved in the inflammatory response and regulated by hASC-exosomes was identified. Our study provides transcriptomic evidence supporting the regulatory role of hASC-derived exosomes in attenuating the expression of inflammatory and neurodegenerative markers, positioning them as potential candidates for antiaging neurotherapeutics.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-34661-5.

Introduction

Neuronal degeneration is a central pathological hallmark of brain aging and is a major contributor to age-related cognitive decline^1,2. The process of neuronal aging originates at the molecular level, where disturbances in gene expression and transcriptional regulation propagate through the central dogma, ultimately disrupting protein synthesis, synaptic function, and higher-order brain activities^3–5. These transcriptomic alterations render neurons vulnerable to stress and accelerate the progressive deterioration of neurological function. Consequently, strategies for mitigating brain aging must address not only macroscopic clinical outcomes but also the molecular events that precede and drive neuronal decline⁶.

At the transcriptomic level, brain aging is characterized by an interplay of molecular hallmarks that undermine neuronal homeostasis⁴. Among these factors, DNA damage and genomic instability are primary drivers, arising from accumulated mutations, defective DNA repair mechanisms, and alterations in chromatin architecture. These events create aberrant transcriptional landscapes that perturb the regulation of genes critical for neuronal survival, synaptic plasticity, and stress resistance. In parallel, neuroinflammation emerges as a dominant pathological feature. The transcriptomic signatures of aging brains consistently demonstrate the upregulation of proinflammatory mediators, which reinforces a chronic inflammatory environment^7–9. Oxidative stress further compounds these deleterious processes by altering redox-sensitive transcription factors and suppressing antioxidant defense gene expression¹⁰. Collectively, these converging mechanisms highlight transcriptomic dysregulation as a pivotal mediator of neuronal aging.

Experimental models are indispensable for dissecting these molecular changes. The D-galactose (DG)-induced aging model has been extensively utilized as a robust in vitro and in vivo system to mimic neuronal aging. DG administration leads to oxidative stress, mitochondrial dysfunction, and DNA damage, thereby replicating key features of neurodegenerative progression^11–13. In neuronal cell lines, DG exposure induces senescence-like phenotypes and transcriptomic alterations consistent with those observed in aged brain tissue, making it a suitable platform for evaluating therapeutic interventions^14–18.

In recent years, exosomes derived from mesenchymal stem cells (MSCs) have gained considerable attention as potential therapeutic agents for age-related neurological conditions. These nanoscale extracellular vesicles (30–150 nm) carry diverse biomolecules and are capable of modulating gene expression in recipient cells^19–22. Exosomes derived specifically from human adipose tissue–derived stem cells (hASCs) have strong anti-inflammatory, antioxidant, and pro-regenerative properties^23–27. Despite these promising findings, the transcriptomic-level mechanisms through which hASC-derived exosomes exert neuroprotective effects remain largely unexplored.

Unbiased and high-throughput mRNA sequencing (mRNA-seq) with computational methods provides a comprehensive platform for investigating transcript abundance in RNA samples. By profiling differential gene expression, identifying signaling pathways, and annotating biological processes, mRNA-seq allows precise mapping of the molecular impact of therapeutic interventions²⁸. When applied to exosome research, this approach has the potential to elucidate how exosomes influence stress-related and aging-associated transcriptional modulation in recipient cells^29,30.

In the present study, we applied next-generation mRNA sequencing to examine the transcriptomic regulation mediated by hASC-derived exosomes in an in vitro model of DG-induced neuronal cells. We hypothesized that hASC exosomes would counteract DG-induced transcriptomic dysregulation, particularly by suppressing pathways related to neuroinflammation, cellular senescence, and apoptosis, while enhancing the expression of genes associated with neuronal protection. By integrating differential gene expression analysis with functional pathway enrichment, this work provides new evidence of the transcriptomic networks regulated by hASC-derived exosomes in neuronal cells, advancing their potential application as therapeutic candidates for neurodegenerative and aging-related disorders.

Materials and methods

Neuronal CAD cell culture

The mouse CNS-derived CAD cell line was obtained from the European Collection of Authenticated Cell Cultures (ECACC, accession number: 08100805) and cultured in Dulbecco’s modified Eagle’s medium (DMEM): nutrient mixture F-12 (F12) (Gibco, USA) supplemented with 8% fetal bovine serum (FBS) (HyClone, USA) and 1% antibiotic-antimycotic 100X (Gibco, USA). The cells were maintained in a humidified incubator at 37 °C with 5% CO₂ and 95% air (v/v). The cultured cells were regularly tested for mycoplasma contamination via a Mycoplasma PCR detection kit (Abcam, UK).

Human ASC culture and exosome isolation

Human primary adipose stem cells (hASCs) were purchased from iCell Biosciences, Shanghai, China (#catalog: HUM-iCELL-s2017). hASCs were cultured in high-glucose Dulbecco′s modified Eagle′s medium (DMEM) (D5648, Sigma‒Aldrich, USA) supplemented with 10% FBS, 1 mM sodium pyruvate (Gibco, USA), and 1% penicillin‒streptomycin solution (Gibco, USA) in a humidified incubator containing 5% CO₂ at 37 °C.

To collect the exosomes, the hASCs were washed three times with 1× PBS before being cultured in serum-free medium for 48 h to collect the conditioned medium (CM). The cell debris was removed via centrifugation at 2000 × g for 10 min and then passed through a 0.22 μm filter to remove large vesicles. The supernatants were concentrated via Macrosep advance centrifugal devices with an Omega 100 K membrane (product code: MAP100C37, Cytiva, USA) according to the manufacturer’s instructions. Briefly, CMs were added to the tube, and centrifugation at 3000 × g for 15 min was subsequently conducted to collect 100× concentrated CM. Subsequently, buffer exchange was performed to replace the culture media with PBS. The purified exosomes (500×) in PBS were passed through a 0.22 μm filter and stored at − 80 °C. To quantify the protein concentration, a BCA protein assay kit was used. The size, concentration and particle size distribution of the exosomes were determined via nanoparticle tracking analysis. The morphologies of the collected exosomes were observed via cryo-EM. To analyze exosome marker expression, exosome panel antibodies against the Calnexin, CD9, CD63, CD81, and TSG101 proteins were used for western blotting.

Nanoparticle tracking analysis

The size distribution and concentration (particle/mL) of the isolated ASC-exosomes were assessed with NanoSight Pro (Malvern Panalytical, UK) according to the manufacturer’s instructions. The samples were diluted to an appropriate dilution with PBS and then injected with a sterile syringe.

Cryogenic electron microscopy (cryo-EM)

The isolated exosomes were submitted to the Research Frontier Facility (FRF), Mahidol University, for morphological analysis of hASC-derived exosomes. An aliquot of each sample (3 µL) was applied to a plasma-cleaned copper grid (C-flat R 1.2/1.3, 300 mesh, Electron Microscopy Sciences) and prepared via an automated grid plunger (Vitrobot Mark IV, Thermo Fisher Scientific) with the environmental chamber set at 100% humidity and 4 °C. The grid was blotted for 4.5 s and vitrified in liquid ethane. The frozen hydrated exosomes were imaged via a cryogenic electron microscope (Glacios, Thermo Fisher Scientific) equipped with a Falcon 3EC. Cryo-EM imaging was performed at 73,000x and 120,000x magnification at 200 kV.

Cellular internalization of hASC-exosomes

A PKH26 red fluorescent cell linker mini kit for general cell membrane labeling (MINI26, Sigma‒Aldrich, USA) was used to observe the uptake of hASC-exosomes by neuronal cells. The mixture was combined with 100 µLof exosomes (50 µg), 4 µl of PKH26 dye, and 1 mL of Diluent C and incubated for 5 min at room temperature. The reaction was terminated by adding 10 mL of 8% FBS-containing cell culture media. The collection of labeled exosomes and removal of excess free dye were performed via Macrosep advance centrifugal devices as described previously. Then, the concentrated and labeled exosomes were collected, resuspended in complete culture media, and incubated with the neuronal cells for 16 h. The cells were fixed, permeabilized, and stained for βIII-tubulin and DAPI as described in the immunocytochemistry section. The same samples were used for 3D reconstitution and visualization via a high-end Leica Stellaris STED microscope (Leica Microsystems CMS, Mannheim, Germany) via Z-Stack images and LasX 3D visualization software.

Cell viability assay

To study the effect of isolated hASC-exosomes on cell viability, the cells were plated on a 96-well plate at a density of 2500 cells/well. Then, the cells were exposed to different concentrations of the exosomes (particle/mL) for 72 h. A control group was included, in which cells received the same final volume of PBS used to dilute the exosomes. To study the toxicity of D-galactose in CAD cells, D-galactose (G5388, Sigma‒Aldrich, USA) at concentrations ranging from 50 to 800 mM was applied to the cells. CAD cells were plated in a 96-well plate at a density of 7 × 10³ cells/well. After 24 h of plating, the cells were exposed for 24 and 48 h to observe the concentration‒cellular toxicity relationship. At the end of the experiment, cell viability was determined via an MTT assay (M5655, Sigma‒Aldrich, USA). The MTT stock reagent (5 mg/mL) was directly added to the conditioning media to obtain the final concentration (0.5 mg/mL) of the MTT reagent. The plates were further cultured for an additional 2 h. The MTT solution was carefully removed, and insoluble formazan was dissolved in DMSO. Colorimetric determination was performed at 570 nm via a spectrophotometer (EZ Read2000 microplate reader, Biochrom, USA). The data were normalized to the control data and are presented as the relative absorbance.

Immunocytochemistry

The cells were fixed with 4% paraformaldehyde (PFA) in PBS for 15 min at room temperature and rinsed three times with PBS. The cells were then permeabilized in 0.2% Triton X-100 at room temperature. The plates were washed with PBS three times. Nonspecific binding was blocked with 1% bovine serum albumin (BSA) in PBS for 30 min at 37 °C. The cells were stained with anti-phospho-γH2AX-S139 (cat no. SAB5700329, Sigma‒Aldrich, USA) on parafilm for 30 min at 37 °C before being washed with PBS three times, each for more than 5 min. The goat anti-rabbit IgG (H + L) cross-adsorbed secondary antibody, Alexa Fluor 488 (A32790, Invitrogen, USA), was incubated for 30 min at 37 °C. The incubated cells were washed with PBS three times, dipped in clean water, and mounted on slides using Fluoromount-G Mounting Medium with DAPI (cat no. 00495952; Invitrogen, USA). Images were acquired via a Zeiss LSM 800 confocal laser scanning microscope.

Western blot analysis

Proteins were isolated with RIPA lysis buffer containing 10% Triton X-100. The concentration of the extracted proteins was measured via a Pierce BCA protein assay kit (cat no. 23225; Thermo Scientific, USA). An identical concentration of total extracted protein was prepared in 1X LaemmLi SDS sample buffer (Thermo Scientific, USA), heated at 95 °C for 10 min, separated by SDS‒PAGE, and transferred to a PVDF membrane (1620177, Bio-Rad Laboratories, USA). The membranes were blocked in 2.5–5% nonfat dry milk in TBST for 30 min and incubated with primary antibodies against an exosome panel (ab275018, Abcam, USA), SIRT1 (9475 S, Cell Signaling, USA), or GAPDH (ABS16, Merck Millipore, USA). HRP-conjugated secondary antibodies for anti-mouse (7076, Cell Signaling, USA) or anti-rabbit (7074, Cell Signaling, USA) IgG were used for detection with enhanced chemiluminescence (ECL) reagents (GE Healthcare). The signals were imaged with a Fusion FX chemiluminescence system (Vilber, France).

Quantitative real-time polymerase chain reaction (qRT‒PCR)

Total RNA was extracted with TRIzol reagent (Invitrogen, Carlsbad, USA) according to the manufacturer’s instructions. The concentration and purity of the extracted RNA were determined via a NanoDrop One Spectrophotometer (Thermo Fisher Scientific, USA). cDNA synthesis was conducted via an iScript cDNA Synthesis Kit (Bio-Rad Laboratories, USA). Total cDNA (100 ng in 2 µl) was mixed with 8 µl of iTaq Universal SYBR Green Supermix (Bio-Rad Laboratories, USA) containing specific forward and reverse primers (400 nM each) to perform real-time PCR amplification, following the manufacturer’s instructions. The relative mRNA expression of the target genes was normalized to that of GAPDH, which served as the internal control and was calculated via the 2^−ΔΔCt method. A list of primers and their sequences for targeting genes in mice (Mus Muscularis) is provided in Supplementary Table 1.

mRNA sequencing-based transcriptomic analysis

RNA quantification and qualification

Total RNA was extracted according to the instruction manual of TRlzol Reagent (Life Technologies, California, USA). The RNA concentration and purity were measured via a NanoDrop 2000 (Thermo Fisher Scientific, Wilmington, DE). RNA integrity was assessed via both agarose electrophoresis and the RNA Nano 6000 Assay Kit of the Agilent Bioanalyzer 2100 system (Agilent Technologies, CA, USA). The software used in this study is listed in Supplementary Table 2.

Library preparation for transcriptome sequencing

A total of 1 µg of RNA per sample was used as input material for the RNA sample preparations. The sequencing libraries were generated via the NEBNext UltraTM RNA Library Prep Kit for Illumina (NEB, USA) following the manufacturer’s recommendations, and index codes were added to attribute the sequences to each sample. Briefly, mRNA was purified from total RNA via poly-T oligo-attached magnetic beads. Fragmentation was carried out using divalent cations under elevated temperature in NEBNext First Strand Synthesis Reaction Buffer (5X). First-strand cDNA was synthesized via random hexamer primers and M-MuLV reverse transcriptase. Second-strand cDNA synthesis was subsequently performed via DNA polymerase I and RNase H. The remaining overhangs were converted into blunt ends via exonuclease/polymerase activities. After adenylation of the 3’ ends of the DNA fragments, NEBNext adaptors with hairpin loop structures were ligated to prepare for hybridization. To preferentially select cDNA fragments 240 bp in length, the library fragments were purified with the AMPure XP system (Beckman Coulter, Beverly, USA). Then, 3 µl of USER Enzyme (NEB, USA) was used with size-selected, adaptor-ligated cDNA at 37 °C for 15 min, followed by 5 min at 95 °C before PCR. Then, PCR was performed with Phusion High-Fidelity DNA polymerase, universal PCR primers and Index (X) primers. Finally, the PCR products were purified (AMPure XP system), and library quality was assessed on an Agilent Bioanalyzer 2100 system.

Clustering and sequencing

Clustering of the index-coded samples was performed on a cBot Cluster Generation System via the TruSeq PE Cluster Kit v4-cBot-HS (Illumina) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina platform, and paired-end reads were generated.

Quality control and data analysis

Raw data (raw reads) in fastq format were first processed through in-house Perl scripts. In this step, clean data (clean reads) were obtained by removing reads containing adapters, reads containing poly-N sequences and low-quality reads from the raw data. Moreover, the Q20, Q30, GC content and sequence duplication level of the clean data were calculated. All the downstream analyses were based on high-quality, clean data.

Comparative analysis

The adaptor sequences and low-quality sequence reads were removed from the datasets. The raw sequences were transformed into clean reads after data processing. These clean reads were then mapped to the reference genome sequence (GRCm38.p6). Only reads with a perfect match or one mismatch were further analyzed and annotated on the basis of the reference genome. HISAT2 tools were used for mapping with the reference genome.

Gene functional annotation

Gene function was annotated on the basis of the databases listed in Supplementary Tables 3–4.

Quantification of gene expression levels

Gene expression levels were estimated by the number of fragments per kilobase of transcript per million fragments mapped. The following formula was used: FPKM = cDNA Fragments/(Mapped Fragments (Millions)* Transcript Length (kb)).

Differential expression analysis

For the samples with biological replicates, differential expression analysis of two conditions/groups was performed via DESeq2. DESeq2 provides statistical routines for determining differential expression in digital gene expression data via a model based on the negative binomial distribution. The resulting P values were adjusted via Benjamini and Hochberg’s approach for controlling the false discovery rate. Genes with an adjusted P value < 0.01 according to DESeq2 were considered differentially expressed. For the samples without biological replicates, differential expression analysis of two samples was performed via edgeR. An FDR < 0.01 and a fold change ≥ 2 were set as the thresholds for significantly differential expression.

GO enrichment analysis

Gene Ontology (GO) enrichment analysis of the differentially expressed genes (DEGs) was implemented via the GOseq R package-based Wallenius noncentral hypergeometric distribution³¹, which can adjust for gene length bias in DEGs.

KEGG pathway enrichment analysis

KEGG^32–35 is a database resource for understanding high-level functions and utilities of biological systems, such as cells, organisms and ecosystems, from molecular-level information, especially large-scale molecular datasets generated by genome sequencing and other high-throughput experimental technologies (http://www.genome.jp/kegg/). We used KOBAS³⁶ software to test the statistical enrichment of DEGs in KEGG pathways.

PPI (protein–protein interaction)

The sequences of the DEGs were subjected to BLAST (blastx) against the genome of a related species (the protein–protein interaction of which is present in the STRING database: http://string-db.org/) to obtain the predicted PPIs of these DEGs. The PPIs of these DEGs were subsequently visualized via Cytoscape³⁷.

Statistical analysis and reproducibility

GraphPad Prism software was used for statistical analysis, and the outcomes are reported as the means ± SDs. Comparisons among different groups were performed via one-way ANOVA. A significant difference was considered when the p value was less than 0.05 or as indicated. Replications are mentioned in the relevant figure legends.

Results

Isolation and characterization of hASC-exosomes

As shown in Supplementary Fig. 1, human ASCs were expanded in large-scale culture. The hASCs exhibited the typical morphological characteristics of mesenchymal stem cells, with a spindle-shaped, fibroblast-like appearance (Fig. 1a). The conditioned media (CM) were collected for exosome isolation. The extracted hASC-exosomes were then characterized and used for cellular and molecular evaluation.

Fig. 1.

Characterization of hASC-derived exosomes isolated via ultrafiltration. (a) Spindle-shaped, fibroblast-like appearance of human ASCs. (b) Size distribution and concentration (particles/mL) of extracted hASC-exosomes measured via nanoparticle tracking analysis. (c) A representative image from cryo-EM to visualize the nanoscale, membrane-bound vesicular structures of the isolated hASC-exosomes. Scale bar: 50 nm. (d) Western blot analysis of the expression of positive (CD9, CD63, TSG101, HSP70, and GAPDH), negative (calnexin), and sirtuin-1 (SIRT1) markers in hASC lysates and their isolated exosomes.

The results of the nanoparticle tracking analysis revealed that the average size of the nanoparticles was 114.5 nm, which is within the range of the size of the exosomes used in most studies (Fig. 1b). The morphology of hASC-derived exosomes was examined via cryo-EM, which revealed nanoscale, membrane-bound vesicular structures (Fig. 1c). To confirm whether the isolated nanoparticles were exosomes, the expression of various MSC exosome markers was detected via western blot analysis. As shown in Fig. 1d, exosomal markers, including CD9, CD63, TSG101, and HSP70, were detected, whereas the negative marker calnexin was absent, confirming the successful isolation of the exosomes. Interestingly, we also detected a high level of sirtuin-1 (SIRT1), a molecule that plays a significant role in the anti-aging process. Together, the data from the characterization methods confirmed the isolation of hASC-exosomes.

Uptake of hASC-exosomes by neuronal CAD cells

The neuronal CAD cell line used in this study was derived from the CNS and expresses neuron-specific markers^38–40. To investigate whether hASC-exosomes are internalized into the neuronal CAD cells, the cells were exposed to PKH26-tagged exosomes for 16 h. Fluorescent signals were observed within the cells, which appeared as nanoscale punctate dots (Fig. 2a,b). 3D visualization was performed to confirm that the exosomes were internalized within the cytosol rather than remaining attached to the cell membrane (Fig. 2a and Supplementary File 1). These data suggest that hASC-derived exosomes are taken up by neuronal CAD cells.

Fig. 2.

Uptake of hASC-derived exosomes by neuronal CAD cells. (a) PKH26 labeling combined with immunocytochemistry for βIII-tubulin (neuronal marker) and nuclear staining (DAPI) demonstrated the internalization of hASC-exosomes in neuronal cells after 16 h of incubation. Scale bar: 50 μm. (b) Enlarged 2D image from panel (a) showing PKH26-labeled exosomes localized within neuronal cells. (c) 3D reconstruction confirming the intracellular internalization of PHK 26-labeled exosomes. Scale bar: 5 μm.

D-galactose-induced CAD cells were used as an in vitro model for testing transcriptomic modulation by hASC-derived exosomes

Next, the cytotoxicity of hASC-exosomes was measured via the MTT assay. The cells were exposed to various concentrations of exosomes ranging from 10⁵ to 10¹¹ particles/mL. The MTT assay revealed that there was no toxicity caused by hASC-exosomes at the selected concentrations (Fig. 3a). To establish a cellular model for evaluating the effects of exosomes on transcriptomic modulation, we applied D-galactose (DG), a widely used inducer for studying aging and neurodegeneration^17,18,41. To select the concentration of DG, the cells were exposed to various concentrations of DG (50–800 mM). The results of the cytotoxicity assay revealed that DG at concentrations greater than 200 mM reduced cell viability (Fig. 3b). DG at concentrations of 200, 400, and 800 mM reduced cell viability by approximately 20%, 50%, and 90%, respectively. Additionally, treatment with 400 mM DG led to upregulation of oxidative stress–responsive genes and induced DNA damage, as evidenced by the pronounced γ-H2AX signal (Supplementary Fig. 2). Therefore, 400 mM DG was selected as the optimal concentration for eliciting gene expression responses in CAD model cells in subsequent experiments.

Fig. 3.

Effects of hASC-exosomes on neuronal CAD cell viability and gene expression in a D-galactose-induced neuronal model. (a) Relative absorbance of CAD cells incubated with increasing concentrations of hASC-exosomes (particles/mL) for 72 h, showing no cytotoxic effect. (b) Relative absorbance of CAD cells exposed to different concentrations of D-galactose (mM) for 24 h and 48 h, demonstrating a dose- and time-dependent reduction in cell viability. Relative mRNA expression of senescence- and inflammation-associated genes in CAD cells treated with D-galactose (200 mM, 24 h) and hASC-exosomes at two different concentrations (1 × 10¹¹ and 1 × 10¹⁰ particles/mL). The expression levels of p21 (c), p16 (d), nf-kb (e), Il6 (f), and gdf15 (g) were quantified. All the data are presented as the means ± SEMs. Each dot represents an individual value from each biological replicate. The data were collected from at least three biological samples. Statistical significance was analyzed via one-way ANOVA. **p < 0.01, ***p < 0.001 (vs. control at 24 h); + p < 0.01, ++ <0.001 (vs. control at 48 h); ^#p < 0.01, ^##p < 0.001 (vs. D-galactose group, DG).

hASC-exosomes suppressed the upregulation of genes associated with cellular stress response and neuroinflammation in DG-induced CAD cells

To assess the efficacy of the gene regulation of hASC-exosomes in DG-induced neuronal cells, we utilized conventional RT‒PCR to observe the expression of genes associated with aging, DNA damage, and inflammation. DNA damage induces the expression of p16 and p21, which further promote cellular senescence and apoptosis^7,42. The upregulation of nf-kb, an inflammatory regulator, can induce the production of proinflammatory cytokines such as interleukin-6 (il6) and growth differentiation factor 15 (gdf15), which are also molecular hallmarks that are increased in aging plasma^8,43. As shown in Fig. 3c–g, DG significantly promoted the upregulation of these genes. The presence of hASC-exosomes (10¹⁰ and 10¹¹ particles/mL) significantly suppressed the upregulation of these genes, indicating the ability of hASC-exosomes to regulate gene expression in neuronal cells.

Illumina-based mRNA sequencing analysis identified 3019 genes that were differentially expressed in the presence of hASC-exosomes in DG-neuronal cells

As a state-of-the-art and high-throughput method, Illumina-based mRNA sequencing (mRNA-seq) was applied to profile transcriptomic changes, revealing massive amounts of information on gene regulation by hASC-exosomes in neuronal cells. Total RNA from three groups: untreated cells; CTL (A), D-galactose-treated cells; DG (B), or DG- and hASC-Exo (10¹¹ particles/mL)-treated cells; DGE (C) were extracted. A diagram of the mRNA-seq workflow, including sample preparation, library construction, library quality control and sequencing, is shown in Supplementary Fig. 3a,b. The analysis quantified expression profiles (FPKM) for more than 27,000 genes. (Supplementary File 2). The sample information, sequencing data statistics, and data mapping statistics are provided in Supplementary Tables 5–7.

To ensure the reliability of the RNA-seq dataset, key quality metrics were assessed. All samples showed high sequencing depth, ranging from 39.6 to 44.2 million total reads per library (Supplementary Table 7). Alignment rates were consistently high. Unique mapped reads accounted for 82–89% of total reads across samples, further confirming mapping quality. Quality control analyses were performed, and the corresponding plots were generated to assess data reliability. The PCA plot revealed that the mRNA expression profiles among the three groups were different, as indicated by separate clusters among the groups (Fig. 4a). Additionally, three biological replicates derived from each group were clustered together, meaning that the data were reproducible among the submitted replicates. Pearson correlation analysis demonstrated strong consistency among biological replicates, with correlation coefficients ranging from r = 0.93 to 0.99 (Fig. 4b). Sample-to-sample distance clustering confirmed the difference in the gene expression profiles among the three groups, while the three biological replicates were similar. Interestingly, the clustering pattern revealed that groups A (control, CTL) and C (exosome + DG treatment) presented greater transcriptomic similarity, while both were distinctly separated from group B (DG treatment alone). This observation suggests that hASC-derived exosomes may mitigate the transcriptomic alterations induced by DG. Overall, both the PCA and the clustering analysis indicate that there is transcriptional distinction among the three sample groups, whereas three biological replicates submitted from each sample share high similarity in their transcriptomic profile, indicating that the subsequent analyses of differentially expressed genes (DEGs) and the bioinformatic assessment of functional enrichment can be considered reliable.

Fig. 4.

Principal component analysis and sample correlation of transcriptomic profiles. (a) Principal component analysis (PCA) plot showing the distribution of samples across three experimental groups; A, B, and C. (C). Three biological samples (1–3) from each experimental group were subjected to mRNA-seq. The first two principal components (PC1 and PC2) account for 46.62% and 27.99% of the variance, respectively, indicating clear separation among groups. (b) Heatmap of sample‒to‒sample Pearson correlation coefficients with hierarchical clustering. High intragroup similarity and distinct intergroup clustering confirm the reliability and reproducibility of the RNA sequencing data. A; Untreated cells (CTL), B; DG-treated cells (DG), and C; DG- and hASC-exosome-treated cells (DGEx).

The number of DEGs was statistically identified and is summarized in Table 1 and Supplementary File 3. In total, 3951 DEGs were detected between groups A and B, whereas 3019 DEGs were identified between groups B and C. Volcano plots and MA plots were used to visualize the overall distribution of gene expression and the fold change in expression levels between the two compared samples (Fig. 5a,b and Supplementary Fig. 4). Together with the heatmap hierarchical clustering presented in Fig. 4c,d, the results revealed a clear distinction between comparative groups on the basis of the abundance profile of DEGs. In DG-treated cells, many genes were upregulated in response to cellular stress, as reflected by the greater number of upregulated DEGs (2540) than downregulated DEGs (1411). However, in the presence of hASC-derived exosomes, there were 1948 downregulated DEGs, suggesting that the exosomes suppressed stress-induced gene expression. Additionally, Venn diagram analysis of DEGs between samples revealed the number of unique DEGs between each comparison pair and the overlap of DEGs between different comparison pairs (Supplementary Fig. 5).

Table 1.

Differentially expressed gene counts between compared pairs.

Comparing set	Number of differentially expressed genes (DEGs)
	Total	Upregulated genes	Downregulated genes
Untreated (A) vs. DG-treated (B)	3951	2540	1411
B vs. DG + hASC-exosome-treated(C)	3019	1071	1948

Fig. 5.

Differential gene expression analysis between the experimental groups. Volcano plots to visualize the differentially expressed genes (DEGs) among the compared groups; (a) A vs. B, and (b) B vs. C. Each dot represents a gene. X-axis: log₂(fold change) of expression; Y-axis: -log₁₀(FDR). The blue dots represent downregulated DEGs, the red dots represent upregulated DEGs, and the black dots represent genes whose expression did not significantly differ. Heatmaps representing hierarchical clustering of DEGs between each comparison group; (c) A vs. B, and (d) B vs. C. Red and blue indicate up- and downregulated DEGs, respectively.

The total DEGs obtained from differential expression analysis were annotated and subsequently subjected to functional enrichment analysis, with the enriched GO terms presented in Supplementary Fig. 6. In the comparison of untreated cells (A) with DG-treated cells (B), a greater number of upregulated DEGs than downregulated DEGs were observed, and these DEGs were annotated to be associated with various biological processes, cellular components, and molecular functions. In contrast, in the comparison between DG-treated cells (B) and DG plus exosome-treated cells (C), the number of downregulated DEGs exceeded that of the upregulated DEGs. This shift suggests that hASC-derived exosomes exert a modulatory influence on gene expression in neuronal cells, counteracting DG-induced transcriptional alterations.

KEGG pathway analysis further revealed that the DEGs from the A vs. B comparison were enriched in the p53 signaling pathway, the cell cycle, DNA replication, purine and pyrimidine metabolism, and mismatch repair. These results confirm the impact of DG on neuronal cell function, particularly highlighting its deleterious effects on cell cycle regulation (Supplementary Fig. 7a). On the basis of the results of the GO enrichment analysis, the biological process associated with the DG-treated cells that was most significantly enriched was the cellular response to DNA damage stimulus, in which the H2AX gene was enriched (Fig. 5a,b). These findings are consistent with the confocal imaging of γ-H2AX and confirm that DG induces a DNA damage response, thereby contributing to deleterious effects on cell cycle regulation.

The KEGG pathway enrichment bubble chart from comparisons B and C revealed several pathways associated with neuroprotection and cellular resilience, including MAPK signaling, cAMP signaling, the longevity-regulating pathway, FoxO signaling, and estrogen signaling. The enrichment of these pathways is particularly noteworthy, as they are known to mediate cell survival, stress adaptation, and transcriptional programs that preserve neuronal integrity (Supplementary Fig. 7b). As shown in Fig. 6, the protein–protein interaction (PPI) network of the DEGs revealed a large cluster centered around mitogen-activated protein kinase 13 (MAPK13), also known as stress-activated protein kinase 4 (SAPK4). This clustering highlights MAPK13 as a potential hub gene mediating hASC-exosome stress-responsive signaling and suggests that hASC-derived exosomes may influence transcriptomic networks involving MAPK13.

Fig. 6.

D-galactose induced the upregulation of various genes, including those associated with the DNA damage response, in neuronal CAD cells. (a) GO term enrichment analysis of total DEGs between groups A and B. Bar chart showing the number of DEGs associated with significantly enriched biological process GO terms. The bars are color-coded by the q value, and the x-axis represents the number of genes involved in each process. (b) Circular network diagram illustrating gene‒function associations of total DEGs from groups A vs. B. Genes are arranged around the circle and connected to their respective GO terms. Node color represents the fold change (from blue to red), and black circles indicate gene counts per term. This visualization highlights that the H2AX gene plays a key role in the cellular response to DNA damage stimuli.

D-galactose promoted the upregulation of genes associated with aging and neurodegeneration but downregulated DNA replication, translation and cell cycle-related processes in neuronal CAD cells

To identify transcriptomic changes induced by DG treatment in the neuronal CAD cell model, sets of upregulated and downregulated DEGs from the A vs. B comparison pair were separately subjected to pathway annotation. As shown in Supplementary Fig. 8a, DG treatment promoted the upregulation of genes involved in cellular senescence, apoptosis, and several aging-associated signaling pathways, including the MAPK, PI3K-Akt, cAMP, mTOR, and JAK-STAT signaling pathways. In contrast, DG markedly downregulated genes associated with cell cycle progression, DNA replication, translation, and DNA repair, the majority of which were localized in the nucleus (Supplementary Fig. 8b-c). Collectively, these results suggest that DG induces a dual transcriptomic response characterized by the activation of stress- and aging-related pathways together with the suppression of fundamental processes required for genomic stability and cellular renewal. This pattern reflects the deleterious impact of DG on neuronal cells and is consistent with previous evidence indicating that DG is a widely used inducer of aging and neurodegeneration in experimental models.

hASC-exosomes induced the upregulation of genes associated with metabolic processes, DNA replication, and neuronal signal transduction but suppressed global translation in the DG-treated model

In the comparison of B vs. C, the upregulated DEGs were enriched in several metabolic and neuronal processes, including cholesterol biosynthetic processes, oxidative phosphorylation, regulation of cell junction assembly, receptor internalization, neuronal signal transduction, mitochondria–nucleus signaling, calcium ion transport, synapse organization, the cellular response to potassium ions, and negative regulation of axon regeneration, as well as pathways related to DNA replication and regulation of cell cycle arrest (Fig. 7a). Notably, these upregulated genes were predominantly associated with neuronal compartments such as the postsynaptic density, neuronal cell body, glutamatergic synapse, dendritic shaft, neuron projection, presynaptic active zone, and voltage-gated sodium channel complex. Furthermore, genes involved in the DNA replication complex, factor C complex, histone deacetylase complex, and CD40 receptor complex were also upregulated in the presence of hASC-exosomes (Fig. 7b,c). Together, these findings indicate that hASC-derived exosomes transcriptionally promote multifaceted protective effects on neurons exposed to DG-induced stress.

Fig. 7.

Protein–protein interaction (PPI) network of the total identified DEGs revealing a large cluster centered around MAPK13 (mitogen-activated protein kinase 13). This clustering highlights MAPK13 as a potential hub gene mediating hASC-exosome stress-responsive signaling and suggests that hASC-derived exosomes influence transcriptomic networks involving MAPK13.

Phylogenetic classification of DEGs via Clusters of Orthologous Groups of Proteins (COG) was performed to determine the frequency of genes assigned to each functional category on the basis of orthologous groups. Using the list of upregulated DEGs in the DG-treated cells, the bar corresponding to K: transcription shows a marked increase in frequency, indicating that DG treatment alone upregulated transcription-related genes in neuronal cells, which could be a cellular stress response to cope with D-galactose-induced oxidative/genotoxic stresses (Supplementary Fig. 9a). Conversely, analysis of downregulated genes in cells treated with hASC-derived exosomes in addition to DG revealed that the same functional category (K) was substantially lower in frequency than in those treated with DG alone (Supplementary Fig. 9b). These findings suggest that hASC-exosome treatment reduces the overactivation of transcription-related genes caused by D-galactose, reducing aberrant gene activation and potentially protecting against premature aging or neurodegeneration.

A number of genes involved in inflammatory responses are regulated by hASC-exosomes

Next, the prominent genes that were significantly downregulated by hASC-exosomes in DG-induced neuronal cells were identified. GO enrichment analysis was performed to identify the top biological processes that may underlie the set of downregulated genes (Fig. 8). More than 100 downregulated genes were categorized into a pathway regulating translation and a DNA-templated pathway, supporting previous results concerning the role of hASC-exosomes in transcriptional regulation. Interestingly, we detected a set of genes involved in the cellular response to interleukin-13 and in the positive regulation of the inflammatory response that were also regulated by hASC-exosomes. The transcriptomic expression (FPKM) values of the genes involved in the inflammatory response are plotted in Fig. 9. The list includes Ankyrin Repeat Domain 42 (Ankrd42), Cyclin-Dependent Kinase 19 (Cdk19), ETS proto-oncogene 1 (Ets1), Interleukin 33 (Il33), NFKB Inhibitor Alpha (Nfkbia), NF-kappa-B inhibitor zeta (Nfkbiz), Programmed Cell Death 4 (Pdcd4), Transglutaminase 2 (Tgm2), Toll-like receptor 2 (Tlr2), Thymic Stromal Lymphopoietin (Tslp), ADAM Metallopeptidase Domain 8 (Adam8), Complement component 3 (C3), E74 Like ETS Transcription Factor 3 (Elf3), EPH receptor A2 (Epha2), Interleukin-15 (Il15), Interleukin 17 C (Il17c), Interleukin-17 F (Il17f), Myeloid differentiation primary response 88 (Myd88), Nuclear Protein 1 (Nupr1), S100 Calcium Binding Protein A7A (S100a7a), Arachidonate 15-lipoxygenase (Alox15), Fem-1 homolog A (Fem1a), and sedoheptulokinase (Shpk). These data indicate that hASC-derived exosomes not only regulate transcriptional processes but also suppress proinflammatory gene expression, underscoring the regulatory effects of hASC-exosomes on inflammation in DG-induced neuronal cells.

Fig. 9.

A number of genes involved in the regulation of translation and inflammatory responses are regulated by hASC-exosomes in DG-induced neuronal cells. (a) GO term enrichment analysis of downregulated DEGs between group B and group C, with a focus on biological process categories. The bar chart displays the number of genes (GeneNum) associated with each enriched process, with bars color-coded by statistical significance (q value). Notably, the process regulating transcription, DNA templating, is prominently represented among the downregulated genes, indicating a shutdown of transcriptional activity. An additional enriched process was positive regulation of the inflammatory response (square), suggesting the impact of hASC-exosomes on the suppression of inflammatory genes in D-galactose-induced neuronal cells. (b) Circular network diagram illustrating gene-to-process associations. Genes are associated with various biological processes, such as the regulation of transcription, DNA templating, protein refolding, positive regulation of ion transmembrane transport, positive regulation of the inflammatory response, and the cellular response to interleukin-13, highlighting the gene regulatory roles of hASC-exosomes.

Fig. 8.

hASC-exosomes induced the upregulation of genes associated with metabolic processes, DNA replication, neuronal signal transduction, and synapse compartments in neuronal CAD cells. GO term enrichment analysis of upregulated DEGs between B and C, focusing on (a) biological process categories and (b) cellular component categories. Bar chart showing the number of genes associated with each enriched category. The bar colors represent statistical significance on the basis of the q value ranges. (c) Circular network diagram illustrating gene‒component associations. The genes are related to their respective cellular components, i.e., postsynaptic density, the neuronal cell body, glutamatergic synapses and the dendritic shaft, highlighting the role of hASC-exosomes in the upregulation of genes associated with synapses and neuronal compartments.

Genes associated with aging and neurodegeneration are suppressed by hASC-exosomes

Stem cell–derived exosomes have been proposed to modulate several pathways associated with aging and neurodegeneration. However, their global effects on transcriptomic regulation remain unclear. The present mRNA-seq analysis reveals the involvement of such pathways, thereby underscoring the regulatory role of hASC-derived exosomes at the transcriptomic level. As presented in Supplementary Fig. 10, the downregulated DEGs obtained from the comparison between groups B and C were subjected to pathway enrichment analysis. The data highlight a broad spectrum of pathways in which the downregulated genes are involved. In particular, the downregulated genes were significantly enriched in cellular senescence and apoptosis. Multiple signaling cascades, including the PI3K-Akt, MAPK, FoxO, Wnt, mTOR, and longevity-regulating pathways, were also downregulated. These pathways are tightly linked to aging regulation, neuronal survival, and synaptic plasticity. The expression levels of genes associated with apoptosis and cellular senescence across the experimental groups are presented in Fig. 10. The genes included Astrotactin-2 (Astn2), B-cell lymphoma 2-like 11 (Bcl2l11), cIAP1 (Birc2), Caspase-3 (Casp3), CASP8 and FADD-like apoptosis regulator (Cflar), Cathepsin K (Ctsk), DNA damage-inducible transcript 3 protein, also known as CHOP (Ddit3), DNA fragmentation factor subunit beta (Dffb), protein kinase RNA-like endoplasmic reticulum kinase, also known as PERK (Eif2ak3), c-fos, forming part of the AP-1 transcription factor complex (Fos), growth arrest and DNA damage inducible alpha, beta, and grammar (Gadd45a, Gadd45b, and Gadd45g, respectively), inositol 1,4,5-trisphosphate receptor type 2, acting as an intracellular calcium channel (Itpr2), Jun proto-oncogene, AP-1 transcription factor subunit (Jun), mitogen-activated protein kinase kinase 14 (Map3k14), regulatory subunits of phosphatidylinositol 3-kinase (Pik3r1, Pik3r3), Tyrosine-protein phosphatase nonreceptor type 13 (Ptpn13), The suppressive effect on gene expression suggested that hASC-exosomes counteract DG-induced cellular stress responses that would otherwise accelerate cellular aging and death.

Fig. 10.

Expression modulation of genes involved in the inflammatory response by hASC-exosomes in DG-induced neuronal cells. Bar graphs showing the transcriptomic expression (FPKM) of selected genes (a–w) across three conditions: control, D-galactose (DG)-treated, and DG- and hASC-exosome-treated (DG + Exo) cells. Each value represents a biological replicate (n = 3). Statistical significance was determined via one-way ANOVA. *p < 0.05, **p < 0.01, ***p < 0.001 (vs. control) and ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 (vs. DG).

Discussion

The present study provides transcriptomic evidence that exosomes derived from hASCs exert a modulatory influence on neuronal cells exposed to DG, an established inducer of aging- and neurodegeneration-like phenotypes. Using high-throughput mRNA sequencing, we identified DEGs and functional pathways that were significantly altered by DG treatment and subsequently attenuated by hASC-exosomes. These findings advance the current knowledge of the neuroprotective properties of hASC exosomes by elucidating their impact on global gene expression programs relevant to neuronal survival, stress responses, and inflammatory regulation.

DG exposure promoted substantial upregulation of genes associated with the DNA damage response, senescence, and apoptosis but downregulated genes involved in DNA replication, translation, and cell cycle progression. This dual transcriptomic profile reflects the vulnerability of neuronal cells under genotoxic stress, which is consistent with previous reports that DG induces oxidative damage, mitochondrial dysfunction, and genomic instability^44–46. The upregulation of senescence markers such as p21 and p16, alongside proinflammatory mediators, including nf-κb, il6, and gdf15, aligns with known molecular signatures of neuronal aging. The identified DEGs obtained from our mRNA-seq studies serve as a reference dataset underlying D-galactose-induced neuronal stress responses. Together, these results validate the DG model as a reliable platform for exploring transcriptomic perturbations underlying brain aging and for assessing interventions designed to mitigate such alterations.

The presence of hASC-derived exosomes markedly reshaped the transcriptomic landscape of DG-induced neuronal cells. Nearly two thousand DEGs whose expression was upregulated in response to DG were downregulated upon exosome treatment, suggesting a suppressive effect on stress-induced gene expression. The results of functional enrichment analyses highlighted several biological processes modulated by hASC exosomes, including neuroinflammation, apoptosis, and cellular senescence. Particularly notable was the downregulation of proinflammatory signaling networks, as reflected by the decreased expression of genes such as Il33, Tlr2, C3, and Nfkbiz. These findings are consistent with accumulating evidence that stem cell–derived exosomes exert anti-inflammatory effects^47–49. In addition to their ability to suppress detrimental pathways, hASC exosomes appeared to modulate transcriptomic programs associated with neuronal vulnerability. The upregulated DEGs in the exosome-treated groups were enriched in processes related to neuronal signal transduction, synaptic organization, oxidative phosphorylation, and DNA replication. Genes associated with neuronal compartments, such as the postsynaptic density, dendritic shafts, and glutamatergic synapses, were particularly increased, suggesting that exosomes play a supportive role in maintaining neuronal connectivity. These results suggest that hASC exosomes not only mitigate stress-induced damage but also transcriptionally facilitate neuroprotective mechanisms. However, future studies incorporating functional assays such as apoptosis measurement, ROS quantification, or neurite morphology analysis will be necessary to determine whether these transcriptional changes translate into measurable neuroprotective outcomes (Fig. 11).

Fig. 11.

Modulation of the expression of apoptosis-related genes in response to hASC exosomes in DG-induced neuronal cells. Bar graphs showing the transcriptomic expression (FPKM) of selected genes (a–v) across three conditions: control, D-galactose (DG)-treated, and DG- and hASC-exosome-treated (DG + Exo) cells. Each value represents a biological replicate (n = 3). Statistical significance was determined via one-way ANOVA. *p < 0.05, **p < 0.01, ***p < 0.001 (vs. control) and ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 (vs. DG).

The involvement of signaling pathways such as the MAPK, PI3K-Akt, cAMP, and FoxO pathways further underscores the pleiotropic regulatory roles of hASC-derived exosomes. Similar results were reported in an in vivo model of acute myocardial infarction⁵⁰. The identification of MAPK13 as a central node in the protein–protein interaction (PPI) network suggests that this kinase may serve as a critical hub mediating the cellular response to exosomal signals under stress conditions⁴⁶. MAPK family members are well known to integrate oxidative, inflammatory, and apoptotic stimuli, and their modulation by exosomes could represent a key mechanism of neuroprotection. In parallel, the enrichment of longevity-regulating and estrogen signaling pathways highlights potential cross-talk between exosome-mediated regulatory factors and systemic factors known to influence neuronal aging.

Several limitations of the present study should be acknowledged. First, this study employed only a single neuronal model. Although murine CAD cells exhibit catecholaminergic features and originate from the central nervous system, their species background and restricted neuronal subtype limit direct generalization of the transcriptomic findings to broader neuronal populations. Future work should extend these observations by validating the exosome-induced transcriptomic signatures in additional neuronal systems, including human-derived neuronal cell lines or iPSC-derived neurons, to strengthen the translational relevance of the findings. Second, while mRNA sequencing provides a comprehensive overview of transcriptomic changes, it does not capture posttranscriptional regulation, protein modifications, or functional outcomes at the cellular or behavioral level. Future studies integrating proteomic analyses will be essential to validate the biological significance of these transcriptomic modulations. Third, the exosome preparations in this study were derived from hASCs under basal culture conditions. Preconditioning strategies, such as exposure to hypoxia or small molecules, have been reported to increase the therapeutic potential of exosomes^51–56. Comparative analyses of exosomes from differently conditioned MSCs could further clarify the underlying mechanisms and optimize their application. Lastly, MSC-derived exosomes contain diverse cargo, including miRNAs, proteins, and metabolites. In this study, we detected SIRT1 protein within ASC-derived exosomes, indicating possible regulatory involvement. However, the specific cargo responsible for the observed transcriptomic changes remains undefined. Future OMIC-based profiling will be required to identify these components.

Despite these limitations, the present findings provide compelling transcriptomic-level evidence supporting the potential of hASC-derived exosomes in mitigating neuronal aging. By downregulating stress- and inflammation-related genes while enhancing pathways linked to neuronal survival, energy metabolism, and synaptic organization, these exosomes orchestrate a multifaceted protective effect. These properties underscore their promise as a novel class of nanoscale agents for targeting age-related neurodegenerative conditions.

Conclusion

Through mRNA sequencing and a model of D-galactose-induced neuronal cells, we profiled the effects of hASC-exosomes on transcriptomic regulation in neuronal cells. Several genes, including those involved in inflammation and neurodegeneration, are regulated by hASC-exosomes. This high-throughput screening revealed the neuroregulatory roles of hASC-exosomes at the transcriptomic level. Future comparative studies of exosomes obtained from various MSC sources are warranted to clarify their differential effects on neuronal gene regulation.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

E.K. developed the concept, designed the experiments, collected and analyzed the data, and wrote the manuscript. K.P. performed the in vitro studies, cultured hASCs, and isolated the exosomes. J.M. and P.M. performed the Cryo-EM and data processing. W.C. performed gel electrophoresis prior to mRNA-seq. N.S. collected the confocal images and performed 3D image visualization. B.C. supervised all aspects of this project. All the authors had the opportunity to discuss the results and comment on the manuscript.

Funding

This research project is supported by Mahidol University (Strategic Research Fund): 2024.

Data availability

The data will be made available upon proper request to the corresponding author. The datasets generated and/or analysed during the current study are available in the ArrayExpress repository accession E-MTAB-15753.

Declarations

Competing interests

The authors declare no competing interests.

Declaration of generative AI and AI-assisted technologies in the writing process

This manuscript was prepared with the assistance of generative AI and AI-assisted technologies to improve its clarity, grammar, and overall readability. The tools utilized included Grammarly, and Curie which are employed for language editing and text generation.