Laser-induced skin microenvironment remodeling via exosomal CEBPA-mediated crosstalk between keratinocyte and fibroblast Authors: Qin Zou, Hao Wang, Yanjun Zhou, Xiang Wen Published: 2026 ABSTRACT AbstractLaser treatments are widely applied in aesthetic and clinical dermatology to promote skin rejuvenation by inducing controlled tissue disruption, activating repair processes involving keratinocyte proliferation, fibroblast migration, and extracellular matrix remodeling. However, the cellular specialization and intercellular communication mechanisms, particularly in the early stages following laser treatment, remain poorly understood. Using single-cell RNA sequencing, we analyzed laser-treated and untreated skin, revealing significant differences in cell populations and functions. Keratinocytes and fibroblasts displayed the most prominent shifts in cell number and transcriptional diversity following laser exposure. In early-stage laser-induced skin remodeling, distinct keratinocyte subpopulations actively orchestrated skin repair, immune responses, and regenerative signaling. Cell-cell communication analysis uncovered a dynamic crosstalk between keratinocytes and fibroblasts. Functional validation through co-culture revealed this crosstalk was mediated by exosomes derived from laser-treated keratinocytes, which were enriched with CEBPA. Mechanistically, these exosomes enhanced keratinocyte proliferation through YAP/TAZ activation, while concurrently suppressing fibroblast activity by upregulating IGFBP3 expression during the early phase of laser-induced skin remodeling. Our findings provide new insights into how laser treatments modulate cellular behavior and intercellular signaling, emphasizing the therapeutic potential of exosome-based strategies for promoting skin regeneration and rejuvenation.Graphical abstract Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12951-026-04709-3.Keywords: Laser, Skin extracellular matrix remodeling, Single-cell transcriptome sequencing, Keratinocytes, Fibroblasts, Exosomes, CEBPA Introduction Skin aging, driven by genetic and environmental factors, involves structural degradation and visible signs like wrinkles and pores [1]. Laser rejuvenation remodels the dermal matrix by inducing controlled tissue disruption, promoting repair and anti-aging [2, 3]. During the skin repair process, two key events are involved, which include the activation of keratinocytes to initiate the restoration process, the proliferation and migration of fibroblasts, and collagen production. To fully leverage the benefits of laser therapy for anti-aging and enhancing the precision and efficiency of treatment, it is essential to elucidate key events in skin tissue repair post-laser treatment. Some studies showed highly complex and heterogeneous nature of skin repair necessitates dynamic regulation and information transfer between keratinocytes and fibroblasts [4, 5]. Studies have shown that epidermal-derived factors stimulate collagen matrix contraction by regulating fibroblast numbers or altering their states [5, 6]. However, the intricate cellular specialization, interactions among different cell types, and the underlying mechanisms in response to the early stages of laser treatment remain underexplored. A deeper understanding is critical for facilitating rapid skin repair and maintaining a rejuvenated state following laser treatment. Dynamic cellular events following skin laser treatment depend on bidirectional intercellular communication to achieve effective skin rejuvenation. Traditionally, this crosstalk was thought to occur through paracrine signaling [7]. A series of researches highlighted extracellular vesicles (EVs) played a crucial role in intercellular communication [8, 9]. They are nanosized vesicles released by cells carrying biomolecules such as nucleic acids and proteins, serving as carriers for intercellular communication to regulate the behavior and function of recipient cells [10, 11]. Studies showed that keratinocyte-derived EVs played a crucial role in transforming myeloid cells into fibroblast-like cells within the granulation tissue at the injury site [12]. Other researches demonstrated that keratinocyte-derived exosomes deliver miRNAs that regulated macrophage numbers and functions within granulation tissue, playing a vital role in effective wound closure [13]. However, the precise mechanisms of cell-cell communication following laser treatment remain poorly understood. Recent advances in single-cell RNA sequencing (scRNA-seq) provide a powerful tool to uncover the cellular heterogeneity and mechanistic interactions involved in tissue [14, 15]. We selected pigs as model animals due to their close resemblance to human skin in terms of structure, thickness, physiology, and response to stimuli [16, 17]. Here, in present study we clarified the cell populations, the heterogeneity of the keratinocyte and fibroblast, and underlying cellular mechanisms by performing 10X scRNA-seq from laser-treated skin and normal skin. Analysis of the results revealed significant differences in both the number and functional properties of cells between the two groups. Additionally, cellular communication analysis highlighted a close interaction between keratinocytes and fibroblasts. Interestingly, cell coculture experiments revealed that keratinocytes and fibroblasts, influenced by laser treatment, engage in crosstalk mediated by keratinocyte-derived EVs with high expression of CEBPA. Mechanistically, exosomes enhanced keratinocyte proliferation via the YAP/TAZ axis, while concurrently suppressed fibroblast activation by upregulating IGFBP3 expression. These findings revealed a regulatory model in which early keratinocyte proliferation after laser treatment inhibited fibroblast activity while maintaining ECM homeostasis, contributing to improved skin remodeling and rejuvenation. Materials and methods Experimental design A porcine skin model was used to evaluate the phenotypic and transcriptional responses to laser treatment. Eight laser parameter settings were tested on the back skin of each pig. Each parameter was applied to a spatially separated skin region, resulting in eight independent treatment sites on the same animal. Skin samples at d 0 (immediately after treatment), d 7, d 14, d 30, and d 60 were collected from different pigs, with one pig assigned to each timepoint. At each timepoint, NC samples were collected from multiple spatially separated untreated dorsal skin regions of the corresponding pig. The entire treated skin square was excised and subdivided into matched portions for downstream analyses, including histological evaluation, RNA extraction, and protein analysis. During the initial phenotypic screening, parameter 7 showed a relatively more consistent tissue response at d 7 and was therefore selected for subsequent single-cell RNA sequencing. Experimental animals All animal experiments were approved by the Animal Welfare Committee of West China Hospital, Sichuan University and conducted in compliance with their ethical standards and guidelines (20240401011). Six female Large White pigs (50 kg) were housed under controlled conditions for seven days before the study, with free access to food and water and regular health monitoring. Laser treatments were performed under isoflurane inhalation anesthesia, and euthanasia was conducted following institutional ethical standards for sample collection. Detailed information on the laser treatment, sampling timepoint, and collected samples for each pig is provided in Table 1. Table 1Summary of animal use, sampling timepoints, and sample analysesIDSampling timepointLaser-treated samples collectedNC samples collectedSample analysesPig 1d 0Samples from the same eight-site dorsal laser-parameter mapMultiple spatially separated untreated dorsal skin regionsRT–qPCR, WB, Sirius red staining, immunofluorescencePig 2d 7Pig 3d 14Pig 4d 30Pig 5d 60Pig 6d 7Three spatially separated parameter-7-treated dorsal skin regionsThree spatially separated contralateral untreated NC regions10X scRNA-seq Summary of animal use, sampling timepoints, and sample analyses Laser treatment Laser treatments were performed using a long-pulse 1064 nm Nd: YAG laser (Cynosure, USA) and a picosecond laser (Cynosure, USA) as an auxiliary technique, with parameters divided into eight groups, which showed in Table 2. Each treatment covered a range of 10 × 10 mm area using a uniform sweeping method, with cooling air set to level 2 in mode II and the highest available frequency. Procedures were conducted under aseptic conditions, ensuring consistent laser energy, frequency, and treatment mode. Skin surfaces were sterilized before treatment, and post-treatment observations were recorded. All procedures were carried out by professionally trained personnel to guarantee standardized operations and the accuracy of the results. Table 2Recommendations for combined animal experiment parameters for laserGroupNo.Spot size (mm)Pulse width (ms)Energy (J/cm²)Shots (n)1064 nm laser112300403002105050300370.4153001064 nm laser + picosecond412/F6mm30040300/1000510/F6mm5050300/100067/F6mm0.415300/1000Combined laser7123004030010505030070.415300Combined + picosecond812/F6mm30040300/100010/F6mm5050300/10007/F6mm0.415300/1000 Recommendations for combined animal experiment parameters for laser Skin sample collection and cell isolation for scRNA-seq Skin samples were collected from six different regions of large white pig, with three regions from the laser-treated group and three for the untreated group. Fresh skin samples were thoroughly shaved to remove hair and subcutaneous fat, then cleaned with ice-cold 1× Dulbecco’s Phosphate Buffered Saline (DPBS). The samples were cut into small pieces of approximately 1 cm × 1 cm, placed in 4 mL centrifuge tubes, and transferred to tubes containing an enzyme mixture. The samples were dissociated using a gentleMACS Octo Dissociator for 120 min at 37 °C. After dissociation, the enzymatic reaction was terminated by adding 3 times the volume of DMEM/F12 complete medium. The cell suspension was filtered through a 70 μm MACS® SmartStrainer into a 50 mL centrifuge tube, then centrifuged at 300 rcf for 7 min at room temperature. The supernatant was discarded. Next, 1 mL of pre-cooled 1× Red Blood Cell Removal Solution was added, and the cells were gently mixed and resuspended. The cells were incubated at 4 °C for 10 min, centrifuged, and washed with a pre-cooled wash buffer. The final cell concentration was adjusted to 1000 ~ 2000 cells/µL. Finally, the cells were stained with 0.4% Trypan Blue Solution to evaluate cell viability, and cell counting was performed using the Countess® II Automated Cell Counter. Single-cell library construction and sequencing Single-cell transcriptome profiling was performed on approximately 10,000 cells per sample using the 10X Genomics 3’ Transcriptome Platform. Cells, enzymes, and gel beads were encapsulated in a microfluidic “double-cross” system, forming GEMs (Gel Beads-in-Emulsions). Within the GEMs, the gel beads dissolved, releasing the barcode sequences, while reverse transcription converted mRNA into cDNA. The cDNA served as a template for PCR amplification, generating sufficient material for library construction. These libraries were then subjected to high-throughput sequencing on the Illumina platform in PE150 mode by Genedenovo Biotechnology Co., Ltd (Guangzhou, China). Single-cell data processing and analysis Initial data processing was performed using Cell Ranger to ensure sequencing quality by removing low-quality reads, which aligned to the reference genome (Ensemble assembly: Sus scrofa11.1) to generate a transcript matrix. Subsequently, cells were further filtered based on multiple metrics using Seurat, including gene counts per cell, UMI counts, and the percentage of mitochondrial genes. Dimensionality reduction and clustering analysis were performed using Uniform Manifold Approximation and Projection (UMAP) and t-SNE. Cell annotation was conducted using SingleR, which compared expression patterns with reference datasets and validated cell identities with known marker genes and literature. Additionally, Seurat’s rank-sum test was applied to identify marker genes in different subpopulations, with log2FC ≥ 0.36 and P ≤ 0.01 as the screening criteria. Functional enrichment analyses Differentially expressed genes (DEGs) from various cell types were identified between the laser-treated and control groups using the criteria of |log2FC| > 0.36 and P < 0.01. These DEGs were further subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional enrichment analyses by the Genedenovo Cloud Platform (Guangzhou, China). Pseudotemporal analysis Cellular pseudotemporal trajectories were constructed using the R package Monocle2. This method organizes cells along a trajectory that represents the entire differentiation process, capturing the quasi-temporal sequence of cells as they progress through their asynchronous differentiation states. Cell cycle analysis Cell cycle scoring was performed using the R package Seurat, based on the expression levels of marker genes associated with each phase of the cell cycle. The cell cycle phase with the highest score was assigned as the cell’s current phase. Cells with scores below 0.3 for all phases were classified as non-cycling cells. Cellular communication analysis Since the CellChat database is based on human species, pig genes were screened for analysis based on their homology with human genes. The probability of intercellular communication was simulated by inputting single-cell gene expression matrices and cell population information, integrating gene expression levels with interactions between signaling ligands and receptors. Ligand-receptor interactions with a P-value < 0.05 were considered significant. The probability of communication at the signaling pathway level was inferred by calculating the communication probabilities for all pathways associated with ligand-receptor interactions. Subsequently, the weight and number of ligand-receptor interactions between cells were compared based on these calculated probabilities. CellChat provides a range of visualization images based on these data, including the number of ligand-receptor interactions, interaction weights, and the distribution of signaling pathways. Targeted transcription factors (TF) interaction among cells We utilized Single-cell Regulatory Network Inference and Clustering (SCENIC) software to perform transcription factor regulatory network analysis. SCENIC comprises three sub-modules: GENIE3, RcisTarget, and AUCell, which collectively enable the systematic analysis of regulons. The transcription factor list for porcine species was obtained from the AnimalTFDB database. By analyzing the regulon activity in individual cells, SCENIC facilitates the identification of differences in transcription factor regulatory activity across various experimental groups and cell types. Cell co-culture Human keratinocytes (HaCaT) and human dermal fibroblasts (HDF) were used in the cell experiments. Keratinocytes were seeded in the upper chambers of 12-well, 24-well, and 96-well plates, while either keratinocytes or fibroblasts were seeded in the lower chambers. The upper-layer keratinocytes were treated with a 1064 nm laser at energy density of 60 J/cm², followed by incubation under hypoxic conditions for 4 h. After 24 h, the effects of laser-treated keratinocytes on the lower-layer keratinocytes or fibroblasts were assessed. Co-culture chambers without any treatment were used as the control group. RNA extraction and quantitative reverse-transcription PCR Total RNA was extracted using TRIzol reagent according to the manufacturer’s specifications. The quality and purity of the extracted RNA were assessed using ‌NanoDrop (Thermo Scientific, MA, USA). Reverse transcription of mRNA was performed using commercial kits strictly following the manufacturer’s protocol. Specific primers for amplifying fragments of selected genes were designed using NCBI Primer-BLAST. All samples were amplified in triplicate, and relative changes in gene expression levels were analyzed using ΔΔCt method. Detailed reagents, consumables, and software were listed below in Table 3. Table 3The detailed reagents, consumables, and softwareReagent or resourceSourceIdentifierAntibodiesRabbit anti-HSP47NovusbioCat#:NBP1-97491Rabbit anti-HSP70ProteintechCat #:10995-1-APRabbit anti-Collagen ⅠNovusbioCat #:NB600-408Rabbit anti-Col ⅢAbcamCat #:Ab184993Rabbit anti-DecorinAbcamCat #:ab277636Rabbit anti-MMP1AbcamCat #:Ab137332Rabbit anti-MMP3BiorbytCat #:Orb12895Rabbit anti-MMP9ProteintechCat #:10375-2-APRabbit anti-CD36ProteintechCat #:25682-1-APRabbit anti-TSG101ProteintechCat #:28283-1-APRabbit anti-CEBPABosterCat #:A00386-1Rabbit anti-YAP1ProteintechCat #:13584-1-APRabbit anti-TAZProteintechCat #:23306-1-APRabbit anti-IGFBP3ProteintechCat #:10189-2-APRabbit anti-GAPDHProteintechCat #:10494-1-APDeposited datasc-RNA-seq raw dataThis paperGenome sequence archive: CRA019722Critical commercial assaysMultiplex fluorescence immunohistochemical staining kitAbsinCat #:abs50014Tissue/Cell Total RNA isolation kitVazymeCat #:RC113-01HiScript IV RT SuperMix for qPCRVazymeCat #:R423-01Universal SYBR qPCR master mixVazymeCat #:Q712-02/03EdU cell proliferation kitBeyotimeCat #:C0071SReactive OxygenSpecies assay kitBeyotimeCat #:S0033SFluo-4 calcium assay kitBeyotimeCat #:S1061SCell counting kit-8BeyotimeCat #: C0038BCA protein assay kitBeyotimeCat #: P0010VerteporfinTargetMolCat #: T3112Chemicals, peptides, and recombinant proteinsSirius red stainingSbjbioCat #:BP-DL029MEM mediumGibcoCat #:11,095,080Fetal bovine serumExCellCat #:FSP5000.25% TrypsinHyCloneCat #:SH30042DMEM mediumGibcoCat #: 6,124,230Collagenase type ISigmaCat #: SCR103Collagenase type IISigmaCat #: C2-BIOCCollagenase type IVSigmaCat #: C4-BIOCCollagenase type VSigmaCat #: C9263ElastaseCoolaberCat #: CE5001DNase ICoolaberCat #: SL20761Oligonucleotidessi-IGFBP3TsingkeCustom synthesizedNegative control siRNA (si-NC)TsingkeCustom synthesizedSoftware and algorithmsSTARDobin et al. [18] https://github.com/alexdobin/star R packageR core team https://www.R-project.org/ Monocle2Trapnell et al. [19]https://github.com/cole-trapnell-lab/ monocle-releaseSeuratHao et al. [20] https://github.com/satijalab/seurat SCENISAibar et al. [21] https://github.com/aertslab/scenis Cell chatJin et al. [22] http://www.cellchat.org/ The detailed reagents, consumables, and software Reactive Oxygen Species assay kit Protein extraction and western blot analysis Proteins from EVs and approximately 7 × 10⁶ cells were extracted using Protein Extraction Kit (Servicebio, Wuhan, China) according to the manufacturer’s instructions. Protein concentrations were measured using the BCA Assay Kit (Beyotime, Shanghai, China). Total proteins were denatured by 5× SDS loading buffer. PVDF membranes were incubated with primary antibodies specific to the target proteins at 4 °C overnight. Secondary antibodies were applied and incubated for 1 h at room temperature with gentle shaking. Protein bands were visualized using ECL solution (Servicebio, Wuhan, China) and were captured images by ChemiDoc™ MP Imageing System. Expression levels were analyzed using a gel imaging system, quantified with ImageJ, and normalized to GAPDH expression as an internal control. H&E staining Fixed tissue sections were deparaffinized, rehydrated, and stained with hematoxylin solution for 5 min to visualize the nuclei. The sections were then rinsed with distilled water and stained with eosin solution for 1 min to color the cytoplasm. Finally, images were captured using a VS-200 full-spectrum section scanning instrument. Multiple immunofluorescence staining Fixed tissue sections were subjected to antigen retrieval using enzymatic methods to expose antigenic sites. The sections were then blocked with a blocking solution, and incubated at room temperature for 30 min to minimize non-specific binding. Next, a mixture of primary antibodies was added and incubated overnight at 4 °C. Following incubation, the samples were washed with PBS, and a mixture of fluorescently labeled secondary antibodies was added. The sections were incubated for 60 min in the dark. Then, DAPI was used to stain the nuclei. Finally, the sections were sealed with an anti-fade mounting medium, and the staining results were observed and imaged using a confocal microscope. Sirius red staining Tissue samples were fixed, dehydrated, paraffin-embedded, sectioned, dewaxed, and rehydrated. The sections were then stained with Sirius red solution, followed by washing with distilled water, dehydration, and transparent processing. Finally, the sections were sealed with neutral gum and images were captured using a VS-200 full-spectrum section scanning instrument. Immunofluorescence staining Samples were fixed with 4% paraformaldehyde and permeabilized with 0.5% Triton X-100. After blocking with 1% BSA to reduce non-specific binding, samples were incubated with primary antibodies overnight at 4 °C. Then, fluorescently labeled secondary antibodies were applied and incubated for 1 h in the dark. Finally, nuclei were stained with DAPI, and images were captured using a fluorescence or confocal microscope. CCK-8 cell viability assay The viability of treated cells was evaluated using the CCK-8 kit (LABLEAD, Beijing, China). Briefly, 10 µL of CCK-8 working solution was added to each well containing treated cells, followed by incubation at 37 °C for 1 h. The absorbance at 450 nm was then measured to compare the experimental and control groups. EdU cell proliferation assay The EdU solution was prepared and diluted according to the instructions of the EdU detection kit (Beyotime, Shanghai, China). After incubating the cells with the EdU solution for 2 h, the supernatant was removed, and the cells were fixed. Subsequent staining was performed following the steps outlined in the kit’s instruction manual. Finally, a fluorescent microscope was used to visualize the staining and capture the images, and these EdU results were analyzed statistically using ImageJ software. Cell scratch assay A scratch was created in the cell monolayer in the lower chamber, and images were captured at 0 h and 24 h by fluorescent microscope. The scratched area was then statistically analyzed using ImageJ software to assess the extent of cell migration. ROS staining The DCFH-DA solution was diluted with cell culture medium at a ratio of 1:1000 to achieve a final concentration of 10 µmol/L. After removing the cell culture medium, 500 µL of the diluted DCFH-DA solution was added to the cells, followed by incubation at 37 °C for 20 min in cell culture incubator. The cells were then washed three times with PBS to thoroughly remove any excess DCFH-DA that had not entered the cells. Subsequently, the cells were fixed, and nuclear staining was performed using DAPI. Fluorescent signals were visualized and captured using a fluorescent microscope. Ca2+ staining Prepare a 2 µM Fluo-4 AM working solution by diluting the Fluo-4 AM master mix with PBS. For the cultured cells to be analyzed, remove the culture medium and wash the cells three times with PBS. Add 500 µL of the Fluo-4 AM working solution, ensuring the solution fully covers the cells. Incubate the cells at 37 °C for 60 min, then fix the cells and perform nuclear staining using DAPI. Fluorescent signals were visualized and captured by fluorescent microscope. Exosome extraction Culture supernatants from laser-treated and untreated keratinocytes were collected, and exosomes were isolated and purified using sequential centrifugation protocol. First, the supernatants were centrifuged at 300 ×g for 10 min at 4 °C, and the supernatants were collected. Next, the collected supernatants were centrifuged at 2,000 ×g for 10 min at 4 °C, followed by another centrifugation at 10,000 ×g for 30 min at 4 °C, with the supernatants retained after each step. Finally, the retained supernatants underwent ultracentrifugation at 120,000 ×g for 90 min at 4 °C, and this step was repeated twice to ensure purity. The isolated exosomes were resuspended in PBS and stored at -80 °C for future study. Transmission electron microscopy (TEM) Samples were fixed in 2.5% glutaraldehyde for 2 h, followed by fixation in 1% osmium tetroxide for another 2 h. After fixation, the samples were dehydrated through a graded ethanol series and finally with pure acetone. The dehydrated samples were then embedded in resin for osmotic infiltration and polymerized in an oven at 60 °C. After polymerization, the samples were sectioned into 50–90 nm slices using an ultrathin microtome and placed onto copper grids. The sections were stained with uranyl acetate and lead citrate before being observed and imaged under TEM. Nanoparticle tracking analysis (NTA) EV samples were diluted with PBS. The diluted samples were injected into the NTA detector and analyzed at a controlled temperature. A laser beam illuminated the particles, and their Brownian motion was captured by a camera. The recorded data were subsequently analyzed to obtain the NTA results. Phagocytosis assay Exosomes were labeled with DiD dye following the manufacturer’s instructions. The DiD-labeled exosomes were co-incubated with target cells for 24 h at 37 °C in 5% CO₂ incubator. The cells were subsequently fixed, and nuclear staining was performed using DAPI. Finally, the samples were observed and imaged using a confocal microscope. Statistical analysis All data analyses were conducted using GraphPad Prism software. Independent samples were analyzed using Student’s t-test, and all experimental data are presented as mean ± standard error (Mean ± SEM). Statistical significance was defined as *P < 0.05, **P < 0.01, ***P < 0.001. Result Quantitative analysis of molecular markers and histological observations in the laser-treated skin Using a clinically relevant porcine model, we applied eight different laser treatment parameters to evaluate skin regeneration. As an immediate response to laser exposure, varying degrees of redness and bruising were observed across the different parameters (Fig. 1A). Parameters 1, 2, 3, 5, and 6 produced mild erythema, whereas parameters 4, 7, and 8 induced more pronounced erythema, with localized bruising observed in a few areas. To investigate the molecular and structural changes, skin samples were collected at d 0, d 7, d 14, d 30, and d 60 post-treatment for histological and molecular markers analysis (Fig. 1B). Gene expression analysis using RT–qPCR showed that, compared with the corresponding time-matched NC group, COL1A1, COL1A2, COL3A1, ELN, DCN, and FN1 were significantly downregulated on day 7, followed by upregulation on days 14 and 30, with some parameters maintaining this effect until day 60 (Fig. S1A). WB analysis revealed parameter- and time-dependent changes in collagen I/III and MMP1 expression after laser treatment (Fig. 1C and Table S1). Collagen I decreased in most treatment groups at day 7 and increased in most groups at days 14 and 30, whereas collagen III remained relatively stable at days 7 and 14 but increased more consistently at later time points. MMP1 showed early parameter-dependent changes, suggesting dynamic ECM remodeling after laser treatment. Histological analysis further supported these findings. H&E staining showed that different laser parameters induced varying degrees of tissue denaturation at d 0 and d 7 (Fig. 1D). Among them, parameter 7 exhibited a more consistent histological change across these two timepoints, indicating a representative early tissue response to laser exposure. Furthermore, we performed Sirius red staining and immunofluorescence at different time points in laser parameter 7 for laser treatment and normal skin samples. Sirius red staining suggested time-dependent collagen remodeling after laser treatment (Fig. 2A). Compared with the corresponding time-matched NC regions, laser-treated regions showed relatively reduced collagen signals at day 7, followed by increased collagen deposition and more organized collagen fibers at days 14 and 30. Moreover, immunofluorescence analysis showed time-dependent changes in collagen I/III and DCN signals after laser treatment. Compared with the corresponding time-matched NC samples, collagen I/III and DCN signals showed dynamic changes across d 7, d 14, and d 30, suggesting collagen-associated remodeling during tissue repair (Fig. 2B and Fig. S1B). Notably, MMP9 was markedly increased at day 30, indicating the transition to tissue remodeling (Fig. S2). Collectively, the morphological and quantitative analyses suggested that multiple laser parameters induced time-dependent skin remodeling changes. However, the cellular dynamics and intercellular communication mechanisms underlying these effects remain unclear. Although porcine skin has been widely used in laser-related studies, most previous reports have focused on ex vivo skin, wound/scar models, or microablative tissue responses, whereas systematic evaluation of regeneration-related remodeling after laser treatment in intact in vivo porcine skin remains limited. Effects of laser treatment on skin tissue in porcine model. (A) Illustration of laser treatment applied to the porcine model. Scale bar, 0.5 cm. (B) Schematic showing experimental workflow following laser treatment in porcine model. (C) Western blot analysis of related proteins at different time points and under varying laser parameters. (D) H&E staining of skin samples treated with different laser parameters on day 0 and day 7 post-treatment (n = 3). Scale bar, 20 μm. Statistical comparisons were performed using the total denaturation score. Data are presented as mean ± SEM. Statistical significance was determined using Student’s t-test. *P < 0.05, **P < 0.01, ***P < 0.001 Histological and immunofluorescence analysis of skin tissue after laser treatment. (A) Sirius red staining of skin samples collected at different time points of laser parameter 7 post-laser treatment (n = 3). Scale bar, 20 μm. (B) Multiplex immunofluorescence staining of related genes in skin samples from parameter 7 at different time points post-laser treatment (n = 3). Scale bar, 100 μm. Data are presented as mean ± SEM. Statistical significance was determined using Student’s t-test. *P < 0.05, **P < 0.01, ***P < 0.001 A census of single cells from laser-treated and normal skin shows the diversity of cell populations Given the pronounced early tissue response observed in region 7 on day 7, characterized by marked structural and molecular changes after laser treatment, this condition was selected for further analysis using 10X scRNA-seq (Fig. 3A). The resulting quality-controlled single-cell atlas obtained 51,068 cells, with an average of 2,123 genes and 4,661 unique molecular identifier (UMI) counts per cell (Fig. S3A and S3B). Additional quality-control and integration analyses showed overall consistency among the six spatially separated scRNA-seq samples (Fig. S3C and S3D). All cells were identified into 24 cell clusters using uniform manifold approximation and projection (UMAP) (Fig. 3B). The 24 cell clusters were identified based on the expression matrix of marker genes and are shown in the heatmap for each cluster (Fig. 3C), revealing unique transcriptional features. These cell clusters categorized into nine cell types [23] including smooth muscle cells (SMCs), endothelial cells (ECs), keratinocytes (KCs), fibroblasts (FBs), glia cells (GCs), secretory cells (SCs), mast cells (MTCs), macrophages (MACs) and T cells (TCs) (Fig. 3D and E and Table S2). The total distribution and number of cell types were comparable between laser-treated and normal skin (Fig. 3F and G and S3E). While the overall cell distribution was comparable in both groups, significant changes were observed in keratinocytes and fibroblasts. GO enrichment analysis of marker genes identified distinct differences in enriched pathways among various cell types between the two groups (Fig. S3F), indicating substantial functional heterogeneity in the skin microenvironment after laser treatment. Specifically, laser-treated keratinocytes were enriched in pathways related to extracellular matrix organization, whereas fibroblasts were enriched in pathways associated with negative regulation of cell proliferation. Single-cell atlas in laser-treated and untreated skin samples. (A) Overview of the workflow for single-cell RNA sequencing. (B) The t-SNE plot visualization of 24 cell clusters. (C) Heatmap displaying the high-expression genes in each cell cluster. (D) Dot plot illustrating representative genes for each cell type. (E) The t-SNE plots showing the distribution of marker genes across cell types. (F) UMAP depicting the distribution of cell types in the laser-treated (LA) and untreated (NC) groups, respectively. (G) Stacked area plot comparing cell numbers across cell types in both groups Following laser treatment, we further investigated transcriptomic alterations by performing differential expression analysis across cell types between the laser-treated and untreated groups using the integrated single-cell RNA-seq data (Fig. 4A and Table S3). The gene set enrichment analysis (GSEA) revealed the biological processes enriched in the DEGs of each cell type (Fig. 4B and S4A-E). SMCs showed enrichment in intermediate filament and positive chemotaxis, suggesting an injury-responsive state characterized by cytoskeletal remodeling and chemotactic responsiveness during the early phase after laser treatment. Keratinocytes were enriched in proliferation pathways, critical for re-epithelialization and restoring epidermal integrity. Fibroblasts were enriched in diversity pathways, reflecting their essential role in ECM remodeling, immune regulation, and signaling modulation. Moreover, the upregulated and downregulated DEGs in different cell types were primarily enriched in distinct KEGG pathways (Fig. S5A and S5B). Moreover, the upregulated DEGs in different cell types were mainly enriched in pathways related to signal transduction, inflammatory communication, and tissue remodeling, whereas the downregulated DEGs were primarily associated with metabolic regulation and ECM interaction, suggesting early microenvironment remodeling after laser treatment. Differential gene expression, cell communication and transcription factor analysis in laser-treated and untreated skin. (A) Multi-panel volcano plots displaying differentially expressed genes (DEGs) across cell types in both groups. (B) The GSEA enrichment analysis of the DEGs from SMCs, ECs, KCs, and FBs. (C) Circle plot illustrating differential interaction strength among cell types. (D) Heatmap presenting the ligand-receptor interaction strength and quantity between cell types. (E) Bar chart comparing the relative information flow of signaling pathways in both groups. (F) Heatmap showcasing the overall signaling patterns in both groups. (G) Heatmap of regulon activity comparing in both groups. Red indicates relatively activated regulon activity, whereas blue indicates relatively reduced regulon activity. (H) Heatmap displaying transcription factor activity across different cell types Subsequently, we identified 51 signaling pathways mediating intercellular interactions across all cell populations, with fibroblast-keratinocyte interactions being the most prominent (Fig. 4C and D and Table S4). Laser-treated skin showed upregulation of key signaling pathways including OCLN, CDH1, RELN, NRG, EGF, VEGF, NOTCH, and CADM (Fig. 4E and Table S5), which regulate cell adhesion, proliferation, and tissue repair [24, 25, 26]. Further comparison between LA and NC revealed cell-type-specific changes in selected signaling pathways. Notably, ncWNT [27], CSF [28], and NECTIN [29] signaling showed increased activity in fibroblasts after laser treatment (Fig. 4F). Lastly, transcriptional regulatory network analysis revealed significant upregulation of ATF3, KLF4, KLF7, KLF10, EGR1, FOSL2, and HIF1A in laser-treated skin (Fig. 4G), which regulate cell stress adaptation, proliferation, immune and tissue repair [30, 31]. Moreover, cell type-specific analysis further identified KLF3, KLF10, FOSB, and HIF1A as key transcription factors in keratinocytes, KLF4, ZNF148, and EGR1 played roles in ECM remodeling in fibroblast, while JUN, FOS, KLF family members, and STAT2 as critical players in immune regulation and vascular remodeling in endothelial cells (Fig. 4H). The differential activation of molecular mechanisms promotes the diverse functions of different cell types in early-stage laser-induced skin, ultimately driving skin repair, immune responses, and tissue remodeling in a coordinated manner. Heterogeneity and dynamic changes of keratinocyte subclusters The role of keratinocytes during the early stages of laser treatment is primarily focused on immune regulation, barrier restoration, and epidermal repair. To explore the heterogeneity of keratinocytes, UMAP analysis categorized all keratinocytes into seven distinct subpopulations (Fig. 5A). The relative composition of keratinocyte subclusters changed after laser treatment, with several subclusters showing increased proportions in the laser-treated group compared with the NC group (Fig. 5B). Based on previous studies [32] and the expression of known marker genes, we classified keratinocytes into three functional subpopulations: basal keratinocytes (clusters 0, 3, 4, 5; KRT15, SLC2A1, BGN, KRT5), spinous keratinocytes (cluster 1; KRT1, KRT10), and EMT keratinocytes (clusters 2, 6; VIM, DCN, COL3A1) (Fig. 5C). The laser-treated group showed a significant increase in basal, suggesting enhanced proliferation and upward migration for epidermal reconstruction. GO analysis further highlighted functional differences among keratinocyte subclusters (Fig. 5D). Clusters 0, 3, 4, and 5 were mainly associated with translation, proliferation, and adhesion, whereas cluster 1 was enriched in migration-and adhesion-related processes, and clusters 2 and 6 showed enrichment of cytoskeletal remodeling, and integrin-mediated adhesion. Moreover, Pseudotime trajectory analysis showed that basal keratinocytes were at the starting point, differentiating into spinous keratinocytes and EMT keratinocytes (Fig. 5E). Gene expression dynamics along the pseudotime trajectory were organized into several modules with distinct temporal patterns (Fig. 5F). Early pseudotime modules were mainly associated with fatty acid metabolism, cell adhesion, cell proliferation, skin barrier establishment, and keratinocyte differentiation, whereas later modules were associated with interferon response, EGFR regulation, cell migration, cell-cell adhesion, and skin morphogenesis, indicating progressive functional changes during keratinocyte state transition. Consistently, cell cycle analysis showed that KC0 and KC1 were predominantly in S and G2/M phases, supporting their active proliferative state (Fig. 5G). Heterogeneity and dynamic changes of keratinocyte subclusters. (A) UMAP showing keratinocyte subclusters. (B) Histogram showing the distribution of cell number of subclusters. (C) Heatmap of marker gene expression across subclusters. (D) The enriched GO terms of biological process for each subcluster sorted by q-value. (E) Pseudotime ordering of subclusters using Monocle 2. (F) Heatmap illustrating the expression dynamics of genes associated with biological processes along the pseudotime trajectory. (G) Proportion of cells in different cell cycle phases (G2/M, S, and non-cycling) across subclusters. (H) Circular heatmap showing the distribution of gene expression across subclusters Subsequently, we further analyzed the expression patterns of cytokine-related genes across keratinocyte subpopulations (Fig. 5H and Table S6). IL1A was highly expressed in KC0, whereas IL1R2 was enriched in KC1. IL18 and its receptor were mainly detected in basal keratinocytes. Growth factor-related genes showed differential but overlapping expression across keratinocyte states: TGFB3 and PDGFA were relatively enriched in EMT keratinocytes, whereas TGFB1, TGFB2, and PDGFB were also expressed in basal keratinocytes. FGF-related genes were mainly observed in basal keratinocytes. In addition, CCL-related genes were more prominent in EMT keratinocytes, whereas CXCL-related genes were distributed across both basal and EMT keratinocytes. Identification and characterization of fibroblast subpopulations Fibroblasts are essential for maintaining and repairing tissue homeostasis [33]. Single-cell transcriptome analysis identified four fibroblast subpopulations (Fig. 6A). A comparison of fibroblast subclusters between the two groups showed that all fibroblast clusters displayed a downward trend at the early stage after laser treatment (Fig. 6B and S5C). Based on previously published marker genes for skin fibroblasts [23, 34], FB1 and FB4 were classified as reticular fibroblasts (COL1A1, COL3A1, COL5A1, ELN, MGP, MFAP5), FB2 as papillary fibroblasts (COL6A5, COLEC12, AXIN2, PDGFRA), and FB3 as mesenchymal fibroblasts (COL11A1, TNN, SFRP1) (Fig. 6C). While no distinct inflammatory fibroblast subpopulation was identified, FB2 and FB3 showed enrichment of inflammation-related genes (CXCL12, CCL19, CCL2), and FB3 predominantly expressed ECM-related genes (MMPs, ADAMTSs, CTSS) (Fig. 6D). GO enrichment analysis revealed distinct roles for each subtype (Fig. 6E). Interestingly, vesicle-mediated transport was enriched across all fibroblast subpopulations. Further analysis of the transcriptomic changes in fibroblast subpopulations after laser treatment revealed a significant non-overlap of DEGs among these subpopulations, indicating highly specific responses (Fig. 6F). GSEA showed that FB1 was enriched in oxidative stress-related pathways, while FB2 and FB3 were associated with epidermal structure, intermediate filament organization, and immune regulation. FB4 exhibited enrichment in ion channels and stimulus perception, cell cycle regulation and chromosome-related processes (Fig. 6G). Identification and characterization of fibroblasts subpopulations. (A) UMAP visualization of fibroblast subpopulations. (B) Bar plot showing the proportion of fibroblasts in both groups. (C) Dot plot representing the expression levels of marker genes for each fibroblast subpopulation. (D) Dot plot showing the expression of chemokines, inflammatory markers, and matrix-related genes across fibroblast subpopulations. (E) GO enrichment analysis of biological processes across fibroblast subpopulations. (F) Heatmap showing the DEGs of fibroblast subpopulations. (G) The GSEA enrichment analysis of DEG in each fibroblast subpopulation Laser-induced keratinocyte-fibroblast crosstalk and exosome-mediated signaling in vitro Our analysis revealed that keratinocytes and fibroblasts exhibit the closest communication following early laser treatment. To further explore the dynamics of this interaction, we established a simulated co-culture system under a hypoxic condition, where laser-treated keratinocytes were placed in the upper chamber and fibroblasts or keratinocytes in the lower chamber (Fig. 7A). Under laser stimulation, compared to control group, fibroblasts showed downregulation of COL1A1, ELN, FN1, and DCN, while keratinocytes exhibited increased FLG, CDA1, MMP9, and IL1A expression (Fig. S6A). CCK-8 assays indicated a significant increase in keratinocyte numbers, while fibroblast numbers were reduced (Fig. S6B). In the scratch assay, fibroblast migration was significantly inhibited, whereas keratinocyte migration was accelerated (Fig. 7B). Furthermore, EdU staining showed that fibroblast proliferation was suppressed, whereas keratinocyte proliferation was promoted (Fig. 7C). Immunofluorescence staining demonstrated elevated levels of ROS and Ca²⁺ in fibroblasts, indicating cellular stress and fibroblast activity in response to laser stimulation and hypoxic conditions (Fig. 7D-E). These results suggest that during the early phase of laser treatment, keratinocytes positively regulate their own proliferation and migration to accelerate epidermal barrier repair. Simultaneously, keratinocytes appear to negatively regulate fibroblast activity, likely through elevated ROS and Ca²⁺ levels, which impair fibroblast function. In many studies, EVs have been reported as key mediators of intercellular communication [8]. Consistently, our single-cell analysis of fibroblasts showed enrichment of vesicle-mediated transport-related functions, providing a potential clue that EV-mediated signaling may participate in intercellular communication after laser treatment. Laser-induced keratinocyte-fibroblast crosstalk and exosome-mediated signaling in vitro. (A) Schematic diagram of co-culture model. (B) Scratch assay images of fibroblasts and keratinocytes in both groups at 0 h and 24 h. Scale bar, 200 μm. (C) EdU images of fibroblasts and keratinocytes in both groups. Scale bar, 200 μm. (D-E) ROS/Ca2+ staining showing fibroblasts in both groups. Scale bar, 200 μm. (F) Transmission electron microscopy images of EV_treat and EV_NC. Scale bar, 100 nm. (G) Nanoparticle tracking analysis of EV_treat and EV_NC. The value of 9.1 × 10¹¹ particles/mL represents the peak particle concentration observed in the EV_NC group. (H) Western blot analysis of marker proteins in EV_treat and EV_NC. (I) The observation of colocalization of Did-labeled EV_treat with keratinocytes. Scale bar, 50 μm. (J) Scratch assay images of fibroblasts and keratinocytes treated with EV_treat or EV_NC at 0 h and 24 h. Scale bar, 200 μm. (K-L) EdU/ ROS staining images of fibroblasts treated with EV_treat or EV_NC. Scale bar, 200 μm. Data are presented as mean ± SEM. Statistical significance was determined using Student’s t-test. n = 3. *P < 0.05, **P < 0.01, ***P < 0.001 To explore the role of EVs as signaling mediators, we isolated EVs from the supernatants of laser-treated and hypoxia-stimulated keratinocytes (EV_treat) and untreated keratinocytes (EV_NC, serving as control). Characterization of EVs revealed that both EV_treat and EV_NC exhibited typical teacup-like morphology under electron microscopy (Fig. 7F). Nanoparticle tracking analysis (NTA) showed that EV_treat had a peak diameter of 125.3 nm, while EV_NC peaked at 106 nm (Fig. 7G). WB analysis confirmed exosomal markers CD63, TGS101, and HSP70 (Fig. 7H). Immunofluorescence further demonstrated that Did-labeled EVs uptake by keratinocytes and fibroblasts (Fig. 7I and S6C). To confirm the functional effects of keratinocyte-derived EVs, keratinocytes and fibroblasts were treated separately with EV_treat (laser and hypoxia-stimulated EVs) and EV_NC, respectively. Compared to the control, EV_treat significantly downregulated the expression of COL1A1, ELN, and FN1 in fibroblasts, and led to the upregulation of FLG, CDA1, and MMP9 in keratinocytes (Fig. S6D). Scratch assays demonstrated that fibroblast migration was significantly inhibited, while keratinocyte migration was accelerated in the EV-treat group, comparing with EV_NC (Fig. 7J). EdU assays further confirmed the significant increase in proliferating keratinocytes (Fig. 7K). Additionally, immunofluorescence staining revealed elevated ROS and Ca²⁺ levels in EV-treat fibroblasts, consistent with the findings observed in the co-culture system (Fig. 7L and S6E). These results underscore the critical role of keratinocyte-derived EVs in regulating the functional dynamics of keratinocytes and fibroblasts during laser-induced skin repair. CEBPA-containing exosomes secreted by laser-treated keratinocytes mediate the keratinocyte and fibroblast activation To investigate EV-mediated intercellular communication during early laser treatment, we analyzed DEGs in keratinocytes and fibroblasts. In keratinocytes, KEGG analysis provided potential functional clues for the DEGs, with several pathways related to metabolism and Hippo signaling appearing among the enriched terms (Fig. 8A), which regulates YAP/TAZ activation and promotes proliferation and migration [35]. RT–qPCR analysis of key genes within this pathway showed a significant upregulation of YAP1/TAZ in the laser-treated keratinocytes and EV-treat keratinocytes (Fig. S7A). Consistently, western blot analysis showed that total YAP1 expression remained largely unchanged after laser treatment or EV_treat exposure, whereas p-YAP levels showed a decreasing trend, particularly after laser treatment (Fig. S7B and Table S1). Further validation of selected keratinocyte DEGs with relatively large fold changes and high expression levels showed significant upregulation of CEBPA, KRT2, and GJB2 mRNA levels (Fig. 8B). Notably, CEBPA is a well-established transcription factor, that regulated targeted genes transcription [36]. In fibroblasts, KEGG analysis of DEGs revealed enrichment in pathways such as cGMP-PKG signaling, Hedgehog signaling, Wnt signaling, axon regeneration, and cellular senescence (Fig. 8C). Moreover, the RT–qPCR data of DGEs revealed a significant upregulation of IGFBP3, a key regulator in regulating cell proliferation and growth arrest [37], which was involved in the cellular senescence pathway [38] (Fig. 8D). Motif analysis identified the canonical CEBPA-binding sequence (Fig. 8E), and promoter sequence analysis further revealed putative CEBPA-binding motifs in the regulatory regions of YAP1, TAZ, and IGFBP3 (Table S7), suggesting that CEBPA may potentially participate in their transcriptional regulation. Combined with previous studies, CEBPA has also been shown to exhibit antifibrotic properties in pulmonary fibrosis and keloids [39, 40]. Based on this evidence, we hypothesized that during the early laser phase, keratinocytes released CEBPA-enriched EVs, which simultaneously promote keratinocyte proliferation through YAP1/TAZ activation while inhibiting fibroblast activity through IGFBP3 upregulation. CEBPA-containing exosomes secreted by laser-treated keratinocytes mediate the keratinocytes and fibroblasts activation. (A) KEGG analysis of the DEGs of keratinocytes between the treated and untreated groups. (B) Relative mRNA expression levels of genes in treated and untreated keratinocytes. (C) KEGG analysis of the DEGs of fibroblasts between the treated and untreated groups. (D) Relative mRNA expression levels of genes in treated and untreated fibroblasts. (E) Motif sequence of CEBPA-binding sites. (F) Immunofluorescence staining for CEBPA expression levels in fibroblasts (right) and keratinocytes (left) after EV_treat or EV_NC treatment. Scale bar, 200 μm. (G) CCK-8 assay results illustrating the cell viability in keratinocytes and fibroblasts treated with overexpression CEBPA or control. (H) Cell cycle analysis showing cell cycle distribution in keratinocytes and fibroblasts treated with overexpression CEBPA or control. (I) Scratch assay images for keratinocytes and fibroblasts treated with overexpression CEBPA or control at 0 h and 24 h. Scale bar, 200 μm. Immunofluorescence staining of YAP1 (J) and TAZ (L) in keratinocytes treated with overexpression CEBPA or control. Scale bar, 200 μm. (K) Representative confocal images showing the subcellular localization and nuclear translocation of YAP1 in keratinocytes under the indicated treatments. Nuclei were stained with DAPI. Scale bars, 20 μm. (M-N) Western blot analysis of related protein levels in keratinocytes/fibroblasts treated with overexpression CEBPA or control. Data are presented as mean ± SEM. Statistical significance was determined using Student’s t-test. n = 3. *P < 0.05, **P < 0.01, ***P < 0.001 To validate this hypothesis, we examined the expression of CEBPA in keratinocyte-derived EVs. WB results showed that CEBPA expression was significant upregulated in EV_treat group compared to EV_NC group (Fig. S7C). Subsequently, EV_treat was applied to keratinocytes and fibroblasts, immunostaining confirmed that CEBPA expression was significantly increased in both cell types compared with EV_NC (Fig. 8F). To confirm the functional role of CEBPA, we constructed a CEBPA overexpression vector and performed overexpression experiments in keratinocytes and fibroblasts. CCK-8 and cell cycle assays demonstrated that CEBPA overexpression significantly promoted keratinocyte proliferation ability, while showed inhibition of fibroblast proliferation (Fig. 8G and H). And scratch assays further showed prominently inhibited fibroblast migration but enhanced migration ability in keratinocytes (Fig. 8I). To investigate whether CEBPA functions through YAP/TAZ signaling, immunofluorescence and WB assays were performed in keratinocytes. The results showed that CEBPA overexpression affected YAP/TAZ-related signaling, characterized by reduced YAP phosphorylation and increased YAP accumulation (Fig. 8J, L and M and Table S1). EV_treat also altered YAP/TAZ-related protein expression patterns in keratinocytes, as shown by WB analysis (Fig. S7D and Table S1). Moreover, immunofluorescence revealed enhanced nuclear localization of YAP (Fig. 8K), supporting the involvement of YAP1/TAZ signaling in promoting keratinocyte activity. However, after addition of a YAP inhibitor to the CEBPA-overexpression and EV_treat groups, keratinocyte migration and proliferation were significantly reduced (Fig. S7E and S7F), further indicating that YAP1/TAZ signaling contributes to the effects of CEBPA/EV_treat on keratinocyte behavior. Likely, WB analysis further showed that CEBPA overexpression and EV_treat suppressed Collagen I expression while upregulating IGFBP3, indicating its role in fibroblast regulation (Fig. 8N, S7G and Table S1). Notably, when IGFBP3 was silenced, fibroblast proliferation and migration were significantly increased, further indicating that IGFBP3 mediates the inhibitory effects of CEBPA/EV_treat on fibroblast (Fig. S7H and S7I). These findings provide deeper insights into the molecular mechanisms of laser therapy. Discussion Laser treatments are widely used for skin rejuvenation, including improving skin texture and reducing visible signs of aging [41]. Previous research has shown that the 1064 nm Nd: YAG laser effectively focuses energy on the deep dermis while picosecond lasers target the epidermis and superficial dermis, which both have been reported to produce significant improvements in facial aging [42, 43, 44]. In this study, we used a porcine skin model to explore the effects of various laser parameters. Parameter 7 showed the most significant effect, likely due to parameter 8 causing a “dispersive effect” that reduced deep dermal thermal stimulation. Additionally, individual differences in tissue structure may play a role. Although porcine skin closely resembles human skin in tissue architecture, variations in skin thickness, collagen density, vascular distribution, and gene expression across different regions may influence laser energy delivery and its overall effect. Previous studies on laser treatments have primarily focused on clinical and histological analysis, with limited insight into cellular dynamics and intercellular communication during the early remodeling phase. Our results showed that early-stage laser treatment significantly increased keratinocyte numbers while reducing fibroblasts, enhancing their crosstalk to promote skin remodeling. Previous studies demonstrated fibroblasts can stimulate keratinocyte proliferation and immune responses by secreting growth factors such as KGF and FGF10, while in psoriasis, SFRP2 + fibroblasts activated IL-36G signaling in keratinocytes, amplifying inflammation [32]. In our study, laser treatment not only increased keratinocytes but also enhanced their functions, including proliferation, barrier repair, and signal secretion. This enhanced signal secretion mediated the exchange of information between keratinocytes and fibroblasts, particularly through VEGF, EGF, FN1, FGF, TGF-β, and PDGF signaling pathways. VEGF and EGF signaling promoted keratinocyte proliferation while strengthening communication with fibroblasts [45, 46] Conversely, fibroblasts can secrete growth factors and other paracrine mediators to support keratinocyte proliferation, differentiation, and epidermal repair [47]. The increased expression of FGF10 in fibroblasts in the single-cell dataset may provide a potential clue supporting bidirectional communication. The intercellular communication is not limited to receptor-ligand interactions but also involves in macromolecular structures such as EVs (8), lipid rafts [48], and migrasomes [49] as mediators of signaling. Notably, the expression of Myo5b was significantly upregulated in keratinocytes following laser treatment. Myo5b binds actin nucleation factors Spire and Rab11a, forming a ternary complex (Myo5b/Spire/Rab11a) that drives long-distance vesicle transport [50]. Furthermore, Myo5b plays a crucial role in granular layer keratinocytes by regulating the dynamic distribution of intracellular vesicles through actin-dependent transport and anchoring, which its function is essential for maintaining epidermal barrier integrity [51]. Laser treatment may enhance vesicle distribution regulation by upregulating Myo5b expression, thereby optimizing keratinocyte function and accelerating barrier repair. This study identified three distinct keratinocyte subtypes, with basal and spinous keratinocytes significantly increasing post-laser treatment, while EMT keratinocytes showed minimal early-stage changes. EMT is typically associated with embryonic development, various cancers, and trauma healing [52], relies on pathways like TGF-β and Wnt for activation [53]. These findings suggest early skin repair prioritizes barrier restoration through basal and spinous keratinocyte proliferation and differentiation rather than EMT-driven matrix remodeling. Keratinocytes serve dual roles as cytokine secretors and targets, regulating their functions in response to immune signals [54, 55]. In this study, KE0 highly expresses IL1A, suggesting its role as an immune sentinel that initiates localized inflammatory responses. Previous studies have shown that PRR stimulation induces IL-1 signaling, which promotes rapid immune responses and stimulates IL-6 expression [54], and enhanced keratinocyte proliferation for barrier repair [56]. Additionally, IL-18 was predominantly expressed in basal keratinocytes, suggesting a potential role in immune regulation in the post-laser skin microenvironment [57]. Keratinocyte-derived IL-18 may contribute to immune cell activation and subsequent inflammatory regulation through its effects on IFN-γ production and Th1-associated responses [54, 58]. This signaling process underscores the existence of a highly collaborative signaling network among keratinocyte subtypes, which work together to mediate immune responses and barrier repair. Our study revealed a significant reduction in the number of reticular fibroblasts in the laser-treated group, consistent with their role as primary ECM secretors [59] and explaining the early decline in type I/III collagen levels. We did not detect inflammatory fibroblast subtypes in this study, aligning with prior findings on fibroblast heterogeneity in porcine skin [23]. However, chemokine and inflammatory factor expression was enriched in papillary and mesenchymal fibroblasts. Papillary fibroblasts, located in the superficial dermis, may contribute to immune cell recruitment and early repair responses through their secretory activity. Mesenchymal fibroblasts may participate in inflammatory regulation and tissue remodeling in the wound microenvironment through distinct regulatory and secretory programs [60]. This intercellular collaboration underpins the efficiency of tissue repair processes. Functional analysis revealed a notable enrichment in ER-to-Golgi vesicle-mediated transport, suggesting that laser-induced injury signals shift fibroblasts from a quiescent to a hypersecretory state. Enhanced vesicle-mediated transport likely supports increased secretion of signaling molecules and intercellular communication [61], while also preparing fibroblasts for the rapid replenishment of new ECM components, providing the structural foundation for tissue remodeling. Additionally, several genes previously associated with fibroblast proliferation and secretory activity, including FGF10, RUNX2, NOX4, and SMOC2 [62, 63], were significantly upregulated. These changes may indicate a transcriptionally primed state that prepares fibroblasts for subsequent ECM remodeling. We verified the interaction between keratinocytes and fibroblasts during the early stage of laser treatment through in vitro experiments. Our findings revealed that the regulatory effect of keratinocytes on fibroblasts was significantly enhanced following laser treatment. In a co-culture model, keratinocytes from the early laser treatment phase were found to promote their own activity while simultaneously inhibiting the activity of fibroblasts. This regulation is particularly crucial in the early stages, as it prevents fibroblasts from becoming abnormally activated and proliferating due to stimulation by other cytokines or soluble factors, thereby mitigating the risk of tissue fibrosis. This regulatory mechanism appears to be mediated by EVs secreted by keratinocytes. EVs are widely present across various cell types and facilitate intercellular substance exchange and signaling [11, 64]. Interestingly, we observed that the average particle size of EVs secreted by laser-treated keratinocytes was larger than those from untreated cells, suggesting a unique adaptive response. This phenomenon could be likened to “escaping with essential supplies” in a “crisis.” Identifying these key substances represents an important scientific question for future research. Among the EV cargo, we identified a significant upregulation of CEBPA, a transcription factor known to regulate cellular activity and inhibit tissue fibrosis [40]. These CEBPA-enriched exosomes exhibited dual roles: they promoted keratinocyte proliferation by interacting with YAP1/TAZ to enhance barrier repair, and they transported CEBPA to fibroblasts, upregulating IGFBP3 expression, which temporarily suppressed fibroblast activity. This dual functionality likely prevents fibroblasts from abnormal activation and proliferation, reducing the risk of fibrosis. The skin repair process resembles a strategic defense system, where keratinocytes serve as “gatekeepers,” stabilizing the “gate” (skin barrier) to prevent excessive external stimuli and creating a stable environment for fibroblasts, the “inner family,” to prepare repair signals and materials. At the right moment, keratinocytes and fibroblasts collaborate to remodel the tissue and restore function. If the barrier remains unstable, fibroblasts may be prematurely stimulated, leading to abnormal activation and fibrosis, a scenario described as “gate failure.” These findings highlight the intricate interplay between keratinocytes and fibroblasts in regulating skin repair and offer valuable insights into mechanisms that can optimize laser treatment outcomes. We acknowledge that this study has several limitations. The single-cell and mechanistic analyses were primarily focused on the condition of parameter 7 at day 7, and therefore mainly reflect the early response to laser treatment rather than the full temporal- and parameter-dependent landscape. In addition, the scRNA-seq samples were collected from spatially separated regions within a pig, which allowed us to evaluate local tissue heterogeneity but could not fully substitute for independent animal-level biological replicates. Although additional sample-level correlation and integration analyses supported the consistency of the spatially separated samples, future studies with larger animal cohorts will be required to validate these cell-state changes. The mechanistic findings were derived from a porcine model and in vitro co-culture/EV assays, which may not fully capture the complexity of human skin repair in vivo. Moreover, keratinocyte-fibroblast crosstalk after laser treatment is likely shaped by dynamic tissue changes and multidimensional regulatory networks rather than a single pathway [65, 66]. Further studies incorporating broader temporal sampling, additional laser settings, and human-relevant validation will be needed to refine this regulatory framework and clarify its translational relevance. Conclusion In summary, this study employs an integrated strategy combining morphological assessment, molecular markers, and single-cell RNA sequencing to comprehensively uncover how laser treatment reshapes the skin microenvironment. It reveals that laser-induced keratinocyte-derived exosomes enriched with CEBPA mediate intercellular communication, promoting keratinocyte proliferation by interacting with YAP1/TAZ to enhance barrier repair, and suppressing fibroblast overactivation by upregulating IGFBP3 expression. Our findings provide new insights into how laser treatments modulate cellular behavior and intercellular signaling, emphasizing the therapeutic potential of exosome-based strategies for promoting skin regeneration and rejuvenation. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1: Figure S1. The molecular markers and histological analysis in post-laser treatment skin. Figure S2. The molecular markers analysis in post-laser treatment skin. Figure S3. Gene number, UMI counts and functional enrichment across skin samples. Figure S4. The GSEA enrichment analysis of the DEGs from cell types. Figure S5. The KEGG enrichment analysis of the DEGs from cell types. Figure S6-7. Verification of crosstalk between keratinocytes and fibroblasts. Supplementary Material 2: Table S1. Quantitative analysis of western blot bands. Table S2. Marker genes identified for each cell cluster. Table S3. Differentially expressed genes (DEGs) across all skin cell types. Table S4. Cell communication among skin cell populations. Table S5. The relative information flow of signaling pathways. Table S6. Expression levels of diverse cytokine-related genes across keratinocyte subclusters. Table S7. Predicted CEBPA-binding motifs in the regulatory regions of YAP1, WWTR1/TAZ, and IGFBP3. Acknowledgements Thank Yue Ming from Laboratory of Molecular Oncology of West China Hospital for technical assistance in animal experiments. Abbreviations Extracellular vesicles Single-cell RNA sequencing Extracellular matrix Differentially expressed genes Kyoto encyclopedia of genes and genomes Gene ontology Transcription factors Single-cell regulatory network inference and clustering Transmission electron microscopy Nanoparticle tracking analysis Elastin Decorin Fibronectin Western blot Unique molecular identifier Manifold approximation and projection Gene set enrichment analysis Epithelial-mesenchymal transition Author contributions QZ and XW conceived the study and designed the project and experiments. XW acquired funding for this work. QZ, HW, and YZ performed the experiments. QZ and XW analyzed the data. QZ wrote the original draft. XW reviewed and edited the paper. XW supervised the work. Funding This work was supported by grants from Cynosure (Grant No. 311231022). Data availability Raw sequence data have been deposited in the Genome Sequence Archive at the National Genomics Data Center, part of the China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences. The data are publicly accessible under the accession number CRA019722 and can be accessed at https://ngdc.cncb.ac.cn/gsa. Declarations Ethics approval and consent to participate All animal experiments were approved by the Animal Welfare Committee of West China Hospital, Sichuan University and conducted in compliance with their ethical standards and guidelines (20240401011). Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Qin Zou and Hao Wang contributed equally to this work. References Associated Data Supplementary Materials Supplementary Material 1: Figure S1. The molecular markers and histological analysis in post-laser treatment skin. Figure S2. The molecular markers analysis in post-laser treatment skin. Figure S3. Gene number, UMI counts and functional enrichment across skin samples. Figure S4. The GSEA enrichment analysis of the DEGs from cell types. Figure S5. The KEGG enrichment analysis of the DEGs from cell types. Figure S6-7. Verification of crosstalk between keratinocytes and fibroblasts. Supplementary Material 2: Table S1. Quantitative analysis of western blot bands. Table S2. Marker genes identified for each cell cluster. Table S3. Differentially expressed genes (DEGs) across all skin cell types. Table S4. Cell communication among skin cell populations. Table S5. The relative information flow of signaling pathways. Table S6. Expression levels of diverse cytokine-related genes across keratinocyte subclusters. Table S7. Predicted CEBPA-binding motifs in the regulatory regions of YAP1, WWTR1/TAZ, and IGFBP3. Data Availability Statement Raw sequence data have been deposited in the Genome Sequence Archive at the National Genomics Data Center, part of the China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences. The data are publicly accessible under the accession number CRA019722 and can be accessed at https://ngdc.cncb.ac.cn/gsa.