目的：建立黄褐斑皮损及皮损旁外观正常皮肤组织的scRNA-seq图谱，探讨特征成纤维细胞亚群在黄褐斑中的作用机制。

方法：

1. 收集黄褐斑患者面部皮损区（MEL组）以及远离皮损旁外观正常的皮肤组织（NC组）进行scRNA-seq，使用降维聚类的方法对细胞亚群以及主要细胞的子亚群进行分群鉴定，分析成纤维细胞子亚群细胞频率，对CCL19⁺促炎成纤维细胞子亚群进行GSVA炎症基因评分。取黄褐斑患者皮肤组织进行免疫荧光染色，体外验证CCL19⁺促炎成纤维细胞亚群。

2. 利用Seraut软件分析促炎成纤维细胞差异基因，选取在黄褐斑组高表达CXCL2、CCL2、CXCL1、CXCL3、IL-6等细胞因子进行后续验证。采用UVB联合黄体酮以及情绪应激的方式构建C57BL/6J黄褐斑小鼠模型，使用组织病理及血清酶联免疫吸附实验评估模型。进一步利用免疫荧光染色验证小鼠CCL19⁺促炎成纤维细胞亚群，流式细胞染色验证黄褐斑样小鼠促炎成纤维亚群相关细胞因子表达差异。

3. 利用SCENIC以及Readme数据库预测黄褐斑中特征促炎成纤维细胞上游转录因子，筛选出关键调控转录因子NF-κB、REL、RELB，在体外人原代成纤维细胞中过表达转录因子构建特征促炎成纤维细胞亚群，qPCR验证过表达效率以及CXCL2、CCL2、CXCL1、CXCL3、IL-6差异细胞因子表达水平，酶联免疫吸附实验检测上清中细胞因子分泌情况。最后通过生物信息学建立细胞间通讯网络探索促炎成纤维细胞对黑素细胞的影响，并利用体外构建的特征促炎成纤维细胞上清条件培养人黑素瘤细胞MNT1及人原代黑素细胞，检测黑素含量，qPCR检测TYR及MITF基因表达变化情况。

结果：

1. 对研究中6例黄褐斑患者scRNA-seq数据进行处理及质量控制，筛选出符合标准的细胞进行下一步分析。使用Seurat软件对细胞进行降维和无监督聚类分析，确定了12个主要细胞簇，根据亚群上调基因进一步分析鉴定出角质形成细胞子亚群14个、黑素细胞子亚群5个以及成纤维细胞子亚群7个。通过Cellchat软件对主要细胞亚群构建细胞间通讯网络，发现黄褐斑组成纤维细胞、黑素细胞以及血管内皮细胞三者之间的细胞间通讯较对照组显著增强。

2. 利用Seurat软件分析细胞频率，提示成纤维细胞各子亚群细胞占比在组间无明显差异，进一步筛选出子亚群占比最多的CCL19⁺促炎成纤维细胞亚群，通过GSVA炎症基因集评分表明黄褐斑组CCL19⁺促炎成纤维细胞亚群炎症功能较对照组显著增强，利用黄褐斑患者组织病理切片免疫荧光染色验证CCL19⁺促炎成纤维细胞亚群。Seurat软件进一步分析CCL19⁺促炎成纤维细胞组间差异基因，发现CXCL2、CCL2、CXCL1、CXCL3、IL-6等炎症因子在黄褐斑组显著升高。体外构建C57BL/6J黄褐斑样小鼠模型，造模部位表现为黄褐斑样褐色色素沉着斑，组织病理示真表皮黑素增多，造模部位皮肤及血清TYR含量较对照组升高，血清SOD/MDA氧化应激水平失衡，符合临床中黄褐斑患者特征。黄褐斑样小鼠造模部位皮肤也存在CCL19⁺促炎成纤维细胞，且成纤维细胞中CCL2、CXCL1、IL-6表达量较对照组小鼠显著升高。

3. SCENIC数据库预测调控黄褐斑中特征促炎成纤维细胞的上游转录因子为NFIL3、NF-κB、ATF3、REL、RELB、EGR1、CEBPB、CEBPD、BCL3，Readme数据库再次预测与SCENIC取交集转录因子筛选为NF-κB、REL、RELB。体外模拟构建黄褐斑中特征促炎成纤维细胞，该群细胞CXCL2、CCL2、CXCL1、CXCL3、IL-6 mRNA表达水平及分泌水平均显著上调（P＜0.05）。利用体外构建的特征促炎成纤维细胞上清培养人黑素瘤细胞MNT1及人原代黑素细胞后，人黑素瘤细胞MNT1及人原代黑素细胞黑素含量增加，TYR及MITF mRNA表达水平显著增加（P＜0.05）。

结论：

1. 本研究首次建立了黄褐斑及皮损旁皮肤组织的单细胞图谱，解析了细胞组分，并对主要细胞进行子亚群鉴定和细胞通讯构建，提示黄褐斑中存在细胞异质性，并且存在可能与疾病相关的细胞亚群。

2. 本研究首次发现黄褐斑中CCL19⁺促炎成纤维细胞炎症功能显著增强，并在黄褐斑患者组织切片中进一步得到验证。在黄褐斑中，该群细胞可能通过上游NF-κB、REL、RELB等转录因子调控分化，高表达CXCL2、CCL2、CXCL1、CXCL3、IL-6等炎症因子，进一步促进黑素细胞黑素合成。

3. 本研究成功构建了一种与临床黄褐斑十分相似的C57BL/6J动物模型，表型及相关指标符合临床黄褐斑特征，该模型为后续黄褐斑的研究提供应用价值。在黄褐斑样小鼠中验证同样存在一群炎症功能增强的促炎成纤维细胞亚群，高表达相关炎症因子，提示促炎成纤维细胞亚群在黄褐斑中发挥重要作用。

【Objective】 To construct an scRNA-seq profile of melasma lesions and adjacent normal skin, elucidating the functional mechanisms of distinct fibroblast subsets in melasma development.

【Methods】1. Skin samples were collected from facial lesions of melasma patients (MEL group) and normal-appearing skin distant from lesions (NC group) for scRNA-seq. Dimensionality reduction and clustering methods were used to identify cell subpopulations and major cell subsets. Fibroblast subset frequencies were analyzed, and GSVA inflammatory gene set scoring was performed on CCL19⁺ pro-inflammatory fibroblast subsets. Immunofluorescence co-staining was conducted on melasma patient skin tissues to validate the CCL19⁺ pro-inflammatory fibroblast population.2. The Seurat software was used to screen differentially expressed genes in pro-inflammatory fibroblasts, selecting cytokines highly expressed in the melasma group (CXCL2, CCL2, CXCL1, CXCL3, IL-6) for further validation. A C57BL/6J mouse model of melasma was established using UVB irradiation combined with progesterone administration and emotional stress. The model was evaluated through histopathological examination and serum ELISA. Flow cytometry was used to verify differences in cytokine expression related to pro-inflammatory fibroblast subsets in melasma-like mice.3. The SCENIC and ReMap databases were employed to predict upstream transcription factors of characteristic pro-inflammatory fibroblasts, identifying key regulatory transcription factors NF-κB, REL, and RELB. These transcription factors were overexpressed in primary human fibroblasts in vitro to construct characteristic pro-inflammatory fibroblast subsets. qPCR was used to verify overexpression efficiency and differential expression of cytokines (CXCL2, CCL2, CXCL1, CXCL3, IL-6), while ELISA detected cytokine secretion in supernatants. Finally, cell-cell communication networks were established through bioinformatics to explore the effects of pro-inflammatory fibroblasts on melanocytes. Conditioned medium from in vitro-constructed characteristic pro-inflammatory fibroblasts was used to culture MNT1 cells and primary human melanocytes, followed by measurement of melanin content and qPCR analysis of TYR and MITF expression changes.

【Results】1.scRNA-seq data processing and quality control were performed on samples from six melasma patients. After filtering cells meeting quality criteria, subsequent analyses were conducted. Using Seurat software, cells underwent dimensionality reduction and unsupervised clustering, identifying 12 major cell clusters. Further subpopulation analysis based on upregulated marker genes revealed 14 keratinocyte subsets, 5 melanocyte subsets, and 7 fibroblast subsets. Cell-cell communication networks constructed by CellChat demonstrated significantly increased ligand-receptor interactions among fibroblasts, melanocytes, and vascular endothelial cells in the MEL group compared to NC controls. 2. Fibroblast subset frequency analysis using Seurat showed no significant intergroup differences in cellular proportions. The predominant CCL19⁺ pro-inflammatory fibroblast subset was selected for further investigation. GSVA of inflammatory gene sets revealed enhanced inflammatory functionality in melasma-derived CCL19⁺ fibroblasts compared to controls, which was validated by immunofluorescence staining of melasma tissue sections. Differential gene expression analysis identified significantly elevated levels of inflammatory factors (CXCL2, CCL2, CXCL1, CXCL3, IL-6) in melasma-derived pro-inflammatory fibroblasts. 3. A C57BL/6J mouse model of melasma was successfully established, exhibiting characteristic hyperpigmented macules with histopathological evidence of increased epidermal melanin. The model demonstrated elevated cutaneous and serum tyrosinase (TYR) activity, along with imbalanced SOD/MDA oxidative stress markers, recapitulating clinical melasma features. Model skin showed significantly increased expression of Ccl2, Cxcl1, and Il-6 compared to controls. 4. SCENIC and ReMap database analyses predicted NFIL3, NF-κB, ATF3, REL, RELB, EGR1, CEBPB, CEBPD, and BCL3 as potential upstream transcriptional regulators of pro-inflammatory fibroblasts, with consensus factors NF-κB, REL, and RELB identified through intersection analysis. In vitro-engineered pro-inflammatory fibroblasts exhibited significantly upregulated mRNA expression (P<0.05) and secretion of CXCL2, CCL2, CXCL1, CXCL3, and IL-6. Treatment of MNT1 cells and primary human melanocytes with conditioned medium from these fibroblasts resulted in increased melanin content and significantly elevated expression (P<0.05) of melanogenesis-related genes TYR and MITF.

【Conclusion】1. This study presents the first comprehensive single-cell atlas of melasma lesions and adjacent normal skin, systematically characterizing cellular composition and identifying distinct subpopulations of major cell types. Our cell-cell communication analysis reveals significant cellular heterogeneity in melasma, highlighting the existence of disease-associated cell subpopulations that may contribute to pathogenesis. 2. We have identified a functionally enhanced CCL19⁺ pro-inflammatory fibroblast subpopulation in melasma for the first time, , which was further validated through histological examination. Mechanistically, these fibroblasts demonstrate NF-κB/REL/RELB transcription factor-mediated upregulation and subsequent hypersecretion of inflammatory cytokines (CXCL2, CCL2, CXCL1, CXCL3, IL-6), establishing a pro-melanogenic microenvironment that promotes melanin synthesis in melanocytes. 3. We successfully established a novel C57BL/6J mouse model that faithfully recapitulates key clinical features of human melasma, including characteristic pigmentation patterns and biochemical markers. This model demonstrates significant translational value for future melasma research. Importantly, we confirmed the presence of similarly activated pro-inflammatory fibroblasts in the animal model, which exhibited parallel cytokine overexpression profiles, further supporting their critical role in melasma pathogenesis.