Skin samples were collected from 8 patients diagnosed with PTM and 9 healthy controls undergoing cosmetic surgery (Supplementary Table S1). Skin biopsies were strictly collected from the pretibial region. All PTM patients had a documented history of GD and were diagnosed by skin histopathology. There were no significant differences in sex or body mass index (BMI) between the two groups. This prospective study was approved by the institutional ethics committee. Informed consent was obtained from all participants.

Skin tissues were processed for LC-MS and GC-MS based untargeted metabolomics as previously described (14, 15). Briefly, tissues were homogenized in methanol (1 mg:50 μL), followed by centrifugation (12, 000 × g, 15 min, 4°C). Supernatants were aliquoted for LC-MS or GC-MS analysis. For LC-MS, samples were mixed with cold methanol containing internal standards; for GC-MS, samples were dried under nitrogen, derivatized with methoxamine hydrochloride and MSTFA, and analyzed under standard chromatographic conditions. Raw data were processed using MassHunter and XCMS-Online for peak detection, alignment, normalization, and missing value imputation. Metabolite identification was based on NIST and HMDB databases. Significant metabolites were defined by p ≤ 0.05 and |log₂FC| ≥ 0.5, and pathway enrichment was performed using MetaboAnalyst 6.0.

Frozen tissues were thawed gradually, embedded in Cryo-Gel, sectioned at 20 μm, and mounted on Superfrost Plus slides. One section underwent H&E staining; others were vacuum-dried (~30 min). DESI-MSI was performed in positive and negative ion modes using acetonitrile/water (80:20, v/v) at 1.5 μL/min, with 20 μm spatial resolution, m/z 70–1200, and resolution of 20, 000. Sebaceous gland regions were excluded, and ion images were reconstructed using MassLynx and HDI software.

HFFs (Zhong Qiao Xin Zhou Biotechnology, Cat# ZQ0450, RRID: CVCL_3285) were cultured in F12/DMEM (Gibco, USA) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37°C in a humidified incubator with 5% CO₂. When cell confluence reached 70–80%, the medium was replaced, and cells were treated with 200 μM OA or PA for 48 h.

Total RNA from HFFs was extracted using TRIzol reagent (Invitrogen) and used to prepare strand-specific RNA-seq libraries for sequencing on the Illumina platform. Clean reads were aligned using HISAT2, and differentially expressed genes (|log₂FoldChange| > 1, adjusted p ≤ 0.05) were identified with DESeq2. Functional enrichment was performed using the ClusterProfiler R package, including GO analysis of biological processes, cellular components, and molecular functions, as well as KEGG pathway analysis, with terms or pathways showing FDR < 0.05 considered significant.

After 48 h of fatty acid treatment, cell culture supernatants were collected for cytokine quantification using a Multi-Analyte Flow Assay Kit (Dake Precision Technology, Shenzhen, China), following the manufacturer’s instructions. Flow cytometric analysis was performed on a BD FACSCanto™ system, and data were processed using LEGENDplex™ software.

Cells were lysed in RIPA buffer with protease inhibitors at 4°C for 15 min, then centrifuged at 12, 000 × g for 15 min. Protein concentration was normalized, and samples were separated via SDS-PAGE and transferred to PVDF membranes. After blocking for 15 min, membranes were incubated overnight with primary antibodies: α-SMA (Thermo Fisher Scientific, Cat# 14-9760-80, RRID: AB_2572995), ADFP (Invitrogen, Cat# PA1-16972, RRID: AB_2223607), and tubulin (Proteintech, Cat# HRP-66031, RRID: AB_2687491). After washing, membranes were incubated with HRP-conjugated secondary antibodies, anti-mouse IgG (Epizyme Biomedical Technology, Cat# LF101, RRID: AB_3083706) and anti-rabbit IgG (ABCAM, Cat# Ab6721, RRID: AB_955447), for 1 h at room temperature. Bands were visualized using an ECL detection kit.

Paraffin-embedded tissue sections were deparaffinized, rehydrated, and subjected to antigen retrieval. Endogenous peroxidase was blocked with 3% H₂O₂, followed by blocking with normal goat serum. Sections were incubated overnight at 4°C with anti-IL-8 antibody (Thermo Fisher Scientific, Cat# PA5-79114, RRID: AB_2746230), then with HRP-conjugated secondary antibody, and developed using DAB. Nuclei were counterstained with hematoxylin. Positive and negative controls were included, and staining was evaluated by light microscopy. The percentage of positive cells was calculated by dividing the number of positively stained cells by the total number of cells in four randomly selected high-power fields (HPFs) per section.

To characterize the metabolic profiles of PTM skin lesions, we performed untargeted metabolomics analyses using both LC-MS and GC-MS platforms on pretibial skin from patients with PTM and healthy controls (Figure 1a). Principal component analysis (PCA) showed distinct clustering of PTM and control groups across LC-MS positive/negative ion modes and the GC-MS platform suggesting significant metabolic divergence in PTM lesions (Figure 1b).

LC-MS in positive ion mode identified 53 significantly upregulated and 817 significantly downregulated metabolites. Upregulated metabolites included glyceric acid (fold change [FC] = 1.79), LysoPC (16:1/0:0) (FC = 2.26), and PC (36:1) (FC = 2.00). Downregulated metabolites mainly involved lipids and carbohydrate-related compounds such as α-linolenic acid (FC = 0.63), adenosine (FC = 0.14), and L-acetylcarnitine (FC = 0.41). In the negative ion mode, 22 upregulated metabolites were detected, including citric acid (FC = 36.9), azelaic acid (FC = 1.53), and arachidonic acid (FC = 1.47), alongside 108 downregulated metabolites such as adenine (FC = 0.21), dodecanedioic acid (FC = 0.62), and hexanoylcarnitine (FC = 0.67). GC-MS analysis identified 9 upregulated and 1 downregulated metabolites. Notably, PA was markedly elevated (FC = 15.47) (Figure 1c), highlighting PA as a markedly altered metabolite in PTM lesions. By integrating results from both platforms, we identified a total of 107 potential biomarkers, including significantly upregulated citric acid, PA, and LysoPCs, as well as downregulated L-acetylcarnitine, docosatrienoic acid, and arachidic acid (Figure 1d, Supplementary Table S2).

KEGG pathway enrichment analysis revealed that both upregulated and downregulated metabolites were significantly enriched in key energy and lipid metabolic pathways, including the biosynthesis of unsaturated fatty acid, citrate cycle (TCA cycle), and arachidonic acid metabolism. Additionally, pathways related to fatty acid synthesis, elongation, and degradation were altered (Figure 1e). These findings indicate a pronounced dysregulation of fatty acid metabolism and TCA cycle activity in PTM lesions.

To investigate regional metabolic alterations in PTM, we applied DESI-MSI to pretibial skin samples from a PTM patient and a healthy control, focusing on fibroblast-rich dermal compartments. To minimize variability due to structural heterogeneity, we excluded areas containing sebaceous glands and restricted the spatial analysis to matched dermal regions, thereby reducing bias from lipid-dense structures.

We selected representative metabolites from the TCA cycle, fatty acid metabolism, glycolysis and amino acid metabolism pathways, and visualized their trends through histograms of conventional metabolism and spatial metabolomic ion images (Figure 2a). The results showed a general upregulation of TCA cycle intermediates, with citric acid exhibiting a particularly strong increase (FC = 36.9, p < 0.05). Other intermediates, including cis-aconitate, malic acid, and oxaloacetate, did not reach statistical significance but displayed a consistent upward trend. These findings indicate a profound dysregulation of mitochondrial homeostasis in PTM lesions. In contrast, no significant changes were observed in glycolytic or amino acid metabolites, including glucose, lactate, and most amino acids. PA was significantly elevated in conventional and spatial metabolism, while OA was significantly upregulated in spatial metabolism.

Given that PTM is characterized by fibroblast activation predominantly within the dermis, we co-registered DESI-MSI ion images with matched H&E-stained sections to spatially resolve metabolic alterations in the dermal compartment (Figure 2b). Dermal fatty acid signals were markedly elevated, with focal enrichment patterns most pronounced for PA and OA (Figures 2c, d). These findings suggest a dermal lipid-accumulation signature in PTM lesions, with a potential association with fibroblast activation.

To investigate the pro-inflammatory effects of fatty acids on fibroblasts, we treated HFFs with OA and PA, both of which were identified as elevated in PTM lesions, and performed transcriptomic profiling using RNA-seq.

Compared with the control group, OA treatment led to 489 upregulated genes and 435 downregulated genes, whereas PA treatment resulted in 77 upregulated genes and 51 downregulated genes (Figure 3a). GO analysis showed that OA-upregulated genes were enriched in Toll-like receptor signaling and the positive regulation of fibroblast proliferation. PA-upregulated genes were enriched in the positive regulation of fibroblast proliferation, the ERK1/ERK2 cascade, and the ERBB4 signaling pathway (Figure 3b). KEGG pathway analysis revealed an enrichment of inflammation-related pathways. Following OA treatment, upregulated genes were enriched in the IL-17, p53, and NF-κB signaling pathways. Following PA treatment, upregulated genes were enriched in the TNF, IL-17, and JAK-STAT signaling pathways (Figure 3c). These results suggest that both OA and PA treatments activate inflammation-related signaling pathways in fibroblasts and may influence their differentiation.

Intersection analysis identified 47 commonly upregulated and 16 downregulated genes (Figure 3d). Two key pro-inflammatory chemokines, CXCL1 and CXCL8, were consistently elevated (Figure 3e), implicating that fatty acids may trigger fibroblast-mediated inflammatory responses via these mediators. Pathway enrichment analysis of the shared upregulated genes revealed significant involvement of the IL-17 and NF-κB signaling pathways, both of which play central roles in inflammatory regulation (Figure 3f). These findings suggest that fatty acid accumulation in PTM lesions may promote local immune activation by stimulating inflammatory gene expression in fibroblasts through classical pro-inflammatory signaling pathways.

Based on transcriptomic analysis showing activation of inflammatory pathways in fibroblasts after OA and PA exposure, we further examined their effects on cytokine secretion and fibroblast phenotype. HFFs were treated with OA or PA, and inflammatory cytokines in the culture supernatant were subsequently quantified. Both OA and PA treatment significantly increased IL-8 secretion, whereas PA also significantly increased IL-6 levels (Figure 4a). To assess disease relevance, IL-8 expression in skin sections from PTM patients and healthy controls was examined by immunohistochemistry. The percentage of IL-8 positive cells was significantly higher in PTM lesions than in healthy controls (Figure 4b). These findings suggest that altered fatty acid metabolism may play a role in sustaining local inflammation in PTM skin by promoting fibroblast-derived IL-8 production.

We also examined the effect of dysregulated fatty acid metabolism on fibroblast differentiation by assessing the expression of two classical differentiation markers: α-smooth muscle actin (α-SMA) for myofibroblasts (16), and adipose differentiation-related protein (ADFP) for adipocytes (17, 18) Western blot analysis revealed that ADFP expression remained unchanged following OA or PA treatment; however, PA stimulation significantly upregulated α-SMA expression, suggesting a fibroblast-to-myofibroblast phenotypic transition and potentially contributing to tissue remodeling and fibrosis. (Figure 4c).

To further validate the physiological relevance of these findings, we repeated the key experiments using primary dermal fibroblasts isolated from adult skin. Consistent with the observations in HFFs, PA treatment significantly upregulated the secretion of IL-8 and IL-6 (Supplementary Figure S1a) and increased the expression of α-SMA (Supplementary Figure S1b) in primary fibroblasts.