Lipodystrophy and adipose tissue recovery are mediated by the Wnt/lipogenesis axis during skin fibrosis Authors: Suneeti R Madhavan, Piper Moore, Rohan Varshney, Miles Montegut, Qiannan Ma, Kayla Klatt, Reshmi Parameswaran, Elif G Ertugral, Chandrasekhar R Kothapalli, Michael C Rudolph, Radhika P Atit Published: 2026 ABSTRACT AbstractAcquired lipodystrophy in the dermal white adipose tissue (DWAT) is an early phenotype of skin fibrosis, followed by the accumulation of extracellular matrix (ECM). Lipodystrophy syndromes are estimated to affect 1 in 20,000 people and are associated with metabolic comorbidities. Recently, we showed that fibrosis‐associated lipodystrophy depended on sustained Wnt signaling, although the mechanism remains unclear. Transcriptomic profiling of mature dermal adipocytes in vivo revealed that Wnt activation downregulated the de novo lipogenesis (DNL) axis enzymes within 48 h. We further found that expression of fatty acid synthase (FASN), a key DNL enzyme, depends on sustained Wnt activation in vitro and in vivo. In a bleomycin model, and in human systemic sclerosis (< 1 year disease duration) and keloids, FASN was significantly downregulated. Notably, inhibition of FASN in mice during reversal of Wnt‐induced fibrosis impaired the recovery of DWAT lipid content and ECM architecture. Collectively, these findings demonstrate that acquired lipodystrophy in skin fibrosis is mediated by a previously unrecognized role of the Wnt–DNL axis. These findings underscore the importance of this pathway in lipodystrophy and fibrosis and highlight its potential as a therapeutic target in skin fibrosis. © 2026 The Author(s). The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.Keywords: adipocytes, lipid, fatty acids, extracellular matrix, skin fibrosis, lipodystrophy Introduction The pathological loss of adipose tissue, lipodystrophy, may be congenital or acquired and is frequently observed in association with autoimmune conditions, such as systemic sclerosis (SSc) [1]. Although estimated to affect 1 in 20,000 people in the USA, lipodystrophy often goes undiagnosed [2, 3]. Undetected lipodystrophy increases the risks of morbidity and mortality by disrupting tissue elasticity and structural integrity, cellular signaling, and metabolism [4]. Acquired forms of lipodystrophy usually accompany metabolic abnormalities in adipocytes, such as dyslipidemia and severe insulin resistance, as well as tissue fibrosis of the lung, liver, and skin in SSc [4, 5, 6, 7]. Fibrotic lipodystrophy results from an imbalance of lipid uptake and lipid breakdown in adipocytes, lipofibroblasts, and related stromal cell types [4, 5]. Fibrosis‐related diseases affect approximately one in four people globally, yet no approved therapies currently exist to reverse lipodystrophy or associated fibrosis [1, 8]. Although lipodystrophy precedes fibrosis, the initiating signals and cellular mechanisms driving adipose loss remain unclear. In this study, we investigate the early signals using a novel mouse skin model, a tissue that is composed of matrix‐rich dermis and underlying dermal white adipose tissue (DWAT). DWAT serves diverse physiological roles, including elasticity, wound healing, and hair cycle regulation [9]. The loss of dermal adipocytes or their lipid content is the hallmark of multiple human skin conditions such as SSc, androgenic alopecia, radiation‐ or therapy‐induced cutaneous toxicity, psoriasis, and skin aging [8, 10, 11, 12, 13]. In humans, pigs, and rodents, the dorsal DWAT is anatomically and metabolically distinct from subcutaneous adipose tissue (SWAT) and arises from Engrailed homeobox 1 (EN1)‐expressing fibro‐adipocyte progenitors in mice [14, 15, 16, 17]. Lipodystrophy precedes dermal extracellular matrix (ECM) dysfunction in humans and mice and contributes to fibrosis progression resulting in thickened, stiffened, and scarred skin [13, 18, 19, 20, 21]. The DWAT lipodystrophy phenotype has consistently been recapitulated in mouse models of dermal fibrosis, including Wnt activation, transforming growth factor‐β (TGF‐β) activation, bleomycin treatment, and radiation exposure [21, 22, 23, 24, 25]. While much of the field has focused on fibroblast activation and ECM dysregulation, the mechanisms driving DWAT depletion remain largely understudied, representing a critical opportunity to identify early molecular targets for lipodystrophy and fibrosis. Acquired lipodystrophy in humans is associated with multiple signaling pathways that impair adipose tissue function and its distribution [4, 26, 27]. The Wnt signaling pathway is a conserved profibrotic signal across human skin, lung, and liver tissues and is well documented in corresponding animal models of fibrosis [28, 29, 30, 31]. Canonical Wnt signaling is anti‐adipogenic in adipocyte stem cells, and in mature adipocytes it plays a pro‐lipolytic role [22, 24, 32]. Importantly, in human SSc, lipid metabolic pathways are broadly dysregulated [33, 34]. Recently we demonstrated that Wnt‐induced adipocyte lipolysis was required for both lipid depletion and ECM remodeling [23, 24], providing evidence that DWAT lipid loss is not a bystander but a required event in skin fibrosis progression. However, early mediators and downstream regulators of Wnt‐driven lipid dysregulation in mature dermal adipocytes remain undefined and will be key to reversing DWAT lipodystrophy and fibrosis. Although lipid depletion of DWAT accompanies human fibrosing skin diseases, management has been limited by detailed understanding of early stages. Here, we analyzed duration‐stratified human SSc, localized scleroderma and keloid skin, revealing early suppression of key de novo lipogenesis (DNL) enzyme, fatty acid synthase (FASN). Leveraging our inducible and reversible genetic model, which confines activated β‐catenin to fibro‐adipocyte progenitors expressing En1 and descendants, we activate Wnt signaling to trigger lipodystrophy and ECM accumulation, both of which recover after withdrawal from transgene activation [23, 35]. We tested the hypothesis that activity of the principal DNL enzyme FASN is required to rescue Wnt‐induced DWAT lipodystrophy and fibrotic ECM remodeling. Our studies suggest a putative regulatory cascade in which Wnt/β‐catenin activation suppresses FASN and the DNL axis as early mediators of dermal adipocyte lipid depletion, highlighting this pathway as a potential therapeutic target for dermal disease. Materials and methods Detailed methods are provided in the Supplementary materials and methods. Ethics approval All animal procedures were approved by the Institutional Animal Care and Use Committees at Case Western Reserve University in accordance with American Veterinary Medical Association guidelines protocol 20130156, approved 21 December 2022. Animals The Wnt inducible and reversible mouse model (β‐catistab) was produced by mating En1‐Cre (gift of Alexendra Joyner), Rosa26‐rTA‐EGFP (RRID:IMSR_JAX:005572, The Jackson Laboratory, Bar Harbor, ME, USA), and TetO‐ΔN89‐myc‐β‐catenin mice (gift of John P. Lydon), as previously described and validated with myc‐tagged β‐catenin [23, 35]. Following appropriate regimens, dorsal dermis was harvested for intradermal adipocyte isolation, DWAT biopsy, whole‐skin flash freezing, and paraffin or frozen sectioning. For the bleomycin model, female BALB/c mice, age 8–10 weeks old, were purchased from The Jackson Laboratory. Bleomycin (TLC Standards, Newmarket, ON, Canada; Catalog No.: B‐190006) was solubilized in 0.1% sodium chloride at 1 mg/ml. BLM or PBS (100 μl) were injected subcutaneously into the back of 87‐day‐old (p87) mice daily for 14 days before sacrifice and dorsal skin collection at p105. A minimum of three β‐catistab mice with litter‐matched controls were analyzed for each experiment. Animals of both sexes were randomized in studies and housed in a pathogen‐free animal facility at the CWRU Animal Resource Center. All mice had access to water and food and were kept in the following environmental conditions: 12 h light/dark cycle, 20 °C room temperature, and 40–60% humidity. RNA sequencing (RNA‐seq) of dermal adipocytes RNA was extracted from isolated intradermal adipocytes collected in TRIzol following the manufacturer's protocol for high lipid, low RNA content in tissue samples. RNA‐grade glycogen (Thermo Fisher Scientific, Waltham, MA, USA; FERR0551) was used as a carrier. Total RNA quantity was measured by NanoDrop (Thermo Fisher Scientific), and RNA quality was assessed by Agilent Tape Station 4200 (Agilent, Santa Clara, CA, USA). The Universal Plus mRNA‐SEQ kit was used to input 100 ng of total RNA for sequencing libraries following the manufacturer's protocol (NuGEN/Tecan, Männedorf, Switzerland; Catalog No.: 0520), and sequencing data were collected with an Illumina HiSEQ4000 (Illumina, San Diego, CA, USA). Sequenced reads were trimmed using Trimmomatic version 0.39 [36], aligned using gsnap v2020‐12‐16 (mm10 genome; GRCm38.79) [37], and filtered and sorted using SAMTools version 1.9 [38] to generate merged BAM files. Raw FPKM counts were assembled, and expression levels were calculated using Cufflinks version 2.2.1 [39]. A matrix containing all samples and all genes was imported into R version 4.5.1 (https://www.r-project.org/), filtered for genes > 1 in all samples and a per‐gene row sum ≥ 15, VST normalization was performed, and statistical differences were calculated using the DESeq2 package (version 1.40.1). The data have been deposited to the Gene Expression Omnibus (GEO) database under accession number GSE305616. In vivo FASN inhibition FASN inhibitor TVB3664 (MedChemExpress, Monmouth Junction, NJ, USA; HY‐120062) was formulated in a vehicle containing 10% DMSO, 40% PEG300, 5% Tween‐80%, and 45% saline at a concentration of 3 mg/ml. β‐catistab mice at p21 were doxycycline induced as above, and normal chow and water were provided for 10 days to reverse fibrosis. During reversal, TVB3664 or vehicle was administered daily by gavage (3 mg/kg body weight), after which skin tissue was collected as above for analyses. In vitro FASN inhibition FASN inhibitors TVB2640 (10 μm) (MedChemExpress, HY‐120394) and TVB3166 (250 μm) (MedChemExpress, HY‐112829) were formulated in DMSO and administered to differentiated adipocytes for 5 days. In vitro 13C acetate labeling, lipidomics, tracer studies, and triglyceride assays Differentiated adipocyte cultures undergoing Wnt activation were treated with 250 μm 13C2 sodium acetate (Sigma‐Aldrich, Burlington, MA, USA; 282,014) for 48 h, after which cells were scraped in 1 ml of ice‐cold methanol 6:1 v/v acidified with 10 μl of 0.1 M HCl. Lipidomics for 13C incorporation into fatty acids (FAs) was performed as previously described [40, 41]. Statistical analyses Sample size was determined by power analysis for 80% power. Two‐group comparisons used unpaired two‐tailed Student's t‐test with Welch's correction; paired t‐tests were used for appropriate data. Three group comparisons used corrected one‐way ANOVA. Frequency distributions used the Kruskal–Wallis test. Each individual data point represents the average of one mouse (except AFM data) from three to eight biological replicates. All graphs show mean ± SEM. p values less than 0.05 were considered statistically significant. Results De novo lipogenesis is suppressed in human fibrosing conditions and restored upon WNT inhibition in human SSc Lipodystrophy of dermal adipocytes is an early and conserved feature of human and mouse fibrosis [13, 23, 24, 42, 43]. While emerging studies attributed fibrotic lipodystrophy to increased lipid breakdown [22, 24, 44], we evaluated whether suppression of lipid production pathway DNL occurred in human fibrotic skin. We performed a meta‐analysis of the Prospective Registry of Early Systemic Sclerosis (PRESS) cohort, which performed bulk RNA‐seq of 48 skin biopsies from patients with early, diffuse SSc with disease duration < 1 year (GSE130955; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE130955) [45]. We found significantly reduced expression of principal DNL enzymes FASN, acetyl Co‐A carboxylase alpha (ACACA), and stearoyl Co‐A desaturase (SCD2), while ATP citrate lyase (ACLY) was not statistically different (Figure 1A; supplementary material, Figure S1A). In addition, we found elevated expression of genes from the Wnt signaling pathway without the expense of mature adipocyte signature (supplementary material, Figure S1A). Similarly, in keloids and localized scleroderma – other fibrotic dermal conditions with elevated Wnt signaling – FASN was predominantly downregulated relative to healthy skin controls (GSE90051, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE90051; GSE153011, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE153011) (Figure 1B; supplementary material, Figure S1B,C) [46, 47]. Importantly, gene expression analysis of SSc patient biopsies in a short‐duration clinical trial for topical Wnt/β‐catenin inhibitor C‐82 revealed significant upregulation of FASN and ACACA (GSE94340, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE94340) [48] (Figure 1C; supplementary material, Figure S1D). Together, these findings establish FASN suppression as an early and conserved molecular event in SSc and localized scleroderma and keloids and that FASN expression is recoverable following Wnt pathway inhibition in human dermal fibrosis. Human fibrosing diseases downregulate de novo lipogenesis pathway in a WNT‐dependent manner and is recapitulated in WNT‐activated skin fibrosis mouse model. (A–C) Expression of FASN and ACACA by bulk RNA‐seq in forearm skin biopsies of healthy (n = 33) year 1 human systemic sclerosis (n = 48) (GSE130955), fold‐change in human keloid skin normalized to adjacent normal skin locations (n = 7) (GSE90051), and log2 fold‐change after 28‐day placebo or topical WNT‐inhibitor (C‐82) in diffuse cutaneous SSc forearm biopsies (n = 26) (GSE94340). (D) Schematic of adipocyte isolation for bulk RNA‐seq. Created in BioRender, https://BioRender.com/v49v527. (E) PCA of genes in control and Wnt‐activated adipocytes. (F) Volcano plot highlighting top differentially expressed genes (p < 0.05). (G) Top dysregulated pathways ranked by combined odds ratio (Enrichr Wiki pathways, p < 0.05). (H) Heatmap of DEGs related to lipid and fatty acid metabolism. (I) Schematic for DNL, a crucial pathway in adipocytes to convert excess glucose into triglycerides. Created in BioRender, https://BioRender.com/yvu92wq. (J) Expression (FPKM) values of key DNL enzymes ATP citrate lyase (Acly), acetyl Co‐A carboxylase (Acaca), fatty acid synthase (Fasn), stearoyl‐CoA desaturase (Scd) (n = 4 for 0‐ and 5‐day samples; n = 3 for 2‐day samples). p values were calculated with paired [for (C)] or unpaired, two‐tailed t‐test with Welch's correction. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Wnt signaling downregulates de novo lipogenesis preceding dermal adipocyte lipid depletion in mice To test whether Wnt signaling is sufficient to suppress DNL in murine dermal adipocytes in vivo, we employed bulk RNA‐seq to quantify early transcriptional changes in dorsal dermal adipocytes isolated by enzymatic digestion following 2 and 5 days of βcatistab transgene induction (Figure 1D). Effective upregulation of canonical Wnt target genes, including Axin2 and Lef1, at both time points confirmed βcatistab induction (supplementary material, Figure S2A). Principal component analysis (PCA) indicated that βcatistab adipocytes expressed a distinct transcriptome relative to control adipocytes (Figure 1E,F). Analysis of differentially expressed genes (DEGs) identified 475 DEGs at day 2 and 1,319 DEGs at day 5, with 114 genes shared at both time points (p ≤ 0.05) (supplementary material, Figure S2B). Analyses of the top 100 DEGs and the hierarchical clustering of all DEGs revealed a clear separation of genes responding early at 2 days and those sustained over 5 days of βcatistab activation (supplementary material, Figure S2C–E), suggesting complex temporal transcriptional responses within the dermal adipocytes following Wnt activation. Pathway analysis using Enrichr identified that Gene Ontology (GO) terms FA, cholesterol, and triglyceride synthesis were populated by downregulated DEGs in βcatistab dermal adipocytes (Figure 1G). Consistent with effective activation (supplementary material, Figure S2A), Enrichr identified upregulated DEGs related to Wnt signaling (Figure 1G). Moreover, GO analysis of DEGs using DAVID (https://davidbioinformatics.nih.gov/) similarly identified the enrichment of FA biosynthesis and lipid metabolic processes in 2‐day βcatistab adipocytes (supplementary material, Figure S2F). Genes related to extracellular region and ECM binding were upregulated at 2 days of βcatistab activation, with a 10‐fold increase in enrichment score after 5 days (supplementary material, Figure S2F,G). Importantly, the early molecular changes in lipid metabolism and subsequent robust increase in ECM signature support the kinetics of fibrosis progression observed in human skin fibrosis and animal models [23, 28, 43, 49]. We identified downregulation of the core genes encoding enzymes required for de novo FA synthesis: Acly, Acaca, and Fasn (Figure 1H). In addition, stearoyl‐CoA desaturase 2 (Scd2), which catalyzes desaturation of saturated fatty acyl‐CoAs, was also suppressed. Together, Acly, Acaca, Fasn, and Scd2 form a component of the DNL pathway that converts carbohydrate‐derived carbon into palmitate (16:0) and, subsequently, palmitoleate (16:1) (Figure 1I). Significant suppression of all four genes was observed in 2‐day βcatistab adipocytes and sustained throughout day 5 (Figure 1J). Importantly, key adipocyte markers peroxisome proliferator‐activated receptor gamma (Pparg), CCAAT enhancer binding protein alpha (Cebpa), perilipin 1 (Plin1), adiponectin, C1Q, and collagen domain containing (Adipoq) were not statistically different at both time points, indicating that early Wnt activation suppressed lipid biosynthesis without altering dermal adipocyte identity (supplementary material, Figure S3A). Transcriptional regulators of the DNL program, sterol regulatory element‐binding transcription factor 1 (Srebf1) and MLX interacting protein like (Mlxipl) was not significantly reduced in 2‐day βcatistab adipocytes (supplementary material, Figure S3B). However, genes regulating FA elongation, DNL‐specific glucose and citrate transport, and triglyceride synthesis were significantly downregulated (supplementary material, Figure S3C). In contrast, other prominent lipid metabolism pathways lipolysis and FA uptake and metabolically related pathways like glycolysis and FA oxidation (FAO) remained not statistically different (supplementary material, Figure S3D–G). Together, these findings indicate that selective transcriptional suppression of the DNL axis is an early and specific molecular signature of Wnt‐driven dermal dysfunction during skin fibrosis progression. Wnt‐induced downregulation of de novo lipogenesis is cell‐autonomous to mature dermal adipocytes To determine whether Wnt signaling suppresses the DNL pathway in a cell‐autonomous manner, we activated the Wnt pathway in mature mouse dermal adipocytes following differentiation of stromal vascular fraction (SVF) in vitro, as previously described [23] (Figure 2A). Treatment with WNT agonist (7 μm CHIR99021, Cayman Chemical Company, Ann Arbor, Michigan, USA, Catalog No.: 13122; or 20 μm BML284, Cayman Chemical Company, Catalog No.: 19903) for 5 days resulted in robust upregulation of Axin2, without affecting Plin1, confirming pathway activation without loss of adipocyte identity (supplementary material, Figure S4A). Both pharmacological treatments resulted in a significant reduction in adipocyte lipid accumulation, measured by Oil‐Red‐O staining (Figure 2B). Consistent with our in vivo results, 5 days of CHIR99021 treatment led to a significant decrease of solute carrier family 2 member 4 (Slc2a4, encoding glucose transporter GLUT4), Acly, and Fasn expression, as well as FASN protein (Figure 2C,D; supplementary material, Figure S4B). Mlxipl was dramatically reduced while Srebf1 expression was comparable to controls (Figure 2E). Morphometric analyses showed significant reductions in the number, area, perimeter, and circularity of lipid droplets in CHIR99021‐treated adipocytes (Figure 2F,G; supplementary material, Figure S4C,D). Thus, pharmacological activation of Wnt signaling is sufficient to suppress DNL pathway genes and diminish lipid droplet formation in an adipocyte‐autonomous manner. Cultured adipocytes undergo Wnt‐induced downregulation of de novo lipogenesis. (A) Differentiated adipocyte culture model. Created in BioRender, https://BioRender.com/i8sgrgn. (B) Oil red O staining of undifferentiated (U), differentiated (D), and 5‐day WNT agonist CHIR99021‐treated cells (D + W) (scale bar, 50 μm) and quantification (n = 6). (C) Fasn mRNA normalized to Hprt (n = 3). (D) FASN protein band and quantification normalized to GAPDH (n = 3). (E) Srebf1 and Mlxipl mRNA normalized to Hprt (n = 3–4). (F) Number of lipid droplets per field (n = 8) quantified using CellPose. (G) Histogram of individual lipid droplet areas (n = 1,724 droplets, n = 8). (H) Experimental design of WNT agonism, FASN inhibition, and 13C acetate pulse in vitro. Created in BioRender, https://BioRender.com/jq0ydho. (I) Representative images of undifferentiated (U), differentiated (D), 5‐day CHIR99021‐treated (D + W) and 5‐day FASN inhibitor TVB3664‐treated (D + F) cultures co‐stained with LipidSpot (lipid droplet, green) and DAPI (nucleus, blue) (scale bar, 100 μm) with inset (white square) showing high‐magnification single‐cell morphology (scale bar, 25 μm) and quantification (n = 8). (J) Triglyceride (TAG) quantification from cultures (n = 4). (K) Lipidomics analysis of WNT agonist BML‐284 and FASNi‐treated adipocytes showing NEFA subtypes normalized to protein (n = 4). (L) 13C‐isotope incorporation in 16:0 palmitate (n = 4). Unpaired two‐tailed Student's t‐test for two groups, one‐way ANOVA for three/four groups, and nonparametric Kruskal–Wallis for frequency distributions were applied. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. FASN enzyme catalyzes the seven reactions that convert carbohydrate‐derived carbons in the form of acetyl‐CoA and malonyl‐CoA substrates into the saturated FA palmitate [40, 50]. To determine whether loss of FASN function is sufficient to phenocopy Wnt‐mediated lipodystrophy, we treated adipocytes in vitro with FASN inhibitor (FASNi) TVB‐2640 (10 μm) for 5 days, and a 48‐h [13C]‐acetate pulse (Figure 2H). FASN inhibition significantly reduced neutral lipid accumulation (LipidSpot), triglyceride content, and lipid droplet area and perimeter, closely resembling the phenotype of BML284‐treated adipocytes (Figure 2I,J; supplementary material, Figure S4E–G). The [13C]‐acetate labeling showed that Wnt‐activated and FASN‐inhibited adipocytes exhibited comparable levels of total and nonesterified FAs (NEFAs), but there was significant reduction in labeled nonesterified palmitate (16:0), nonesterified polyunsaturated FAs (PUFAs), and saturation index (saturated: unsaturated FA; reflects SCD enzyme activity) (Figure 2K; supplementary material, Figure S4H,I and Table S1). Furthermore, both Wnt‐activated and FASN‐inhibited adipocytes incorporated significantly less [13C] tracer into the M + 2, M + 4, and M + 6 palmitate ions (Figure 2L). Together, these results demonstrate that FASN is functionally required for lipid accumulation in dermal adipocytes, and its inhibition is sufficient to phenocopy Wnt‐induced dermal adipocyte lipodystrophy. Wnt‐induced downregulation of de novo lipogenesis recovers in Wnt‐reversed mouse skin Given that FASN is suppressed early in human SSc, localized scleroderma, and keloids and recovers upon Wnt inhibition in human SSc (Figure 1H–J), we tested whether the DNL axis gene and protein expression was lost during βcatistab activation and restored following reversal of βcatistab activation in vivo. After 5 days of β‐catistab, DWAT thickness and individual PLIN+ adipocyte area and perimeter were significantly reduced relative to controls (Figure 3A–C; supplementary material, Figure S5A), indicating early and selective lipodystrophy following activation of the β‐catistab transgene. Using DWAT biopsies from intact mouse skin, we found Fasn, Acaca, and Mlxipl mRNA was significantly decreased, while Plin1 and Srebf1 were comparable to controls (Figure 3D,E; supplementary material, Figure S5B). The transcriptional start site (±3 kb from promoter) of human MLXIPL (but not SREBF1 or FASN) was predominantly enriched for TCF7L2, a β‐catenin‐binding partner in the TCF/LEF complex and contained a predicted LEF1 binding site in WNT‐overexpressing HCT116 and K‐562 cells (GSM782123, GSM2424109) [51, 52], as well as in human differentiated adipocytes (GSM3712534) [53, 54] (Figure 3E; supplementary material, S5C,D). These data suggest that MLXIPL can be a direct transcriptional target of Wnt/β‐catenin signaling, offering a mechanistic link between Wnt activation and DNL suppression in dermal adipocytes. FASN is downregulated in genetic and chemical mouse models of fibrosis. (A) Experimental design of 5‐day Wnt‐activated mice. Created in BioRender, https://BioRender.com/j2uswey. (B) Histology of p26 control and 5‐day βcatistab (scale bar, 100 μm) with quantification of dermal and DWAT thickness (n = 6). (C) Average individual adipocyte area in both groups (n = 5). (D) Fasn, Srebf1, and Mlxipl mRNA from DWAT biopsies of both groups normalized to ActB (n = 4–7). (E) Integrative Genomics Viewer snapshot of TCF7L2 binding in regions ±3 kb from transcription start site (arrow) of Mlxipl in WNT‐overexpressing HCT116 cells. (F) Immunofluorescence for FASN (red) and Perilipin1 (green) in control and 5‐day βcatistab DWAT (field of view indicated in schematic); arrows highlight FASN− adipocytes (scale bar, 50 μm). Created in BioRender, https://BioRender.com/a24r3st. (G) Percentage of FASN+ adipocytes in total counted per field (n = 6–7). Each symbol is the mean of three representative fields (crossed symbol) per animal. (H) Immunofluorescence for FASN and Perilipin1 in PBS‐ and 14‐day bleomycin (bleo)‐treated DWAT; arrows indicate FASN− adipocytes (scale bar, 50 μm). (I) Quantification of FASN+ adipocytes (n = 3–6). Statistical analysis was performed using unpaired Student's t‐test. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. To validate that Wnt signaling regulates FASN protein levels in dermal adipocytes in vivo, we assessed FASN protein in PLIN1+ adipocytes within the DWAT in the genetic β‐catistab and bleomycin‐induced skin fibrosis model (supplementary material, Figure S6A). In control mice, nearly all PLIN1+ adipocytes expressed FASN. In contrast, nearly half of PLIN1+ dermal adipocytes lacked detectable FASN in 5‐day βcatistab and 14‐day bleomycin‐treated DWAT (Figure 3F–I). The relative fluorescence intensity of FASN in bleomycin‐treated DWAT was also dramatically reduced compared to age‐matched controls (supplementary material, Figure S6B). To determine whether FASN downregulation depends on sustained Wnt activation, we examined FASN expression in the DWAT during an intermediate stage of recovery, prior to full resolution. After 10 days of reversal from β‐catistab transgene activation, we observed the re‐emergence and recovery of the DWAT layer thickness, individual adipocyte area, perimeter, and circularity (Figure 4A–E). The percentage of 10‐day reversal βcatistab DWAT adipocytes expressing FASN in PLIN1+ adipocytes tended to be lower but was comparable to that of controls (Figure 4F,G). Together, these findings demonstrate that FASN expression is dynamically regulated by elevated Wnt signaling in vivo and that recovery from dermal lipodystrophy is accompanied by restoration of FASN expression in mature adipocytes. FASN recovers after 10 days of reversal from established fibrosis. (A) Experimental design of 10‐day reversal from established fibrosis. Inset picture of established fibrosis at 21 days of Wnt activation. Created in BioRender, https://BioRender.com/631b3l1. (B) Masson's Trichrome‐stained mouse skin of 10‐day reversal βcatistab and age‐matched control with corresponding quantification of dermal and DWAT thickness (scale bar, 100 μm) (n = 6–10). (C) Quantification of individual adipocyte area, (D) perimeter, and (E) circularity in both groups (n = 5–6). (F) Immunofluorescence for FASN and Perilipin1 in control and 10‐day βcatistab reversal skin (scale bar, 50 μm) and (G) respective quantification of % FASN+ adipocytes (n = 4–5). p values were obtained using unpaired Student's t‐test with Welch's correction for bar and scatter plots. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Recovery from Wnt‐induced lipodystrophy and fibrotic ECM accumulation is dependent on FASN‐mediated DNL To test whether FASN‐dependent DNL might be functionally required for recovery from skin fibrosis, we pharmacologically inhibited FASN enzyme activity using TVB3664 (FASNi) during the reversal phase of βcatistab mouse model, when Wnt signaling is withdrawn and DWAT and dermal architecture normally recover (Figure 5A). FASNi‐reversal mice maintained normal body weight relative to untreated and vehicle‐treated controls (supplementary material, Figure S7A), but exhibited significantly thinner DWAT (Figure 5B,C). Interestingly, total skin thickness of FASNi‐reversal mice was significantly lower compared to vehicle‐treated controls (supplementary material, Figure S7B). Subsequent morphological analysis of adipocytes showed that FASNi‐reversal skin had smaller and less circular dermal adipocytes (Figure 5D,E; supplementary material, S7C,D). Biochemical analysis supported this morphological impairment: FASNi‐treated skin had significantly lower triglyceride levels (Figure 5F). Targeted lipidomics of the whole‐skin nonesterified FA (NEFA) fraction revealed marked decreases of PUFAs in FASNi‐reversal mice compared to vehicle control, while the quantitative sums for total FA, saturated FA, medium‐chain FA, and monounsaturated FA (MUFA) remained not statistically different (Figure 5G; supplementary material, Figure S7E,F and Table S2). These data indicate that FASN enzyme function is required, in part, for dermal adipocyte refilling, improvement of adipocyte morphology, and PUFA levels during recovery from established fibrosis‐associated lipodystrophy. FASN is functionally required for dermal adipocyte refilling after fibrotic lipodystrophy. (A) Schematic of FASN inhibitor TVB3664 oral treatment (3 mg/ml, 10 days) during reversal phase in mice. Created in BioRender, https://BioRender.com/glq3xn6. (B) Masson's Trichrome staining of healthy p52 control, 10‐day reversal βcatistab with vehicle gavage, and FASN‐inhibited 10‐day reversal βcatistab mouse dorsal skin (scale bar, 100 μm). (C) Dermal and DWAT thickness measurements (n = 5–7). (D) Immunostaining for FASN (red) and Perilipin1 (green); arrows point to lipid‐depleted dermal adipocytes (scale bar, 50 μm). (E) Histogram of adipocyte area in all three groups (393 random cells, n = 5–6). (F) Triglyceride quantity normalized to tissue weight (n = 4). (G) Lipidomics analysis of NEFA subtypes [total NEFA, medium chain FA (MCFA), saturated FA (SatFA), monounsaturated FA (MUFA), polyunsaturated FA (PUFA)] normalized to weight of tissue (n = 3). Statistical significance was determined by one‐way ANOVA and the Kruskal–Wallis test for frequency distributions. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Given the thickened dermis in FASNi‐reversal mice, we characterized the skin for ECM‐related fibrotic outcomes. We used atomic force microscopy to measure the mechanical stiffness of skin as a functional readout of skin fibrosis. Compared to vehicle treatment, the Young's modulus (kPa) of the lower dermis in FASNi‐reversal skin was significantly higher, suggesting increased ECM accumulation associated with delayed DWAT recovery (Figure 6A). We found increased collagen deposition in the dermis and DWAT of FASNi‐reversal skin compared to the vehicle group (supplementary material, Figure S8A). Specifically, we found that FASNi‐reversal skin contained a significantly lower amount of immature collagen fibers, leading to an elevated ratio of mature to immature collagen (Figure 6B; supplementary material, Figure S8B). PCA of collagen fiber properties analyzed using The Workflow Of Matrix BioLogy Informatics (TWOMBLI) [23, 55, 56] revealed that FASNi‐reversal skin was distinct from the other two groups, driven by high density matrix (HDM), or the amount of matrix in a fixed space (Figure 6C,D; supplementary material, Figure S8C,D). Alignment of collagen, a crucial fiber property involved in fibrosis onset and recovery [55, 56, 57], was significantly decreased in FASNi reversal using Alignment by Fourier Transform (AFT), an independent method to quantify alignment of collagen fibers (Figure 6E; supplementary material, Figure S8E). Excess sulfated proteoglycans are also indicators of skin fibrosis and recovery [58]. The area occupied by sulfated dermal proteoglycans (Alican blue, pH 2.5), but not highly acidic sulfated proteoglycans (pH 1.0), was significantly increased in FASNi‐reversal DWAT (Figure 6F; supplementary material, Figure S8F). Collectively, these data show that FASNi hampers lipogenesis in adipocytes during reversal and subsequently impedes recovery from fibrotic ECM. Our working model highlights the Wnt–DNL–FASN axis in adipocytes as a central regulator of both lipodystrophy onset and fibrosis recovery (Figure 6G). FASN is required to resolve fibrotic matrix during recovery from Wnt‐induced fibrosis. (A) Individual force curves (Young's modulus, kPa) in p52 control, 10‐day reversal βcatistab with vehicle gavage, and FASN‐inhibited 10‐day reversal βcatistab mouse dorsal skin (~300 curves from n = 3). (B) Herovici staining of mouse dorsal skin lower dermis (scale bar, 75 μm) with respective quantification of ratio of area covered by mature collagen (red) and immature collagen (purple) calculated after color deconvolution (n = 4–5). (C) Picrosirius red (PSR) staining visualized with polarized light microscopy for collagen topography in all three groups (scale bar, 200 μm); insets show high magnification of collagen fiber organization (scale bar, 50 μm). (D) Percentage high density matrix (HDM) quantified by TWOMBLI algorithm (n = 5–7). (E) Median order parameter (alignment) calculated using AFT (n = 5–7). (F) Alcian blue (pH 2.5) staining (scale bar, 200 μm) and quantification for sulfated proteoglycans in dermis and DWAT of mouse skin. (G) Working model illustrates Wnt‐induced downregulation of DNL in mature dermal adipocytes and its impact on ECM accumulation during skin fibrosis. DWAT, dermal white adipose tissue; ECM, extracellular matrix; MLXIPL, Mlx interacting protein like; FASN, fatty acid synthase; LD, lipid droplet. Created in BioRender, https://BioRender.com/wze48ft. Statistical significance was determined using one‐way ANOVA with appropriate post hoc tests; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Discussion Lipid depletion accompanies fibrotic skin conditions, as well as organ fibrosis [5, 59]. Current therapeutic approaches largely target immune and ECM dysregulation, yet clinical responses remain variable and incomplete [60, 61, 62]. This underscores the need to identify alternative, tissue‐intrinsic mechanisms that drive adipocyte remodeling and fibrotic progression. Here, we identify suppression of the FASN–DNL axis as an early and reversible event in human skin fibrosis, downstream of sustained Wnt signaling [48]. In mice, we demonstrate that transcriptional downregulation of FASN and other DNL axis enzymes is Wnt‐dependent and adipocyte‐autonomous. Furthermore, we find that FASN‐dependent FA synthesis is required for dermal adipocyte triglyceride replenishment and fibrotic ECM resolution during recovery from Wnt‐induced fibrosis (Figures 5 and 6). This finding is consistent with prior studies showing that FASN inhibition worsens bleomycin‐induced lung fibrosis in mice, while FA supplementation or LXR‐α agonists show protective potential against lung fibrosis [6, 63, 64, 65]. Together, we define a previously unrecognized Wnt–DNL axis in adipocyte‐driven skin fibrosis and prioritize FASN as a candidate target for therapeutic intervention. FAs have been implicated as either protective or profibrotic mediators, with their effects dependent on the tissue context and mode of fibrotic insult. In vitro, palmitic acid suppresses fibroblast proliferation and reduces expression of ECM genes collagen type 1 alpha chain (COL1A1) and elastin (ELN), suggesting antifibrotic potential [66]. Likewise, in vivo models of radiation‐induced skin fibrosis show dermal fibroblasts adopt a profibrotic metabolic program characterized by increased glycolysis and reduced FAO [21]. Enhancing FAO via CD36, a FA transporter, ameliorates fibrosis, although the source of FAs taken up by CD36hi fibroblasts remains unclear [21]. In both bleomycin‐ and Wnt‐induced fibrosis models, DWAT loss occurs via ATGL‐dependent lipolysis, releasing nonesterified ‘free’ FA (FFA) that influences ECM accumulation in divergent ways [5, 22, 23, 24, 44]. For example, systemic inhibition or global knockout of ATGL in bleomycin‐treated mice preserves DWAT but also enhances dermal thickness and collagen deposition, suggesting that FFAs may act to constrain ECM accumulation [22, 44]. However, in our tissue‐specific Wnt activation model, we found that ATGL‐dependent lipolysis in dermal adipocytes was required for fibrotic ECM expansion and dermal remodeling, indicating that dermal adipocyte‐derived FFAs may also promote ECM accumulation [24]. Pnpla3 (Adiponutrin), which has dual roles as a lipase and in lipid droplet formation, is highly downregulated in our dataset [67]. Variants of Pnpla3 accumulate triglycerides in the lipid droplet and is associated with liver fibrosis diseases and severity [67, 68]. Together, these seemingly contradictory findings point to a complex and context‐dependent role of FAs, lipid metabolism, and fibrotic ECM expansion, highlighting the need to further dissect distinct lipid metabolic pathways and identify lipid‐specific profiles that can mitigate poor fibrotic outcomes. The impact of Wnt signaling on the DNL‐axis transcriptional program in mature adipocytes appears to be dependent on tissue type and the molecular context of Wnt pathway manipulation. In vivo, mature adipocyte‐specific ablation of TCF7L2 enhances DNL and leads to significant adipocyte hypertrophy in white adipose tissue [54]. Similarly, Wnt activation suppresses DNL genes in the Drosophila ‘fat body’ adipocyte‐like cells [69]. Conversely, genetic deletion of β‐catenin in differentiated lipid‐filled mesenchymal cells in vitro suppresses DNL regulators and their target genes [53]. Several studies, including our own, have identified TCF7L2 binding at DNL gene loci (Figure 3E), supporting evidence that the DNL axis is a direct Wnt/β‐catenin target [53, 54, 69]. Our in vivo findings refine this understanding by demonstrating that Wnt activation in fibro‐adipocyte progenitors is sufficient to downregulate the DNL axis in mature dermal adipocytes, including FASN and ChREBP. Importantly, in the context of fibrotic lipodystrophy, repression of DNL and FASN is reversible. Future studies will be needed to determine how Wnt signaling alters chromatin accessibility and recruits co‐repressors to modulate the DNL program in dermal adipocytes. While our model provides strong evidence that sustained Wnt activation in skin suppresses DNL in mature dermal adipocytes and impairs adipocyte recovery, the causal operators – including FASN‐dependent lipid species and potential transcriptional co‐repressors of MLXIPL or SREBP1 – remain to be elucidated. We were limited by the absence of tools to selectively activate DNL or FASN in vivo. Since we cannot rule out effects of FASNi on other cell types in the skin in this study, future experiments should test the role of FASN specifically in dermal adipocytes during recovery from lipodystrophic DWAT. Furthermore, our lipidomic analyses were performed at the whole‐skin level, and future studies incorporating untargeted or spatial lipidomics approaches will be essential to map functional lipid changes or to distinguish adipocytes from fibroblastic stroma. Nevertheless, our findings provide a mechanistic context for clinical studies showing restoration of DNL genes, including FASN, in SSc patients treated with the topical Wnt inhibitor C‐82 [48] or with mycophenolate mofetil (MMF) [33]. This underscores the need for translational strategies that move beyond global Wnt inhibition, which may have systemic effects and has shown limited clinical efficacy [70, 71]. Our data support an alternative approach to restoring the lipogenic capacity of dermal adipocytes to reverse lipodystrophy and fibrotic ECM remodeling. Despite the lack of clarity on precise mechanisms, emerging regenerative therapies, such as fat grafting, adipocyte stem cell treatment, extracellular vesicles, and exosomes, have shown promise for treating fibrotic skin conditions [72, 73, 74, 75, 76, 77, 78, 79]. In conclusion, our studies underscore FASN as an important marker of early lipodystrophy preceding fibrotic progression, whose activity or downstream signaling lipids may be useful for promoting fibrotic lipodystrophy recovery. Author contributions statement All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. RPA had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. SRM, RPA and MCR were responsible for study conception and design. SRM, PM, RV, MM, QM, KK, EGE and RPA acquired the data. SRM, RPA, MCR, PM, RV, MM, EE and CRK analyzed and interpreted the data. SRM, RPA and MCR wrote the manuscript, which was edited by all the authors. All the authors approved the final version of the manuscript. Supporting information Supplementary materials and methods Figure S1. Mining different human sequencing datasets for de novo lipogenesis genes and enzymes Figure S2. Validation and visualization of bulk RNA‐seq of 2‐ and 5‐day βcatistab dermal adipocytes Figure S3. Validation of related lipid metabolism signatures from bulk RNA sequencing of 2‐ and 5‐day βcatistab dermal adipocytes Figure S4. Wnt activation in differentiated adipocytes phenocopies FASN inhibition Figure S5. TCF7L2 and LEF1 sites on promoter regions of FASN and DNL transcription factors in adipocytes and cell lines with sustained Wnt activation Figure S6. Downregulation of FASN is accompanied by histo‐morphometric changes in genetic and chemical model of fibrosis in vivo Figure S7. FASN inhibition during Wnt recovery does not allow recovery of dermal adipocytes Figure S8. Collagen and highly sulfated proteoglycans in FASN‐inhibited skin Table S1. Targeted lipidomics in vitro Table S2. Targeted lipidomics in vivo Acknowledgements We thank all the past and present members of the Atit lab for their input and feedback on this project. We thank Morgan Horowitz for quantification of Picrosirius red staining using TWOMBLI, Rachel Kim and Emilia Sanz‐Rios for quantification of Trichrome staining, Megan Gregory for genotyping the mice, and Saada Eid for technical support with the bleomycin mouse model. Special thanks go to Anna Jussila, Valerie Horsley, David Buchner, Timothy Mead, Rodrigo Somoza‐Palacios and William Spencer for technical support and advice. We thank the bio[box] shared instrumental facility, Tissue resources core and Genomics core facility at Case Western Reserve University and CWRU‐ENGAGE program for summer fellowships (PM). This project was funded by the National Institutes of Health, National Institute of Arthritis and Musculoskeletal and Skin Diseases R01 AR076938 (RPA and Valerie Horsley), National Scleroderma Foundation (SRM and QM), NSF‐MRI grant (CK) to purchase atomic force microscope, R56DK139005‐01A1 (MCR), Oklahoma Center for Adult Stem Cell Research (MCR), and Presbyterian Health Foundation Bridge Grant (MCR). No conflicts of interest were declared. Contributor Information Michael C Rudolph, Email: michael-rudolph@ou.edu. Radhika P Atit, Email: rpa5@case.edu. Data availability statement Sequencing and analysis data were deposited in the National Center for Biotechnology Information Gene Expression Omnibus under accession number GSE305616 (mouse dermal adipocyte bulk RNA‐seq; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE305616), GSE130995 (human SSc RNA‐seq data; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE130995), GSE90051 (keloid microarray data; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE90051), and GSE94340 (human C‐82 clinical trial microarray data; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE94340). References References [80, 81, 82, 83, 84, 85, 86, 87] are cited only in the supplementary material. Associated Data Supplementary Materials Supplementary materials and methods Figure S1. Mining different human sequencing datasets for de novo lipogenesis genes and enzymes Figure S2. Validation and visualization of bulk RNA‐seq of 2‐ and 5‐day βcatistab dermal adipocytes Figure S3. Validation of related lipid metabolism signatures from bulk RNA sequencing of 2‐ and 5‐day βcatistab dermal adipocytes Figure S4. Wnt activation in differentiated adipocytes phenocopies FASN inhibition Figure S5. TCF7L2 and LEF1 sites on promoter regions of FASN and DNL transcription factors in adipocytes and cell lines with sustained Wnt activation Figure S6. Downregulation of FASN is accompanied by histo‐morphometric changes in genetic and chemical model of fibrosis in vivo Figure S7. FASN inhibition during Wnt recovery does not allow recovery of dermal adipocytes Figure S8. Collagen and highly sulfated proteoglycans in FASN‐inhibited skin Table S1. Targeted lipidomics in vitro Table S2. Targeted lipidomics in vivo Data Availability Statement Sequencing and analysis data were deposited in the National Center for Biotechnology Information Gene Expression Omnibus under accession number GSE305616 (mouse dermal adipocyte bulk RNA‐seq; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE305616), GSE130995 (human SSc RNA‐seq data; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE130995), GSE90051 (keloid microarray data; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE90051), and GSE94340 (human C‐82 clinical trial microarray data; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE94340).