Rump and tail adipose tissue was aseptically dissected from ovine fetuses (approximately 60–90 days of gestation, E60-E90) of Kazakh sheep (fat-rumped) and Hu sheep (fat-tailed) breeds (22, 23). Fetuses were collected from the Western Pastoral Slaughterhouse in Shihezi, Xinjiang. Specifically, adipose tissue from three Kazakh sheep fetuses at about E75 was for the primary isolation of adipose-derived stem cells (ADSCs) used in subsequent functional experiments. The collected tissues were immediately washed four times with phosphate-buffered saline (PBS) (Solarbio, Beijing, China) supplemented with 1% penicillin–streptomycin (Solarbio, Beijing, China). All animal experimental procedures were approved by the Biology Ethics Committee of Shihezi University (approval number is: A2020-34).

ADSCs were isolated as previously described (24), with the following critical modifications to optimize yield and viability. Briefly, the collected adipose tissues from the tail or rump were immediately transferred to a biosafety cabinet and placed into a new sterile Eppendorf tube. The tissue was washed three times with sterile phosphate-buffered saline (PBS) (Solarbio, Beijing, China) containing 1% penicillin/streptomycin to remove debris and red blood cells. Subsequently, it was minced into fine fragments (approximately 1 mm³ in size) using surgical scissors and digested with 0.5 mL of 0.25% trypsin for 3 min at room temperature. The digestion was terminated by adding twice the volume of complete culture medium. The digest was then filtered through a 100 μm cell strainer, and the filtrate was centrifuged at 1000 × g for 5 min. After discarding the supernatant, the pellet was resuspended in 1 mL of complete medium and plated into a 60 mm culture dish. Cells were incubated at 37 °C in a 5% CO₂ atmosphere. The medium was replaced after 24 h to remove non-adherent cells and debris, and subsequently every 2 days thereafter. Primary cells were passaged upon reaching 80% confluence. Cells from passages 5 (P5) were used for all subsequent experiments to ensure consistency and avoid senescence-related effects.

For passaging, cells were washed once with PBS to remove residual serum. Pre-warmed 0.25% trypsin (TransGen Biotech, Beijing, China) was added to cover the cell layer and incubated at 37 °C for 90 s. Cell detachment was monitored under a microscope until most of the cells had rounded up and detached from the surface. The enzymatic reaction was promptly stopped by adding twice the volume of complete medium. The cell suspension was collected, centrifuged at 300 × g for 5 min, and the pellet was resuspended in fresh complete medium for reseeding.

Cells at passages 5 (P5) were characterized to confirm their identity as mesenchymal stem cells, based on their adherence to plastic, multilineage differentiation potential, and surface marker expression profile. Multilineage Differentiation Potential: ADSCs were induced toward adipogenic and osteogenic lineages as detailed in section 2.5.

Immunofluorescence Staining for ADSCs Surface Marker Identification: To morphologically confirm the stem cell properties of the isolated cells, we employed immunofluorescence staining to detect the expression of the ADSCs surface marker CD44. Specifically, P5 ADSCs were seeded in a 24-well plate at an appropriate density and allowed to grow to 70% confluence before fixation and staining. Cells were first fixed with 4% paraformaldehyde (BioSharp, Hefei, China) at room temperature for 20 min, followed by permeabilization with 0.2% Triton X-100 (Sigma-Aldrich, Germany) for 10 min to increase membrane permeability. To reduce non-specific binding, cells were blocked with 1% bovine serum albumin (BSA, Bioss, Beijing, China) at room temperature for 30 min. After blocking, cells were incubated overnight at 4 °C with a primary antibody against CD44 (rabbit polyclonal antibody, Bioss, Cat No. bs-55039R, dilution 1:1,000). The next day, after thorough washing with PBS, cells were incubated with a CoraLite488-conjugated goat anti-rabbit IgG secondary antibody (Wuhan Sanying, Cat No. SA00013-2, dilution 1:100) at room temperature in the dark for 1 h. Finally, nuclei were counterstained with DAPI (Sigma-Aldrich, Germany) for 10 min. Following staining, images were captured using a BX43 fluorescent microscope (Olympus Corporation, Japan).

Adipogenic Differentiation: P5 ADSCs were induced to differentiate when they reached 95% confluence. The adipogenic induction medium, consisting of 90% DMEM/F12, 10% FBS, 10 ng/mL insulin, 1 μM dexamethasone, 1 μM rosiglitazone, and 0.5 mM IBMX, was applied for 2 days. The medium was then replaced with an adipogenic maintenance medium (90% DMEM/F12, 10% FBS, 10 ng/mL insulin, and 1 μM rosiglitazone), which was refreshed every 2 days until day 6. On day 6, cells were fixed with 4% paraformaldehyde for 20 min and stained with a 0.4% Oil Red O solution (Beyotime, Beijing, China) for 1 h to visualize lipid droplets. The expression of adipogenic genes (PPARr, INHBA, PLIN1 and FABP4). and PDGFD and its receptor genes were analyzed by RT-qPCR at days 0, 2, 4, and 6 of differentiation (Supplementary Tables S2 for primers).

Osteogenic Differentiation: P5 ADSCs were induced at 80% confluence. The growth medium was replaced with an osteogenic induction medium (growth medium supplemented with 100 ng/mL ascorbic acid, 20 mM β-glycerol phosphate, and 200 nM dexamethasone), which was refreshed every 2 days. On days 6 and 12 post-induction, cells were fixed with 4% paraformaldehyde and stained with a 0.2% Alizarin Red S solution (BBI Life Sciences Co., Ltd., Shenggong, Shanghai, China) for 1 h at room temperature in the dark to detect mineralized matrix deposition. Staining was visualized and imaged under a light microscope.

Three small interfering RNAs (siRNAs) targeting PDGFD and a non-targeting scram-bled negative control (NC) siRNA were designed and synthesized by BBI Life Sciences (Shanghai, China). Sequences are listed in Supplementary Tables S3. The sheep P5 ADSCs, which were cultured, purified, and characterized from sheep embryos, were seeded on 12-well plates 1 day prior to transfection to reach 60%–70% confluence. Three PDGFD siRNAs and negative control siRNA (siNC) were transfected at a final concentration of 20 nM per well using EL Transfection Reagent (BBI Life Sciences). The knockdown efficiency of each siRNA was assessed by RT-qPCR as described in section 2.9.4.

For the cell proliferation assay, ADSCs were seeded in 96-well plates at a density of 2.0 × 10⁴ cells per well. The siRNA with the highest knockdown efficiency (si-PDGFD-1427) was selected to establish the PDGFD knockdown model in P5 ADSCs. Cells were transfected using a mixture containing the siRNA and transfection reagent in 10% Opti-MEM (BBI Life Sciences). Cell proliferation was then assessed using a Cell Counting Kit-8 (CCK-8; Beyotime, Beijing, China) according to the manufacturer’s instructions. Briefly, 10% CCK-8 solution was added to the culture medium, followed by incubation at 37 °C for 2 h. The optical density (OD) at 450 nm was measured for both the si-PDGFD-1427 and siNC groups using a microplate reader (Thermo Fisher, MA, United States) at 24, 48, and 72 h post-transfection. Each treatment group was assayed with five replicates.

Total RNA was extracted from ADSCs transfected with si-PDGFD or NC siRNA for 24 and 48 h using Trizol (TransZol Up kit, TransGen Biotech, Beijing, China). Three independent biological replicates were performed for each group and time point. RNA quality was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, United States), and only samples with an RNA Integrity Number (RIN) > 7.0 were used for library preparation. Sequencing libraries were constructed and sequenced on an Illumina platform by a commercial service provider. Raw sequencing reads were quality-filtered using FastQC (v0.11.5) to obtain clean data, which were then aligned to the sheep reference genome (GCF_000298735.2_Oar_v4.0_genomic) using HISAT2 (v2.0.4). Transcript assembly and quantification were performed with StringTie (v1.2.3). Differentially expressed genes (DEGs) were identified using the DESeq2 package (v1.38.3). A model matrix was designed with a single variable indicating the treatment group and si-PDGFD group. To control the false discovery rate (FDR) arising from multiple hypothesis testing, the p-values were adjusted using the Benjamini-Hochberg procedure. Genes with an adjusted p-value (FDR) < 0.05 and an absolute log₂ fold change > 1 were considered statistically significant differentially expressed genes (DEGs).

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses (25) were performed on the differentially expressed genes (DEGs) using the clusterProfiler package in R. Additionally, Gene Set Enrichment Analysis (GSEA) was conducted on the full ranked gene list to identify KEGG pathways with subtle but coordinated expression changes (26). The ggplot2 package is then utilized to visualize the enrichment results. Based on the KEGG enrichment results, the PI3K-Akt and MAPK signaling pathways were selected for further analysis. A protein–protein interaction (PPI) network was constructed using the online STRING database¹ for key DEGs involved in these pathways, including CXCL8.

The specific statistical tests used are detailed in the respective figure legends. Pairwise comparisons between two groups were analyzed using an unpaired two-tailed Student’s t-test. Comparisons among three or more groups were analyzed using one-way ANOVA, followed by Tukey’s post-hoc test for multiple comparisons. Sample sizes (n) for each experiment, representing biological replicates, are indicated in the figure legends.

To define the key developmental period for tail fat deposition, we analyzed fetuses of different tail types at various gestational ages using dissection, Oil Red O staining on frozen sections, and H&E staining on paraffin sections. Our analysis confirmed that visible tail fat formation and lipid droplet accumulation commence at approximately E80–85 days of gestation (corresponding to a fetal body length of 25 cm) (Figure 1; Supplementary Figure S1). Importantly, we identified an earlier, critical time point at E70-75 days of gestation (corresponding to a fetal body length of 16 cm), establishing this as the pivotal window for the onset of rump fat development in fat rumped sheep (Figure 1; Supplementary Figure S1).

Concomitantly, we profiled the expression of PDGFD and its receptors, PDGFRα and PDGFRβ, in embryonic tail and rump adipose tissues, we found that PDGFD expression exhibited a decreasing trend with increasing gestational age in both tail types (Supplementary Figure S2). We also observed that the expression level of PDGFD in the rump fat of fat-tailed sheep was higher than that in long-fat-tailed sheep. The expression pattern of the receptor PDGFRα was largely mirrored that of PDGFD. In contrast, PDGFRβ exhibited an opposing expression trend between the two tail types: it increased with gestational age in fat-rumped sheep but decreased in fat-tailed sheep (Supplementary Figure S2).

To mechanistically dissect PDGFD function, we established a physiologically relevant in vitro model. ADSCs were successfully isolated from embryonic sheep tail tissue and displayed characteristic fibroblast-like morphology (Supplementary Figure S3a–d). Molecular characterization confirmed expression of classic mesenchymal stem cell markers (CD44, CD90, CD29, CD106) and the absence of the endothelial marker CD31. Weak expression of the hematopoietic marker CD45 was also detected (Supplementary Figure S4).

Furthermore, immunofluorescence analysis of the canonical MSC marker CD44 revealed strong positive staining that was localized to the cell membrane and cytoplasm, as expected. Counterstaining with DAPI confirmed nuclear integrity and cell density (Figure 2). Collectively, positive expression of CD44, CD106, CD29, and CD90, and the absence of CD31 and CD45 expression confirm the successful isolation and culture of a purified population of ovine ADSCs.

To assess their multilineage potential, fifth-passage ADSCs were subjected to osteogenic and adipogenic induction. Following osteogenic induction, ADSCs underwent a notable morphological shift, transitioning from a spindle-shaped to a more polygonal and cuboidal morphology, which was evident by day 3 and pronounced by day 6. Alizarin Red S staining on day 12 confirmed successful differentiation, revealing the presence of dense, dark red mineralized nodules, indicating extensive calcium deposition (Supplementary Figure S5). These results demonstrate the robust osteogenic differentiation capacity of the isolated ovine ADSCs.

Similarly, adipogenic induction also induced rapid morphological changes. Within 2 days, cells began to accumulate small, intracellular lipid droplets. By day 6, these droplets had coalesced into numerous large, spherical, and translucent droplets that filled the cytoplasm. Oil Red O staining on day 6 confirmed these structures to be neutral lipids, with bright red staining clearly visible (Supplementary Figure S6). This confirms the strong adipogenic potential of the ADSCs isolated from fetal sheep tail tissue.

To characterize the molecular events during adipogenesis, we profiled the expression of key adipogenic markers and the PDGFD signaling pathway in ADSCs undergoing differentiation (days 0, 2, 4, and 6). As expected, the expression of adipogenic genes was successfully induced, with each marker peaking at a distinct stage of differentiation (Figure 3). The master regulator PPARγ and the INHBA showed the highest expression at day 2. This was followed by a peak in the fatty acid-binding protein FABP4 at day 4, and finally, the lipid droplet-coating protein PLIN1 at day 6 (Figure 3).

Concurrently, we analyzed the expression of PDGFD and its receptors. A striking inverse correlation was observed: as adipogenic commitment progressed, the mRNA levels of both PDGFD and its receptor PDGFRα progressively decreased (Figure 3). In contrast, the expression level of PDGFRβ remained relatively stable throughout the differentiation process and did not show statistically significant variation (Figure 3).

To elucidate the global transcriptional changes and transcriptional network regulated by PDGFD, we performed RNA-seq on ADSCs 48 h after transfection with si-PDGFD and si-NC. Knockdown of PDGFD resulted in 915 differentially expressed genes (DEGs) (p ≤ 0.05, |log₂FoldChange| ≥ 1), comprising 451 upregulated and 464 downregulated genes (Figure 5a; Supplementary Table S4). Notably, the most significantly altered genes included the profoundly downregulated chemokine CXCL8 and the upregulated serine protease inhibitor SERPINE2, and the adipogenic markers PPARγ and FABP4 were also significantly upregulated (Figure 5b).

Gene Ontology (GO) enrichment analysis of all DEGs revealed a strong association with biological processes including cell adhesion, inflammatory response, and MAPK cascade (Figure 5; Supplementary Table S4). When analyzed separately, the downregulated genes were predominantly linked to inflammatory response, cell adhesion, and MAPK cascade, while upregulated genes were enriched in processes such as transcription and phosphorylation (Figure 5d; Supplementary Tables S6, S7).

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis further demonstrated that PDGFD knockdown significantly impacted pathways related to signal transduction and the immune system (Figure 5e). Crucially, this analysis confirmed the enrichment of the PI3K-Akt and MAPK signaling pathways, established downstream cascades of PDGF signaling. A focused analysis showed that upregulated genes were involved in metabolic pathways like the PPAR signaling pathway, whereas downregulated genes were strikingly overrepresented in immune and signaling pathways, including the PI3K-Akt signaling pathway, Rap1 signaling pathway, and Toll-like receptor signaling pathway (Figure 5f).

To further capture more subtle and coordinated expression changes across the entire transcriptome, we performed Gene Set Enrichment Analysis (GSEA) on all genes. This analysis corroborated and extended the findings above, showing a significant positive enrichment (normalized enrichment score, NES = −1.17, p < 0.05) for the MAPK signaling pathway, alongside significant enrichment of metabolic pathways such as PPAR signaling, fatty acid metabolism, and tyrosine metabolism (Supplementary Figure S10; Supplementary Table S8). Notably, the GSEA results specifically highlighted the enrichment of PDGFD’s cognate receptor, PDGFRβ, within the MAPK signaling pathway gene set, providing direct bioinformatic evidence linking PDGFD knockdown to the perturbation of this key downstream cascade. This independent, global analysis underscores the central role of PDGFD in regulating both key signaling and metabolic programs in ADSCs.

To confirm the reliability of our transcriptomic data, we performed RT-PCR on a panel of key genes using the same RNA samples submitted for sequencing. We also included an earlier time point (24 h) to track initial transcriptional changes. The expression trends for all tested genes at 48 h were fully consistent with the RNA-seq data (Figure 6). This included the significant upregulation of the adipogenic markers PPARγ and FABP4, and the pronounced downregulation of CXCL8, INHBA, BCL2A1 and PDGFD itself.

To visualize the impact of PDGFD knockdown on specific signaling cascades, we constructed schematic diagrams of the PI3K-Akt and MAPK pathways based on our KEGG enrichment results, mapping the differentially expressed genes (DEGs) onto their respective pathway positions (Supplementary Figures S7, S8). In the PI3K-Akt signaling pathway, a predominant downregulation of genes was observed. Key affected genes included several phosphatidylinositol kinases such as PIK3CG, PIK3CD, PIK3R1, PIK3R5, and PIK3AP1 (Supplementary Figure S7). The MAPK signaling pathways exhibited a more mixed regulatory pattern. Notably, genes involved in the ERK cascade like ERK1 and MAPK8IP1were upregulated. In contrast, several genes associated with the JNK cascade (MAP3K13, MAP3K5, and MAPK8) were downregulated (Supplementary Figure S8).

Given its status as the most significantly downregulated gene and its presence in the enriched KEGG pathways, we investigated the connectivity of CXCL8. A protein–protein interaction (PPI) network was generated using the STRING database, incorporating CXCL8 and the DEGs from the PI3K-Akt and MAPK signaling pathways. The resulting network revealed a high-confidence (interaction score > 0.7) and interconnected web of proteins (Supplementary Figure S9). CXCL8 was a central node within this network, demonstrating multiple direct and indirect interactions with other downstream effectors, including components of both the PI3K-Akt and MAPK pathways.