Four-Gene Fatty Acid Metabolism Signature Distinguishes Diabetic Foot Ulcers With High Diagnostic Accuracy

A comprehensive integrative transcriptomic analysis has identified a four-gene fatty acid metabolism (FAM)-related signature that distinguishes diabetic foot ulcer (搜索)搜索该疾病 (DFU) tissue from healthy control skin with high discriminatory accuracy, offering candidate diagnostic biomarkers and therapeutic targets for one of the most debilitating complications of diabetes mellitus (搜索)搜索该疾病. The study, combining bulk RNA-seq and single-cell RNA-seq (scRNA-seq) data, found that DFUs are characterized by cell-type-specific FAM dysregulation—primarily enhanced fatty acid oxidation (FAO) and reduced fatty acid synthesis (FAS)—alongside a pro-inflammatory immune landscape.

Metabolic Reprogramming in Diabetic Foot Ulcers

Diabetic foot ulcers are chronic, non-healing wounds primarily affecting the lower extremities that lead to substantial morbidity, lower limb amputations, and increased mortality. Despite progress in wound care modalities and multidisciplinary management, the pathogenesis of DFUs remains poorly understood. Fatty acids play a crucial role in inflammation, oxidative stress, and tissue repair, and dysregulation of fatty acid metabolism is a hallmark of diabetes mellitus (搜索)搜索该疾病.

The study's bulk transcriptomic analysis showed increased FAO-related activity but reduced FAS-related activity in DFU tissues. The authors note this pattern "may reflect a transcriptional shift from lipid synthesis toward FA catabolism under the hypoxic, inflammatory, and metabolically stressed conditions of diabetic wounds." The increase in FAO was mainly attributable to β- and ω-oxidation rather than α-oxidation, and long-chain fatty acid metabolism was more active than medium- or short-chain metabolism.

Importantly, the researchers caution that the pathway scores "represent relative transcriptional activity and should not be interpreted as direct measurements of metabolic flux, FA oxidation rates, or lipid synthesis."

Cell-Type-Specific Heterogeneity

Single-cell analysis of 19,574 cells from five non-healing DFU samples and eleven healthy non-diabetic control samples revealed that FAM-related alterations were heterogeneous across cell populations. FAM levels were elevated in all cell types within the DFU group compared to controls, with the exception of fibroblasts. Most immune-cell populations exhibited higher FAM-related scores in DFU samples, whereas fibroblasts and keratinocytes showed lower FAO- and FAS-related activity.

AUCell analysis demonstrated that epithelial cells exhibited the highest FAM and FAO activities, whereas monocytes showed the highest FAS activity. The authors interpret the higher scores in immune cells as potentially reflecting "metabolic adaptation to the hyperglycemic, inflammatory, and hypoxic wound microenvironment," while reduced FAO- and FAS-related programs in fibroblasts and keratinocytes "may be associated with cell-type-specific metabolic dysfunction, impaired mitochondrial activity, or altered tissue-repair capacity."

Immune Dysregulation and Metabolic-Immune Crosstalk

CIBERSORT analysis revealed distinct immune-cell infiltration patterns between DFU and control groups. In the GSE7014 discovery cohort, DFU samples exhibited significantly increased infiltration of M1 macrophages, resting mast cells, and naïve CD4+ T cells, whereas resting natural killer (NK) cells and activated dendritic cells (DCs) were significantly reduced. Correlation analysis showed an inverse association between M1 macrophages and activated DCs, with a correlation coefficient of −0.45.

The DFU group also demonstrated higher type I interferon responses and Th2-cell-related activity but lower type II interferon responses, T-cell and antigen-presenting-cell co-stimulation, and cytolytic activity than controls. The authors conclude that "DFUs are characterized by enhanced pro-inflammatory immune-cell infiltration accompanied by impaired immune-regulatory and cytotoxic functions."

Comparing diabetes subtypes, T2DM-associated DFU samples exhibited significantly greater M1 macrophage infiltration and lower activated DC infiltration than T1DM-associated DFU samples, alongside lower FAS activity but higher fatty acid catabolism. These findings suggest T2DM-associated DFUs exhibit "a more pronounced shift toward FA catabolism together with a more pro-inflammatory immune profile."

Four-Gene Diagnostic Signature

Using two machine-learning approaches—LASSO logistic regression (10-fold cross-validation repeated 1000 times) and Random Forest—the researchers identified four final marker genes: Factor H (FH) (搜索)搜索该靶点, Leukotriene C4 Synthase (LTC4S) (搜索)搜索该靶点, Glutamate-Ammonia Ligase (GLUL) (搜索)搜索该靶点, and Gamma-Aminobutyric Acid Receptor-Associated Protein Like 1 (GABARAPL1 (搜索)搜索该靶点).

Compared with controls, FH was significantly upregulated in DFU samples, whereas GLUL, LTC4S, and GABARAPL1 (搜索)搜索该靶点 were significantly downregulated. The same expression patterns were reproduced in the independent GSE134431 validation cohort.

A multivariable logistic regression model incorporating the four genes achieved an AUC of 0.992 in the GSE7014 discovery cohort and 0.982 in the GSE134431 validation cohort. Individually, all four genes achieved AUC values greater than 0.90 in the discovery cohort and greater than 0.75 in the validation cohort. Calibration analysis showed good agreement between model-predicted probabilities and observed outcomes in both cohorts.

The authors emphasize that the AUC values approaching 1.0 "suggest strong separation between the DFU and control samples in the analyzed datasets and should be interpreted cautiously," and that the genes "should be regarded as candidate transcriptomic biomarkers."

Biological Roles and Clinical Implications

The four genes capture interconnected alterations in distinct biological pathways. FH is a key regulator of the complement system and was positively correlated with estimated M1 macrophage infiltration (R = 0.35, P < 0.05). GLUL encodes glutamine synthetase and was negatively associated with M1 macrophage abundance (R = −0.59, P < 0.001) while positively correlated with activated DC infiltration (R = 0.47, P < 0.01). GABARAPL1 (搜索)搜索该靶点 participates in autophagy-related processes, and LTC4S encodes a key enzyme in leukotriene biosynthesis.

From a translational perspective, the authors suggest the four-gene signature "could be developed as a tissue-based assay to molecularly stratify DFU lesions according to their metabolic–immune states, thereby complementing conventional clinical assessment and defining biologically distinct wound phenotypes for future prognostic and treatment-response studies."

Study Limitations

The authors acknowledge several limitations. The study was based on three publicly available transcriptomic datasets with relatively limited sample sizes, and systemic metabolic factors—including blood glucose levels, glycemic control, insulin resistance, and circulating lipid profiles—could not be reliably adjusted for because these variables were incompletely reported. All datasets were cross-sectional, precluding assessment of temporal changes during DFU progression or healing. The study did not experimentally validate the functional mechanisms or the scRNA-seq findings in vivo, and tissue lipidomic data were unavailable to quantify individual fatty acid species.

The authors conclude that "larger prospective multicenter studies enrolling diverse DFU populations across different regions and clinical settings and incorporating detailed clinical data, validated patient-reported outcome measures, longitudinal sampling, direct metabolic measurements, and functional validation are required to confirm these findings and clarify their biological and clinical relevance."