Advanced single-cell transcriptomics deciphers cellular complexity and MIF-orchestrated signaling networks in diabetes-induced myocardial disease

Background: Cardiac complications arising from diabetes mellitus manifest as structural and functional myocardial alterations independent of traditional cardiovascular risk factors. The intricate molecular underpinnings driving disease evolution and the spectrum of cellular diversity remain inadequately characterized. Our investigation sought to systematically elucidate the transcriptional architecture and intercellular signaling frameworks in diabetes-associated myocardial dysfunction through integrated omics methodologies. Methods: We executed parallel bulk and single-cell transcriptomic profiling of cardiac specimens from diabetic disease models. Gene expression disparities were determined via DESeq2 employing stringent thresholds (|log₂FC| > 1; FDR < 0.05). Quality control applied specific thresholds (200–6000 genes/cell, <20% mitochondrial content), with clustering resolution optimized at 0.8 for main cell types and 1.2 for fibroblast subclustering. Single-cell datasets underwent Seurat-based processing with t-distributed stochastic neighbor embedding for population delineation. WGCNA employed soft-thresholding (R² > 0.85), minimum module size of 30 genes, and merge cut height of 0.25. Co-expression module detection within fibroblast subsets was achieved through weighted correlation network construction. Functional annotation leveraged GO and KEGG repositories. CellChat analysis incorporated permutation-based significance testing (n = 100, p < 0.05) with CellPhoneDB validation. Intercellular signaling topology was reconstructed using CellChat, emphasizing macrophage migration inhibitory factor circuitry. Results: Transcriptional profiling unveiled 2000 dysregulated transcripts against 28,840 stable genes, demonstrating substantial reprogramming during pathogenesis. Single-cell resolution exposed profound cellular heterogeneity encompassing myocytes, endothelium, fibroblasts, myeloid cells, and specialized populations including metabolic coordinators and stress-activated subsets. Granular fibroblast dissection revealed 21 molecularly distinct subtypes (designated M1-M21), underscoring remarkable intra-lineage diversity. Enrichment analyses highlighted perturbations in matrix architecture, inflammatory cascades, and proliferative control. Network analysis identified co-regulated gene clusters governing matrix remodeling, inflammation, and metabolic homeostasis. Communication mapping positioned MIF signaling as a pivotal intercellular coordination axis, with stress-responsive cells functioning as nodal integrators throughout disease progression. Conclusions: This integrative multi-platform investigation provides comprehensive molecular characterization of diabetes-induced cardiac pathology, revealing extensive cellular heterogeneity and intricate communication networks that extend previous single-cell cardiac studies.