Distinct Drivers of Type 2 Diabetes: A Systems Modeling Approach to Molecular, Genetic, and Metagenomic Insights

Abstract

Type 2 Diabetes (T2D) is a complex and heterogeneous disease characterized by hyperglycemia, insulin resistance, and impaired insulin secretion. While often interrelated, hyperglycemia and insulin resistance represent distinct pathophysiological processes that contribute differently to disease onset, progression, and prognosis. Current methods lack the precision to tease these apart and hence fail to position individual patients accurately in this complex disease-state architecture. This highlights a critical gap where further research is needed. To dissect the specific contributions of these metabolic axes, we established a well-characterized cross-sectional cohort from the Indian subpopulation, encompassing normoglycemic, prediabetic, and newly diagnosed T2D individuals. Through integration of clinical phenotyping, immune cell profiling, and whole blood transcriptomics, we demonstrate that insulin resistance (HOMA2-IR) and hyperglycemia (HbA1c) are associated with distinct immuno-molecular profiles, particularly in granulocyte subsets such as neutrophils and eosinophils. These observations were further validated using independent external cohorts with clinical, proteomic, and metabolomic data, confirming that insulin resistance and hyperglycemia exert divergent effects on the immune-metabolic interface. To gain upstream insight into these distinct phenotypes, we employed a network-based systems approach to identify potential genetic driver mutations associated with HOMA2-IR and HbA1c. Using our in-house ‘PreDD’ framework, we integrated phenotype-associated transcriptomic features onto patient-specific molecular protein-protein interaction networks, leveraging curated mutation databases to identify candidate driver events. Comparative network analysis revealed significant enrichment of driver mutations in insulin resistance-associated subnetworks, particularly those linked to neutrophil activation, reactive oxygen species (ROS), and stress response signaling. These perturbations frequently involved immune cell–associated transcription factors implicated in lipid and glucose metabolism, insulin signaling, and immune cell differentiation. Together, these findings highlight a strong immune-metabolic axis in the pathophysiology of insulin resistance. In addition, we independently assessed the gut microbiome through shotgun metagenomic sequencing to explore its compositional and functional landscape across metabolic health states. Recognizing that enterotypes—such as Firmicutes, Bacteroides, and Prevotella-dominated configurations are primarily shaped by long-term dietary patterns rather than disease status per se, we focused on examining how these diet-influenced microbial community structures are perturbed in the context of insulin resistance and glycemic dysregulation in metabolic disorders. By decoupling enterotype classification from disease categorization, we aimed to identify microbiome alterations robust to dietary background and may serve as independent markers or modulators of metabolic health. This approach allowed us to identify key microbiome host interactions, distinguishing from confounding dietary effects, which contribute to a more nuanced understanding of microbiome involvement in Type-2 Diabetes. Although glucose control remains a clinical priority in Diabetes, insulin resistance, dyslipidemia, and BMI can independently drive chronic low-grade inflammation through distinct molecular pathways, even in individuals with controlled glycemia. Together, this multi-omics, multi-cohort study delineates insulin resistance and hyperglycemia as distinct drivers of inflammation in early T2D and metabolic syndrome. It further identifies specific genetic and microbial perturbations that characterize these phenotypes, offering new avenues for precision diagnosis, mechanistic understanding, and therapeutic targeting in complex metabolic disorders.