Integrative Multi-Omics and Mechanistic Modeling Uncover HK2-Driven Warburg Vulnerabilities in Pancreatic Cancer

Integrative Multi-Omics and Mechanistic Modeling Uncover HK2-Driven Warburg Vulnerabilities in Pancreatic Cancer
Dariush Norouzian,^1,* Vahid Rezaei,² Dr. Mohammad Mahdi Eslami,³ Dr. Reza Mirlohi,⁴
1. Pasteur Institute of Iran, Department of Biochemistry / Head of Cell Culture and Development Laboratory
2. National Biotech School
3. Member of the Bioinformatics Research Group, Nasim Research Institute, Tehran, Iran
4. Member of the Bioinformatics Research Group, Nasim Research Institute, Tehran, Iran
Introduction: Pancreatic ductal adenocarcinoma (PDAC) is a highly lethal malignancy characterized by late-stage diagnosis, poor prognosis, and profound metabolic reprogramming toward aerobic glycolysis (Warburg effect). Targeting glycolytic enzymes such as hexokinase II (HK2) offers a plausible therapeutic strategy given their role in tumor energy metabolism. However, current glycolysis-targeted approaches often rely on indirect biomarkers, lack reproducible flux validated metrics, and provide no standardized thresholds for patient stratification. This study addressed these gaps by developing an integrative, simulation-driven, multi-omics pipeline to quantify glycolytic dependence, identify metabolic vulnerabilities, and benchmark therapeutic responses in PDAC.
Methods: We harmonized transcriptomic data from TCGA‑PAAD and GEO with targeted metabolomics, applying ComBat correction, GSVA/ssGSEA pathway scoring, ANCOVA, and Cox survival models. A mechanistic ODE model incorporating ATP dynamics and oxidative phosphorylation simulated Warburg indices, glycolytic ATP fractions, and drug perturbations for 2‑deoxy‑D‑glucose (2‑DG), 3‑bromopyruvate (3‑BP), and lonidamine.
Results: Cancer cells exhibited 6.2‑fold higher lactate production, depleted glucose pools (0.3× normal levels), and reduced ATP/ADP (0.24×) and NADH/NAD⁺ (0.68×) ratios, indicating mitochondrial compromise. HK2 correlated with lactate (r=0.36) and ECAR (r=0.34) but did not fully explain metabolic heterogeneity. Drug simulations identified 3‑BP as the most potent HK2 inhibitor (IC₅₀=2.75 μM) but with high cytotoxic potential; 2‑DG showed modest inhibition with minimal cytotoxicity, and lonidamine induced mixed metabolic disruption. Combination of 3‑BP and 2‑DG yielded antagonistic effects (CI>1.12).
Conclusion: These findings confirm HK2 as a central metabolic control node in PDAC and demonstrate that therapeutic efficacy depends on inhibition mechanism and pathway context. The proposed pipeline offers a reproducible framework for biomarker‑guided patient selection and rational glycolysis‑targeted therapy in precision oncology.
Keywords: Pancreatic Ductal Adenocarcinoma (PDAC)-Warburg Effect-Hexokinase II (HK2)-Multi‑Omics Integration