Template of Metabolic Reprogramming in Cancer and Healthy Cells for Inferring Oncogenes

Cancer cells exhibit unusual metabolic activity, characterized by high rates of glucose consumption and lactate production even under aerobic conditions, as well as increased glutamine catabolism and amino acid metabolism. Based on the behaviors of metabolic reprogramming, oncogenesis can be inferred from a genome-scale metabolic model of cancer cells. In this study, we incorporated the CORDA method [1] with data of the Human Protein Atlas database [2] and Recon 2.2 network [3] to reconstruct tissue-specific metabolic models for normal and cancer cells, respectively. These models were applied to assess the flux alternation template using flux-sum balance analysis. We developed a bilevel optimization formulation to infer oncogenes for the tissue-specific models. The objective function in the outer optimization problem is the similarity ratio of flux alternations of the dysregulated cell to the template for cancer or health models, or Warburg hypothesis from the literature. The inner optimization problem consisted of the flux balance model and uniform flux distribution model [4]. A nested hybrid differential evolution algorithm was introduced to solve the bilevel optimization problem. The head and neck tissue-specific model was used for a case study to infer oncogenes. Results show that TPK1 achieved the highest similarity ratio of 0.84 and three genes (GNPD1/GNPD2, THTPA, and PTEN) exhibited similarity ratio of 0.839. Literature shows that TPK1 is an oncogene for colon cancer [5] and PTEN is a tumor suppressor gene of squamous cell carcinomas of the head and neck [6].

1. Schulta et al., PLOS Computational Biology, 12(3): e1004808, 2016
2. https://www.proteinatlas.org/
3. Swainston1 et al., Metabolomics, 12:109, 2016
4. Wu et al., PLOS Computational Biology, 13(7): e1005618, 2017
5. Tiwana1 et al., Oncotarget, 6(8) : 5978–89, 2015
6. M. Poetsch et al., Cancer Genetics and Cytogenetics, 132(1):20-4, 2002