Quantitative Systems Biology of Interplay Among Three Transcription Factors NRF2, sMAF, and BACH1 Modulating Cellular Antioxidant Response

Background. The antioxidant response element (ARE) is regulated by three competing transcription factors (TF) that bind as dimeric complexes. The NRF2–MAFG heterodimer activates transcription; the MAFG homodimer, which has no transactivation domain, and the BACH1–MAFG heterodimer silence it. All three TFs compete for the ARE, so the transcriptional output is set by a three-way competition rather than by NRF2 abundance alone. All these TFs are promising therapeutic targets in various diseases, with candidate compounds currently under development. However, successful drug design requires a detailed understanding of their complex interplay at ARE activation. Quantitative systems biology (QSB) modelling of these triple competitive systems provides a means to disentangle these dynamics in different diseases. In this paper, we developed a QSB model of NRF2-sMAF-BACH1-ARE system as a submodule of a larger upstream network that control cellular oxidative stress in various pathologies and aging. Methods. We built a quantitative systems biology model of the three-way competition at the ARE: NRF2, small MAF (sMAFG) and BACH1 interconvert with three dimers (the NRF2-MAFG activator, the MAFG homodimer and the BACH1-MAFG repressor) that compete for one pool of ARE sites; heme sequesters free BACH1 (12 species, 7 reactions; SBML L3V2). Binding constants came from published SPR and EMSA data; four remaining quantities were fitted to a nine-point luciferase titration. Genome-wide classification of antioxidant gene promoters by ARE affinity was carried out, and affinity scale was transferred to a position weight matrix that was applied to the promoters of 172 antioxidant and NRF2-target genes against 394 controls. Results. The dependence of activation and repression of the ARE on NRF2 and sMAF levels was calculated and calibrated against experimental data. Output is set by the shared sMAF pool: activation rises monotonically in NRF2 but peaks in small MAF at 0.173 µM. NRF2 acts as the switch variable and small MAF as a two-sided gain setting. BACH1 inhibits competitively, sliding the dose-response along the NRF2 axis without lowering the saturation level: 0.3 µM holds 54% of the sites and halves activation when BACH1–sMAF dimerisation is tight, and is inert at 10 µM. Heme relieves repression near 0.14 µM by returning sMAF to the activator. Using experimental dissociation constants for various SNP in the ARE, the model results demonstrated that sequence identity alters occupancy three-fold without changing response shape, and sorts real promoters: the high-affinity class is enriched 3.05-fold among antioxidant genes (odds ratio 3.39, p = 1.9∙ 10⁻⁴), recovers the curated NRF2 programme, and is predicted to be 6.5% occupied at resting NRF2 against 1.2% for the low-affinity class. We discussed the obtained results in the context of the potential pharmacological role of NRF2-sMAF-BACH1-ARE system as a therapeutic target for current drug design in various diseases and aging. Conclusions. The ARE step is graded, reading out an upstream sensor rather than creating a switch. Its sensitivity lies not in the well-measured DNA-binding layer but in the dimerisation layer, which is fitted or assumed and decides whether repression occurs at all; measuring the NRF2–sMAF and BACH1–sMAF affinities would provide the strongest constraint on the system.