We propose that in chronically stimulated T cells, the balance between passive and active DNA demethylation reverses: early differentiation relies on replication-dependent dilution of oxidized methylcytosines, whereas persistent TCR signaling induces a metabolic state that favors TDG‑mediated base excision repair at specific enhancers governing inhibitory receptors and cytokine genes. This shift would convert the relatively stable 5hmC marks seen in naïve and memory subsets into transient 5fC/5caC intermediates that are excised, permitting rapid demethylation of loci such as Pdcd1, Lag3, and Havcr2, thereby stabilizing an exhausted transcriptional program. Conversely, loss of TDG would trap 5fC/5caC, leading to aberrant 5hmC accumulation and a failure to fully silence effector genes, which could be rescued by metabolic manipulation that restores α‑KG/succinate ratios and enhances TET oxidation activity.

Testable predictions

1. In a mouse model of chronic LCMV infection, sorted exhausted CD8⁺ T cells will show elevated 5fC and 5caC (measured by oxidative bisulfite sequencing) at Pdcd1 and Lag3 enhancers compared with acute‑phase effector cells, while total 5hmC levels remain unchanged or slightly increased.
2. Conditional deletion of Tdg specifically in T cells (using Cd4‑Cre) during chronic infection will result in:
   • Accumulation of 5fC/5caC at exhaustion‑associated enhancers (oxidative bisulfite seq).
   • Reduced excision of these intermediates, leading to higher 5hmC signal at the same loci.
   • Impaired downregulation of Pdcd1, Lag3, and Havcr2 transcripts (qPCR, flow cytometry).
   • Enhanced proliferative capacity and cytokine production upon ex vivo restimulation, indicating a partial reversal of exhaustion.
3. Metabolic supplementation with cell‑permeable α‑ketoglutarate (α‑KG) or inhibition of succinate dehydrogenase (to raise α‑KG/succinate ratio) in exhausted T cells cultured in vitro will increase TET enzymatic activity (measured by 5hmC production) and synergize with TDG to accelerate demethylation of Pdcd1 enhancer, decreasing PD‑1 surface expression.
4. Conversely, pharmacological inhibition of TDG (e.g., with a small‑molecule inhibitor) in acute‑phase effector T cells will blunt the normal decline of 5hmC at effector gene enhancers (Ifng, Tbx21) and preserve a more effector‑like transcriptome, without affecting proliferation.

Mechanistic rationale Chronic TCR signaling drives a shift toward oxidative phosphorylation and increased mitochondrial ROS, which alters the TCA cycle and elevates succinate—a known competitive inhibitor of TET dioxygenases. However, sustained ROS also activates ATM/ATR pathways that can phosphorylate TDG, enhancing its affinity for 5fC/5caC. We hypothesize that this dual modification creates a feed‑forward loop: moderate TET oxidation generates 5fC/5caC, which, under high‑succinate conditions, are not efficiently further oxidized but become substrates for TDG‑mediated excision. The resulting abasic sites are processed by BER, leading to demethylation. In naïve or memory T cells, low proliferation limits ROS and succinate, favoring passive dilution; in exhaustion, high ROS/succinate flips the balance toward active repair.

Falsifiability If exhausted T cells do not show increased 5fC/5caC at inhibitory receptor enhancers, or if Tdg loss fails to alter 5hmC levels or exhaustion markers, the hypothesis would be refuted. Likewise, if α‑KG supplementation does not enhance demethylation or PD‑1 downregulation despite elevated TET activity, the proposed metabolic coupling would be unsupported.