Energy State as a Shared Lever: A Coupled-Control Model of Recombinant AAV Production in the Producer Cell

Recombinant AAV (rAAV) production in HEK293 cells has advanced through separate optimizations of individual intervention points: raising ATP levels, inhibiting the proteasome, suppressing apoptosis, or increasing Rep52/40 stoichiometry. However, when these results are considered together, each intervention produces counterintuitive trade-offs. This paper argues that these trade-offs are not anomalies but the expected behavior of a coupled system in which intracellular ATP serves as a shared control variable for three manufacturing outputs: particle yield, capsid integrity, and producer-cell longevity. I describe four mechanistic couplings that link ATP to these outputs. First, ATP directly fuels the Rep52/40 packaging motor, making the encapsidation rate sensitive to the cytosolic energy supply. Second, physiological millimolar ATP suppresses 26S proteasome activity in a biphasic, concentration-dependent manner, sparing nascent capsid precursors and Rep proteins (productive arm) while simultaneously stabilizing p53 and sensitizing cells to apoptosis (destructive arm). Third, proteasome suppression also spares the MRN replication-restriction complex, predicting a non-monotonic, window-shaped yield response to increasing ATP. Fourth, the apoptosis node receives at least three mechanistically distinct death inputs, none of which is fully addressed by single-branch interventions. From this coupling structure, the model makes five falsifiable commitments, each specifying experimental results that would challenge it. The framework re-derives recent trade-off findings in the field, including contradictory results from pan-caspase and HIF1α inhibition, as predictable outcomes of a constrained system rather than isolated anomalies. It further predicts that coordinated improvement in yield requires joint control of both the ATP set-point and the apoptosis node.