enetically encoded tools also exist to
directly elevate102 or decrease80,103 the NADH/NAD+ ratio (and
NADPH) in cells and organisms, acting on the system’s f2 to
shape �eR and GDF15 levels.73e ́ R markersThe best available �eR marker in humans is the secreted cytokine
GDF15, a member of the transforming growth factor β (TGF-β)
family.104 GDF15 increases when mitochondrial �eR is elevated
in cells,43,73 mice,105–107 and humans68,81,108 with mtDNA defects that increase the biophysical constraints for electrons to
reach oxygen. Mitochondria-deficient cells and animals experience high �eR and an elevated NADH:NAD+ ratio reflecting reductive stress.68 This activates the nuclear integrated stress
response (ISR)43,72,73,109 and stimulates the production and
release of cytokines signaling excessive energy resistance,
particularly GDF15. The cytokine fibroblast growth factor 21
(FGF21), but interestingly not other traditional cytokines (ILsGlucose, FFAs, HbA1c [Nutrient
oversupply, GLT]Inflammation [Hypermetabolism]Cortisol, NE [Stress hormones] ?mtDNA mutations, mitochondrial
OxPhos defects [Genetic][Poisons and toxins]
[Sleep deprivation]NADH/NAD+ ratio [Direct]Lactate:Pyruvate, Alanine [Indirect]GDF15*, other cytokines? [Cellderived, non-immune cytokines]Oxidative damage [Oxidative stress]Telomeres, DNAm [Secondary]Local temperature [Heat production]Lack of energy, fatigue, anxiety,
psychopathology symptoms[Subjective]Suppressed anabolism,
testosterone [Metabolic shift]Catabolism & atrophy: HPA axis
activation [Metabolic shift]Loss of function in low priority
systems, reproduction [Physiological]Figure 7. Testable predictions and operationalization of key terms in the ERP modelMeasurable factors that contribute to elevate energy potential (EP) or that decrease flux (f2) are �eR drivers. Factors that become elevated directly or indirectly with
high �eR are �eR markers. Signs and symptoms that result from acute or chronically elevated �eR are �eR phenotypes. FFAs, free fatty acids; GDF15, growth differentiation factor 15; HbA1c, glycated hemoglobin; HPA, hypothalamic-pituitary-adrenal; GLT, glucolipotoxicity; NE, norepinephrine; OxPhos, oxidative
phosphorylation.ll2118 Cell Metabolism 37, November 4, 2025Perspectiveand interferons), has also been reported to be elevated in animals and people with mtDNA defects.68,107,110 In ERP terms,
mtDNA-mutant organisms have reduced capacity for flux, f2.
Thus, for a given EP—which is either the same, or may in fact
be elevated in mitochondrial disorders42—reduced f2 should
exponentially elevate �eR (following �eR = EP/f2). This state is
detectable as elevated circulating GDF15,107,110 an e ́ R marker
validated across cellular, preclinical, and clinical studies.e ́ R phenotypesThe best mapped �eR phenotypes are produced by the brain.111GDF15 is expressed in all organs and tissues in the body except
the brain.108 In turn, the receptor for GDF15, GFRAL, is expressed only in a subset of neurons in the brainstem region.104This expression pattern is a robust example of a canonical
body-to-brain metaboception signaling axis conveying somatic
energetic distress to the brain. In parallel with other metaboceptive cues, the brain appears to respond to the �eR signal GDF15
by triggering phenotypes that decrease the system’s EP.111This includes visceral malaise,112 nausea and fasting,113 social
isolation,114 and depression.115 Thus, physiological and behavioral phenotypes of excessive �eR offer opportunities to examine
how subcellular �eR relates to human health.
To facilitate empirical testing of the ERP in various contexts
and systems, we provide an ERP measurement toolkit. Summarized in Figure 8, this toolkit provides established and tentative
approaches to selectively increase or decrease EP and f2 in mitochondria, cells or tissues, and whole organisms. We also highlight an initial set of strategies to monitor �eR at each level.ERP AND HUMAN HEALTHSeveral attempts have been made to define health both biologically and functionally, recognizing that health is more than the
absence of disease.116,117 Aiming to define health from first principles, we recently proposed that health is a field-like state
emerging from the dynamic interplay of energy, communication,
and structure within the organism, giving rise to robustness/resilience, plasticity, performance, and sustainability.118 Extending
this definition to the ERP, we propose that health is the ability
to sustain life at an optimal level of �eR. This then predicts that deviations in �eR, quantified from either �eR drivers or �eR markers,
should predict and/or track with changes in health status. On
the other hand, healing may be defined as the dynamic process
of achieving functional or structural adaptations that optimize �eRdynamics. These proposed operationalizations provide a basis
to develop heretofore missing quantitative metrics and to eventually track human health and healing processes over time.
This is supported by UK Biobank results in >52,000 individuals
where the top circulating protein associated with prevalent and
incident diabetes, cardiovascular, psychiatric, and other disorders
is the �eR marker GDF15.119 Data S1 summarizes UK Biobank results119 validating initial ERP predictions: (1) metabolic oversupply,
immune profiles, and tissue damage markers expected to increaseEP are linked to elevated �eR; (2) �eR is associated with metabolite
profiles of reductive stress including lactate and alanine; and (3)
elevated GDF15 is linked to �eR phenotypes—experiences, behaviors, and physiology—including fatigue, pain, sedentarism, neuroticism, and suppression of growth/anabolic axes.
Considering non-pathological examples where �eR is partitioned across the mammalian body, we suggest that pregnancy
is an example where the ERP integrates heretofore disconnected processes. For reasons that remain unclear, the placenta
expresses exceptionally high GDF15 levels (Figure 9). Based on
evidence reviewed above, this may reflect how placental biology
operates at high �eR. This contention accounts for (1) why
maternal blood GDF15 reaches levels >100-fold higher than in
the non-pregnant state, driven by systemic GDF15 placental
secretion113,120; (2) why the high �eR of the placenta may be
required to counterbalance or ‘‘insulate’’ the proliferative and
low �eR proliferative state of the fetus (as in an electrical circuit,
low-energy-resistance components operate as energy sinks
and must be counterbalanced by resistive elements), balancing
each other to protect the mother121,122; and (3) why the placenta
ages faster than other human tissues,123,124 a predicted necessary consequence of the dissipative losses and pro-aging effects in tissues chronically operating at high-�eR values.