Mitochondrial Stress Commitment as an ARCH-Governed Biological Decision:Cardiolipin as the Ancestral Φ Substrate

1. Introduction

Consider two familiar biological decisions. At a synapse, a neuron encodes a long-term memory trace only when the postsynaptic NMDA receptor is simultaneously unblocked by membrane depolarization and activated by glutamate. Magnesium physically blocks the channel at rest, so maximal glutamate release executes nothing without prior depolarization: the gate must open before the drive can act. In the Venus flytrap, closure requires two mechanosensory triggers within approximately twenty seconds; one touch is insufficient regardless of force. Both systems share a common logic: output depends on the product of jointly necessary conditions, and any single term approaching zero collapses the product regardless of the others.

The ARCH × Φ framework formalizes this grammar as R = Φ(A × D × C) ≥ θ, where A is structural Archetype, D is energetic Drive, C is Context (triggering signal), and Φ is gating permissiveness—a variable that enforces a zero-term veto when absent or dysregulated, must function within a required operating range, and whose recovery timescale (τΦ) scales with the commitment cost of the transition it gates (Rahman et al., 2025). The framework has been applied to behavioral execution (Rahman et al., 2025), DNA replication initiation (Rahman, 2025), sterol gated biological execution (Rahman, 2026a), Venus flytrap closure (Rahman, 2026b), and clownfish sex change (Rahman, 2026c). Across these systems, the same architecture governs whether a costly biological transition executes: four jointly necessary domains, a multiplicative product structure, and a categorical veto imposed by any domain approaching zero.

Mitochondrial stress commitment is not merely one more system to which this framework can be applied. It is the system in which the evolutionary claim is most direct. Cardiolipin, the signature phospholipid of the inner mitochondrial membrane, performs the same function that cholesterol later evolved to perform at the plasma membrane: it separates stored electrochemical drive from discharge permissiveness, making it physically possible for a membrane to be energetically loaded and still not execute. Cardiolipin is present in the plasma membrane of ancestral alpha-proteobacteria and was inherited by mitochondria at endosymbiosis (Hoch, 1992; Schlame & Greenberg, 2017). The Φ architecture therefore did not begin with nervous systems, eukaryotes, or the oxygenated world. It began with a bacterium managing the decision of whether to discharge a stored electrochemical gradient.

This article maps four mitochondrial commitment systems onto the ARCH × Φ domains, establishes the zero-term veto through eight experimental demonstrations, identifies three independently approved therapeutics that converge on Φ-titration as a shared mechanism, shows that the framework retrodicts a recent clinical trial failure that surprised the field, specifies the statistical interaction test that discriminates multiplicative from additive integration, and derives five falsifiable predictions.

2. The ARCH × Φ Framework

2.1. The Readiness Function and Its Properties

The multiplicative readiness function R = Φ (A × D × C) ≥ θ encodes three system-level properties absent from additive models. The first is the zero-term veto: as any domain X approaches zero, R approaches zero regardless of the magnitude of the remaining domains. The second is supra-additive suppression: combined partial reductions suppress R more than the sum of individual reductions, because (1 − δ₁)(1 − δ₂) < (1 − δ₁) + (1 − δ₂) − 1 for δ > 0. The third is threshold nonlinearity: near θ, small changes in any domain produce disproportionate changes in execution probability. Because ln (A × D × C × Φ) = ln A + ln D + ln C + ln Φ, the model is log-linear and testable against additive specifications by standard regression.

2.2. Three Operational Criteria for Φ

Φ is distinguished from limiting factors and amplifiers by three jointly necessary criteria. The first is dynamic separability: Φ recovery is governed by processes causally upstream of and experimentally separable from Drive—the BCL-2/BAX ratio recovers on its own kinetic trajectory when Drive is held constant, and cardiolipin remodeling cannot be restored by exogenous substrate supply in the absence of functional tafazzin. The second is molecular bistability: Φ occupies one of two stable states rather than varying continuously, demonstrated for p16/Rb in the all-or-none commitment to senescence (Narita et al., 2003), for PINK1 stabilization as a function of ΔΨm (Matsuda et al., 2010), and formally for the mitochondrial permeability transition pore, for which bifurcation analysis identifies a bistable domain in calcium concentration space. Cardiolipin itself exhibits bistability at the level of lipid physics, transitioning between lamellar bilayer and non-bilayer inverted hexagonal phases (Birk et al., 2025). The third is non-compensability: the zero-term veto. A conventional scaling parameter can satisfy the third criterion near zero by accident but will not satisfy the first two. The joint three-criterion test distinguishes a Φ-class gating variable from a limiting factor.

3. Domain Mapping Across Four Mitochondrial Commitment Systems

3.1. The Cardiolipin Inner Membrane Gate

Cardiolipin is essential for the structural integrity and proton impermeability that make the mitochondrial membrane potential possible. In its absence, protons leak across the inner membrane, ΔΨm collapses, and ATP synthesis fails regardless of substrate availability or electron transport chain integrity. Jiang et al. (2000) demonstrated this directly: yeast lacking cardiolipin synthase exhibit decreased membrane potential and reduced mitochondrial function despite an intact protein complement. This is a zero-term veto imposed by the lipid layer.

Cardiolipin is synthesized with saturated acyl chains and remodeled post-synthetically by tafazzin, a transacylase that introduces unsaturated fatty acids to produce the mature species supporting electron transport chain supercomplex assembly (Schlame & Greenberg, 2017). This remodeling is activity-dependent and history-dependent: the acyl composition reflects prior metabolic demand, with stressed or aging mitochondria accumulating immature or oxidized forms. The ratio of intact to oxidized cardiolipin is therefore a Φ-state variable written by prior oxidative stress history, gating whether a given ΔΨm perturbation executes cytochrome c release.

Kagan et al. (2005) established a remarkable feedforward architecture in which cytochrome c itself acts as a cardiolipin peroxidase under stress conditions—the apoptotic executor writes the Φ state that gates its own release. Cardiolipin also gates a voltage-dependent channel directly: Rostovtseva et al. (2006) demonstrated that cardiolipin induces significant asymmetry in the voltage gating characteristics of the voltage-dependent anion channel (VDAC), and more recent work confirms that cardiolipin uniquely modulates VDAC isoform 3 gating by preferentially retaining the channel in an open-like conductive state (Rawat et al., 2026). This is the structural homolog of the synaptic argument: a lipid modulating a voltage-dependent channel, separable from the drive it transmits.

Barth syndrome provides the monogenic proof of concept. Caused by tafazzin mutation, it produces accumulation of immature cardiolipin that cannot support normal supercomplex assembly, yielding cardiomyopathy, skeletal muscle weakness, and neutropenia (Schlame & Ren, 2006). The phenotype maps onto a Φ-lock disorder at the inner membrane: structural substrate intact, metabolic drive present, cellular context permissive, but the lipid gate locked in a configuration that vetoes oxidative phosphorylation regardless of substrate supply.

3.2. PINK1/Parkin Mitophagy

Mitophagy, the selective autophagy of damaged mitochondria, is the first commitment level in the cellular stress hierarchy. In healthy mitochondria, PINK1 is continuously imported, translocated inward, and cleaved by the inner membrane protease PARL, maintaining steady-state outer membrane concentration near zero. When ΔΨm collapses below a threshold, import fails, PINK1 accumulates on the outer membrane, autophosphorylates, recruits cytosolic Parkin, and initiates ubiquitin-dependent mitophagy (Narendra et al., 2008). Matsuda et al. (2010) established the binary character of this switch: PINK1 stabilization occurs discretely as a function of ΔΨm rather than proportionally.

Archetype comprises mitochondrial network integrity, the import machinery, and outer membrane receptor complement. Drive comprises ΔΨm collapse magnitude and duration, reactive oxygen species accumulation, and cytosolic calcium overload. Context comprises autophagic flux capacity, lysosomal fusion competence, and LC3-II availability. Φ is the PINK1 stabilization state—a binary switch set by the ΔΨm import threshold and independently tunable from Drive by the rate of PARL-mediated degradation, with τΦ of minutes to hours.

Cardiolipin is not merely upstream of this gate but constitutes part of its recognition machinery. Chu et al. (2013) demonstrated that pro-mitophagy stimuli cause externalization of cardiolipin to the mitochondrial surface and that knockdown of cardiolipin synthase or phospholipid scramblase-3 decreases delivery of mitochondria to autophagosomes. LC3, which mediates autophagosome formation and cargo recognition, contains cardiolipin-binding sites required for engulfment. The ancestral lipid gate writes the mitophagy elimination signal directly.

The zero-term veto operates bidirectionally: when ΔΨm is intact, PINK1 is degraded regardless of cytosolic Drive (Φ too high; execution blocked). When PINK1 is absent, mitophagy is abolished despite ΔΨm collapse and maximal oxidative Drive (Archetype-domain veto; Narendra et al., 2008).

3.3. The BCL-2/BAX Apoptotic Gate

BCL-2 and BCL-xL hold BAX and BAK in inactive states, preventing mitochondrial outer membrane permeabilization (MOMP) and cytochrome c release. Their ratio to the pro-apoptotic effectors sets the commitment threshold, and prior stress writes this ratio: cells experiencing chronic low-level stress accumulate lower BCL-2 and higher BAX, shifting the gate toward commitment (Chipuk et al., 2010; Llambi et al., 2011). This is a sliding modification threshold formally analogous to the Bienenstock–Cooper–Munro (BCM) threshold at the synapse: prior history raises or lowers the threshold for execution independently of the current signal.

The evidence that ratio rather than absolute Drive level determines commitment is direct: the number of mitochondria required to undergo permeability transition before apoptosis is triggered is consistent with the BAX/BCL-2 ratio across cell types, not with total pro-apoptotic protein expression. BCL-2 overexpression provides the categorical falsification of additive models. Every additive account predicts that maximal BH3-only Drive should execute apoptosis or at minimum increase its probability monotonically. In BCL-2-overexpressing cells, maximal Drive fails categorically rather than partially. The logical structure is identical to the Coan et al. (1989) synaptic result, in which removing the Mg²⁺ block abolished rather than facilitated LTP: the additive prediction is unconditional, directional, and wrong (Rahman, 2026d).

3.4. The p16INK4a/Rb Senescence Lock

Archetype comprises genome structural integrity, telomere length, and nuclear lamina organization. Drive comprises CDK4/6 activity, mitogenic signaling, and oncogene-induced replication stress. Context comprises DNA damage signals, the senescence-associated secretory phenotype (SASP) cascade, and the oxidative environment. Φ is the p16INK4a/Rb chromatin lock, in which progressive p16 accumulation inhibits CDK4/6, maintaining Rb in its hypophosphorylated repressive form and enabling formation of senescence-associated heterochromatin foci (SAHF) at E2F target promoters (Narita et al., 2003; Serrano et al., 1997).

The irreversibility of this gate is mechanistically distinct from the mitophagy and apoptotic gates. Initial DNA damage response produces a p21-mediated arrest reversible upon damage repair. Progressive p16 accumulation then inhibits CDK4/6 independently of p53, and SAHF formation renders the chromatin state self-maintaining through subsequent divisions. A fully senescent human cell does not re-enter the cell cycle regardless of growth factor concentration or exogenous CDK activator provision (Beaúśejour et al., 2003; Narita et al., 2003). τΦ approaches the lifetime of the organism because the cost of false-positive reversal is malignancy.

Table 1. ARCH × Φ Domain Mapping Across Four Mitochondrial Commitment Systems. Φ columns are shaded. τΦ rank order matches commitment cost rank order across all four systems. ETC = electron transport chain; ROS = reactive oxygen species; SASP = senescence-associated secretory phenotype; SAHF = senescence-associated heterochromatin foci.

System	Archetype (A)	Drive (D)	Context (C)	Φ — Gate	τΦ
Cardiolipin gate	ETC complement; cristae; mtDNA integrity	Substrate supply; NADH/FADH₂; proton pumping	Matrix pH; ADP; oxygen tension	Cardiolipin acyl composition; remodeled by tafazzin; gates proton impermeability and supercomplex assembly; modulates VDAC gating	Hrs–days
PINK1/Parkin mitophagy	Network integrity; import machinery; outer membrane receptors	ΔΨm collapse; ROS; Ca²⁺ overload	Autophagic flux; lysosomal fusion; LC3-II availability	PINK1 stabilization state — binary switch at ΔΨm threshold; cardiolipin externalization provides LC3 recognition signal	Min–hrs
BCL-2/BAX apoptosis	Outer membrane; cristae; cytochrome c tethering	BH3-only (BIM, BID, PUMA); ΔΨm collapse; cytosolic Ca²⁺	APAF-1; caspase-9 zymogen pool; cytosolic protease machinery	BCL-2/BCL-xL to BAX/BAK ratio — history-dependent gate; independently tunable from acute Drive	Hrs–days
p16/Rb senescence	Genome integrity; telomere length; nuclear lamina	CDK4/6; mitogenic signaling; oncogene-induced stress	DNA damage signals; SASP cascade; oxidative environment	p16INK4a/Rb chromatin lock; SAHF formation is self-maintaining across divisions	Lifetime

Table 1. ARCH × Φ Domain Mapping Across Four Mitochondrial Commitment Systems. Φ columns are shaded. τΦ rank order matches commitment cost rank order across all four systems. ETC = electron transport chain; ROS = reactive oxygen species; SASP = senescence-associated secretory phenotype; SAHF = senescence-associated heterochromatin foci.

System	Archetype (A)	Drive (D)	Context (C)	Φ — Gate	τΦ
Cardiolipin gate	ETC complement; cristae; mtDNA integrity	Substrate supply; NADH/FADH₂; proton pumping	Matrix pH; ADP; oxygen tension	Cardiolipin acyl composition; remodeled by tafazzin; gates proton impermeability and supercomplex assembly; modulates VDAC gating	Hrs–days
PINK1/Parkin mitophagy	Network integrity; import machinery; outer membrane receptors	ΔΨm collapse; ROS; Ca²⁺ overload	Autophagic flux; lysosomal fusion; LC3-II availability	PINK1 stabilization state — binary switch at ΔΨm threshold; cardiolipin externalization provides LC3 recognition signal	Min–hrs
BCL-2/BAX apoptosis	Outer membrane; cristae; cytochrome c tethering	BH3-only (BIM, BID, PUMA); ΔΨm collapse; cytosolic Ca²⁺	APAF-1; caspase-9 zymogen pool; cytosolic protease machinery	BCL-2/BCL-xL to BAX/BAK ratio — history-dependent gate; independently tunable from acute Drive	Hrs–days
p16/Rb senescence	Genome integrity; telomere length; nuclear lamina	CDK4/6; mitogenic signaling; oncogene-induced stress	DNA damage signals; SASP cascade; oxidative environment	p16INK4a/Rb chromatin lock; SAHF formation is self-maintaining across divisions	Lifetime

4. Zero-Term Veto: Eight Experimental Demonstrations

The mitochondrial system provides eight categorical demonstrations across four gating levels—a denser evidentiary base than any prior application of the framework. Four demonstrations establish veto from above, in which the gate is locked high and maximal Drive fails: cardiolipin synthase deletion collapsing ΔΨm despite intact electron transport chain complement (Jiang et al., 2000); tafazzin mutation in Barth syndrome producing multi-system failure despite intact structural and metabolic domains (Schlame & Ren, 2006); maintained ΔΨm preventing PINK1 accumulation regardless of oxidative Drive (Matsuda et al., 2010); and BCL-2 overexpression abolishing MOMP regardless of BH3-only Drive (Chipuk et al., 2010).

Two establish Archetype-domain veto: PINK1 knockout abolishing mitophagy (Narendra et al., 2008) and the p16/Rb senescence lock preventing cell cycle re-entry (Narita et al., 2003). Two establish Φ-titration rescue, in which the gate is restored or lowered and sub-threshold Drive executes: venetoclax executing apoptosis at Drive levels otherwise insufficient (Del Gaizo Moore et al., 2007) and elamipretide restoring cardiolipin function in Barth syndrome (Reid Thompson et al., 2021). Table 2 tabulates all eight.

5. Three Approved Therapeutics Converge on Φ-Titration

Three drugs, approved across three decades for three unrelated indications by three separate development programs, share a single mechanistic description that no additive account provides: each restores a gating variable to its functional operating range without increasing Drive, restoring structural substrate, or modifying Context.

Memantine (approved 2003, Alzheimer’s disease) is a voltage-dependent open-channel blocker of the NMDA receptor. Frankiewicz and Parsons (1999) demonstrated that at therapeutically relevant concentrations it restores LTP in magnesium-free medium without impairing potentiation under normal magnesium conditions. The drug does not add glutamate, modify synaptic structure, or increase stimulation frequency. It reinstates the coincidence-detection gating that magnesium removal had destroyed.

Venetoclax (approved 2016, chronic lymphocytic leukemia and acute myeloid leukemia) is a BH3 mimetic that competitively displaces BIM from BCL-2, lowering the apoptotic commitment threshold. Del Gaizo Moore et al. (2007) established the mechanistic basis: CLL cells require BCL-2 to sequester pro-death BIM, and displacing it executes apoptosis at Drive levels that would otherwise be insufficient. Venetoclax is Φ-titration from below, the mitochondrial counterpart of restoring the synaptic gate from above with memantine.

Elamipretide (granted accelerated approval September 2025, Barth syndrome) is a cell-permeable tetrapeptide that binds reversibly to cardiolipin on the inner mitochondrial membrane, stabilizing respiratory chain supercomplexes and improving electron transport efficiency (Reid Thompson et al., 2021). In the TAZPOWER trial and its open-label extension, treatment produced sustained improvements in muscle strength and cardiac function, with cardiolipin and monolysocardiolipin values improving in parallel with clinical outcomes. Elamipretide is Φ-titration at the ancestral lipid gate.

Three gates, three scales, three approval decisions, one mechanism. The framework predicts this convergence is structural: any biological system implementing threshold-gated commitment through a tunable gating variable will be therapeutically addressable by titrating that variable back into range, and drugs that do so will succeed where drugs targeting Drive or structure fail. Table 3 summarizes.

6. A Retrodiction: The 2025 Integrated Stress Response Trial Failures

The integrated stress response (ISR) operates through four kinases—HRI, PKR, GCN2, and PERK—each responding to a distinct upstream stress signal and converging on phosphorylation of eukaryotic initiation factor 2 alpha (eIF2α), which globally suppresses cap-dependent translation while selectively upregulating ATF4 (Pakos-Zebrucka et al., 2016). The eIF2α phosphorylation state satisfies all three Φ-class criteria: it is written by prior stress history through accumulated kinase activity, it gates whether translation can execute at a given moment independently of ribosome or transcript availability, and—critically—it requires an operating range, not merely a level.

Insufficient phosphorylation impairs stress adaptation, while excessive phosphorylation produces global translation arrest and death receptor 5–mediated apoptosis. Guo et al. (2020) and Fessler et al. (2020) identified the relay connecting mitochondrial stress to this gate: mitochondrial dysfunction activates OMA1, which cleaves DELE1; the cleaved fragment accumulates in the cytosol, binds HRI, and activates eIF2α kinase activity. This is a trans-compartment Φ relay—stress at one cellular locus writes a gating state at a physically separate locus through dedicated machinery, structurally analogous to the astrocytic gap-junction pathway that writes heterosynaptic metaplasticity at synapses receiving no Drive (Hulme et al., 2014).

In January 2025, two eIF2B-activating compounds designed to suppress the ISR—DNL343 and fosigotifator—failed to meet primary endpoints in the HEALEY platform trial for amyotrophic lateral sclerosis (ALS), along with key secondary endpoints of muscle strength and respiratory function, despite good target engagement and strong preclinical rationale.

The ARCH × Φ framework offers a specific account, offered as a retrodiction rather than a post hoc rationalization. If eIF2α phosphorylation is a Φ-class variable with a required operating range, then blanket pharmacological suppression does not restore the gate to range—it drives the gate past range in the opposite direction, which the framework predicts is equally damaging. This is precisely the logic of the Coan et al. (1989) synaptic result: removing the constraint does not maximize the gate but destroys it. Consistent with a window rather than a level, an exploratory high dose of fosigotifator was associated with significant slowing of muscle strength decline—a non-monotonic dose relationship that a simple suppression model does not predict. The framework’s therapeutic prescription differs: ISR modulation should aim at titration into range, with dose individualized to baseline eIF2α phosphorylation state, rather than at maximal suppression.

7. Statistical Test: The Multiplicative Interaction Term

Following methodology applied to photosynthetic measurements (Rahman, 2026c), the framework predicts a significant negative interaction term in the log-linear specification:

ln(R) = β₀ + β₁ ln(D) + β₂ ln(Φ) + β₃ ln(D) ln(Φ) + ε

where R is apoptotic output (percentage mitochondrial depolarization), D is BH3-only peptide concentration, and Φ is the anti-apoptotic gate state manipulated by BH3 mimetic concentration. The additive null predicts β₃ = 0—BCL-2 inhibition suppresses apoptotic threshold equally across all Drive levels. The framework predicts β₃ < 0—as Drive decreases, the penalty from elevated gate state grows disproportionately larger.

Published BH3 profiling data provide the required factorial structure. In a recent analysis of therapy-induced senescence, mitochondrial depolarization was measured across increasing concentrations of activator peptides (BIM, BID, PUMA) and sensitizer peptides (BMF, BAD, NOXA, HRK) in proliferative and treated cells, in parallel with serial dilutions of navitoclax, A1331852, and venetoclax. Peptide concentration normalizes to Drive on the unit interval; BH3 mimetic concentration provides an inverse index of gate state. A negative and significant β₃, with the mimetic effect monotonically largest at the lowest peptide concentration, would confirm multiplicative integration. A null or positive interaction term would falsify the multiplicative specification for this system.

8. The τΦ Hierarchy

The four mitochondrial systems populate the gating recovery timescale hierarchy with mechanistically characterized examples. PINK1 stabilization recovers in minutes to hours, set by ΔΨm restoration kinetics following transient stress, with moderate commitment cost because eliminating a single organelle is metabolically recoverable. Cardiolipin acyl remodeling operates over hours to days, though the structural consequence of tafazzin loss is constitutional. The BCL-2/BAX ratio recovers over hours to days, written by chronic stress through transcriptional change, with high commitment cost because apoptosis is irreversible. The p16/Rb chromatin lock has a recovery timescale approaching the organism’s lifetime, with catastrophic commitment cost because false-positive reversal produces malignancy.

Incorporating these systems, the cross-system rank order runs from milliseconds for magnesium block cycling at the synapse, through minutes to hours for PINK1 stabilization, hours to days for cardiolipin remodeling and the BCL-2/BAX ratio, weeks to months for clownfish sex change, a growing season for vernalization in annual plants, to a lifetime for the senescence lock. This spans approximately eleven orders of magnitude and matches the commitment cost rank order at every step. The concordance constitutes the primary cross-system falsifiable prediction of the Φ-reversibility principle.

9. Testable Predictions

The framework generates five independently falsifiable predictions. Each specifies the direction of the effect, the experimental design needed to test it, and the result that would falsify it.

1. Supra-additive apoptotic suppression. Combined partial elevation of BCL-2 and sub-saturating BH3-only peptide will suppress apoptotic execution more than the sum of individual effects, with a gap of approximately 24 percentage points at 40% perturbation of each domain relative to the additive null. This is testable in the BH3 profiling system with β₃ as the criterion. A null or positive β₃ falsifies multiplicative integration for this system.

2. τΦ scaling with perturbation magnitude. Gating recovery timescale will scale with the magnitude of BCL-2/BAX perturbation. Larger ratio shifts produced by stronger chronic stress will require correspondingly longer recovery before apoptotic threshold returns to baseline, testable by varying stress magnitude and profiling at sequential recovery timepoints after stress withdrawal. A flat relationship between perturbation magnitude and recovery time falsifies the Φ-reversibility principle.

3. Supra-additive oxysterol–BH3 interaction across scales. Sub-threshold 25-hydroxycholesterol combined with sub-threshold BH3-only Drive will produce greater apoptotic execution than the sum of individual effects. The same oxysterol that disrupts synaptic gating through NLRP3 inflammasome activation (Izumi et al., 2025) also promotes mitochondrial cholesterol accumulation and sensitizes the permeability transition pore, predicting supra-additive interaction across scales. An additive response falsifies the cross-scale coupling claim.

4. Barth syndrome neurons show impaired LTP. Neurons derived from Barth syndrome patients will show impaired long-term potentiation at stimulation intensities sufficient in control neurons—a constitutional Archetype-domain vulnerability arising because cardiolipin deficiency impairs mitochondrial ATP production required for sustained CaMKII autophosphorylation during potentiation maintenance. This is testable in iPSC-derived neurons. Normal LTP in Barth syndrome neurons would falsify the ATP-supply pathway of the cascade claim.

5. Inverted-U dose–response for ISR modulation. Integrated stress response modulation will show an inverted-U dose–response relationship with baseline eIF2α phosphorylation. Rescue will be maximal at intermediate phosphorylation states and negligible or harmful at either extreme. This prediction is consistent with the 2025 ALS trial outcomes and is directly testable by stratifying response to eIF2B activators by baseline ATF4 or CHAC1 biomarker level. A monotonic dose–response falsifies the operating-range claim.

10. Discussion

The mitochondrial system closes an argument that prior applications of the framework left open. Each earlier system demonstrated Φ-gating through a lipid substrate of relatively recent evolutionary origin: phytosterols in plants, neurosteroids and cholesterol-derived oxysterols in vertebrates. The mitochondrial system demonstrates it through a substrate that predates eukaryotes entirely. Cardiolipin did not acquire gating function when sterols became available. It already possessed gating function, performing the same separation of electrochemical drive from discharge permissiveness that cholesterol would later enable at the plasma membrane. The Great Oxidation Event therefore represents not the origin of Φ-gating but its elaboration: the addition of a second, independently tunable lipid gate at the outer boundary of a cell that already contained one at its energetic core.

The convergence of memantine, venetoclax, and elamipretide on a single mechanistic description across three biological scales is the framework’s strongest translational claim. These drugs were developed independently, for unrelated indications, by investigators who did not conceive of their targets as instances of a common class. That they nonetheless share a mechanism the framework specifies in advance—and that no additive account provides—suggests the class is real rather than imposed. The prediction that follows is programmatic: other Φ-class gates should be therapeutically addressable by the same logic, and the search for such gates in disease systems is a tractable research program rather than a metaphor.

The 2025 ISR trial failures deserve particular emphasis because they were not anticipated by the field despite strong preclinical evidence and confirmed target engagement. The framework’s account is neither vague nor unfalsifiable: a gating variable requiring a functional operating range cannot be therapeutically normalized by unidirectional suppression, and it specifies the inverted-U dose relationship that should be observed instead. The exploratory high-dose signal in the fosigotifator arm is consistent with this account and inconsistent with a monotonic suppression model. Whether the framework is correct here is determinable by stratified reanalysis of existing trial data, specifically by correlating response with baseline ATF4 or CHAC1 as proxies for eIF2α phosphorylation state.

Finally, the connection between synaptic and mitochondrial gating has implications for neuroinflammatory pathology. When microglia produce 25-hydroxycholesterol in response to pro-inflammatory stimulation, they simultaneously disrupt synaptic gating at the NMDA receptor—impairing potentiation and memory consolidation—and lower the mitochondrial apoptotic threshold, sensitizing neurons to execution under metabolic stress (Izumi et al., 2025). The two disruptions have distinct molecular mechanisms but a single upstream cause. This provides a mechanistic account of why chronic neuroinflammation produces plasticity failure and neurodegeneration as parallel rather than sequential consequences, and predicts that combined intervention across both scales will show supra-additive benefit over single-agent approaches.

11. Conclusion

Before cholesterol, there was cardiolipin. Before eukaryotes, there was gating. The architecture formalized here as R = Φ(A × D × C) ≥ θ is not a property of nervous systems or of the oxygenated world but of lipid membranes with an independently tunable permeability variable, and the first such membrane was bacterial. The four mitochondrial commitment systems examined here are nested instantiations of the same grammar at increasing commitment cost and increasing recovery timescale, from minutes to a lifetime. Eight zero-term veto demonstrations establish the gating variable as formally distinct in each. Three approved therapeutics converge on restoring a gating variable to range across three biological scales. The framework retrodicts a clinical trial failure that surprised the field and specifies the dose relationship that should replace the failed strategy. What remains is the interaction term, and the data required to compute it have already been published.