The Three Layers of Intervention
Output, Signalling, and Structure: A Depth Model for Biological Intervention
Most therapeutic interventions act on what can be most easily measured: the outputs of biological systems. Glucose. Cholesterol. Blood pressure. Inflammatory markers. These are real, important signals, and modifying them produces real, important effects. But outputs are downstream of the systems that generate them, and acting on an output does not necessarily change the system that produced it. When the intervention is withdrawn, the system returns to what it was.
This paper proposes a three-layer model for understanding the depth at which biological interventions act: Output Modulation, which stabilises measurable readouts; Signalling Modulation, which changes how those outputs are generated; and Structural Modulation, which alters the baseline state from which all signals emerge. The structural layer encompasses four distinct targets: adipose tissue distribution, mitochondrial population quality, gut microbiome architecture, and — addressed in the June 2026 revision — haematopoietic and lymphoid niche architecture, the slowly changing tissue infrastructure that generates immune surveillance capacity. The model is illustrated through the quiet biology protocol, which deliberately spans all three layers in a coordinated, temporally sequenced architecture.
The central argument is expressed in a single proposition: for chronic adaptive biological systems, durability tends to increase with intervention depth. Output-layer interventions produce fast, measurable, reversible effects. Structure-layer interventions are slower, harder to measure, and more persistent. Effective long-term biological control likely requires coordination across all three layers — not because any single layer is insufficient, but because the system will tend to return to its structural default unless that default is itself modified. This proposition does not apply to all biological interventions: curative surgery, targeted eradication of an infectious agent, and smoking cessation can all produce durable outcomes through mechanisms that do not fit neatly into the depth model. The framework's domain is chronic adaptive systems — biological conditions that are continuously generated by the system's own architecture rather than imposed by a removable external cause.
01The problem with acting at the surface
Biology has layers. The layer most accessible to measurement — and therefore most accessible to clinical guidelines and trial endpoints — is the output layer: the concentrations, rates, and ratios that blood tests, imaging, and physiological monitoring can capture. These outputs are real and meaningful. They reflect the state of underlying systems. But they are downstream of those systems, not identical to them.
A drug that lowers blood glucose does not necessarily change the insulin resistance, mitochondrial dysfunction, or chronic mTOR activation that is generating the elevated glucose. It modifies the output. The generating system continues. When the drug is withdrawn, the output returns because the system that produced it was never modified.[1]
This is the fundamental limitation of output-layer intervention, and it is not a criticism of any specific drug or guideline. It is a description of what output-layer intervention is and is not. It is a powerful and often necessary tool for managing acute pathology, reducing immediate risk, and buying time for deeper changes to occur. What it is not, on its own, is a strategy for changing what the system tends to return to when the intervention is withdrawn.
The three-layer model this paper describes is an attempt to make this distinction explicit, to map the protocol against it, and to argue that for chronic adaptive biological systems, the most durable outcomes are those achieved by coordinating interventions across all three layers simultaneously.
Most therapies operate at the level of output. Some influence signalling. Few alter structure. For chronic adaptive systems, durability tends to increase with intervention depth.
02The three layers
Figure 1 maps the three layers, their biological targets, the protocol agents acting at each layer, and their role in the overall system. The layers are not discrete compartments — they interact continuously, and an intervention at one layer will have effects at others. The distinction is one of primary mechanism and temporal depth rather than complete separation.
A note on terminology: the word structural is used here in the QB framework sense — meaning slowly changing biological infrastructure that determines the baseline from which outputs and signals emerge — rather than in a strict anatomical sense. The targets it describes are biologically heterogeneous: marrow architecture is genuinely anatomical, mitochondrial population quality is a quality-control process, microbiome composition is ecological, adipose distribution is metabolic, and haematopoietic niche architecture spans tissue and cell-population biology. What they share is not biological category but temporal depth and functional role: they change slowly, they are not captured in standard monitoring, and they strongly influence what the system tends to return to.
Figure 1: The three layers of intervention
Layers interact continuously. Primary mechanism and temporal depth distinguish them.
Layer one: output modulation
Output modulation acts on the most immediately measurable products of biological activity. In the protocol, retatrutide is the primary output modulator: its triple receptor agonism reduces appetite, lowers glucose, improves insulin dynamics, and reduces the inflammatory and metabolic noise that impairs signalling clarity at deeper layers.[2]
Exercise, in its acute phase, contributes to output modulation through insulin-independent glucose uptake driven by AMPK activation in muscle tissue.[3]
A clarification is needed before treating outputs as merely downstream indicators: outputs are both signals and active contributors to disease processes. Hyperglycaemia directly causes glycation damage and vascular injury. Hypertension causes mechanical remodelling of arterial walls. Chronically elevated LDL participates actively in plaque formation rather than simply marking it. Acting on these outputs has genuine therapeutic value independently of whether the generating system is modified. The limitation of output-layer intervention is not that it is superficial in the pejorative sense — it is that it does not, on its own, change what the system will tend to return to when the intervention is withdrawn. For chronic adaptive conditions, output control and structural change are complementary, not competing strategies.
Output modulation is fast and often reversible. Its value in the protocol is not to produce durable change directly but to reduce the noise and pressure that would otherwise impair the clarity of signalling-layer interventions. A system operating under chronic metabolic excess — elevated insulin, high inflammatory tone, fatty liver — is a system in which the deeper signals of the protocol cannot be read cleanly. Output modulation clears the environment that the signalling work operates within.
A note on retatrutide and layer depth: retatrutide appears in both the output and structure rows of Figure 1, and that dual placement is deliberate. Its immediate effects — glucose reduction, appetite suppression, improved insulin dynamics — are output-layer events. But sustained use produces a different order of change: visceral adipose reduction, hepatic fat clearance, and systemic insulin field restoration that accumulate over months and alter the structural baseline from which all other signals emerge. The output effects are the mechanism; the structural effects are the consequence of sustaining that mechanism over time. Its triagonist architecture is relevant to its structural role: the glucagon, GIP, and GLP-1 arms work in concert, and the balance between them has implications for bone and marrow addressed in the dedicated Retatrutide paper (QB Framework Paper 13).
Layer two: signalling modulation
Signalling modulation acts on the pathways through which the cell generates its outputs and makes its decisions. It does not merely change what is measured. It changes how the measurement comes to be. Rapamycin's weekly mTOR inhibition, doxycycline's mitochondrial ribosome stress, the chronic AMPK upregulation produced by exercise, and the hormonal signalling environment maintained by TRT and aromatase inhibition are all signalling-layer interventions: they modify the behaviour of the system, not merely its current readouts.[4]
Signalling modulation is less immediately reversible than output modulation but more so than structural modulation. The mTOR oscillation that rapamycin produces lasts days, not hours, and accumulates across cycles. The mitochondrial selection pressure of doxycycline's stress phase extends into the washout. The AMPK upregulation of chronic exercise training persists for weeks in the absence of continued training. These are not permanent changes, but they are not the rapid-reset pharmacokinetics of output modulation either.
Layer three: structural modulation
Structural modulation acts at the deepest layer: the slowly changing biological infrastructure that determines the baseline from which all signals emerge. The structural layer encompasses four distinct targets, each operating through different mechanisms but sharing the property of slow change and persistence beyond any individual intervention.
Pioglitazone's remodelling of adipose tissue distribution, Urolithin A's renewal of the mitochondrial population through PINK1–Parkin mediated mitophagy, and the long-term architecture of the gut microbiome maintained by PHGG and fermented food are the three targets described in the original paper. The June 2026 revision adds a fourth: the haematopoietic and lymphoid niche — the bone marrow and thymic tissue systems that generate immune surveillance capacity, and whose degradation with age follows a pattern common to the other structural targets.[5]
Before describing each structural target and the agents that work on it, it is worth naming the deeper pattern that connects several of them. Ageing does not simply accumulate fat in random locations. Across multiple organ systems, ageing appears to progressively replace specialised regenerative tissues with adipose tissue — storage infrastructure that is metabolically active but functionally inhibitory to what it displaces:
- Active bone marrow → regulated marrow adipose tissue (rMAT): haematopoietic niche displaced and actively suppressed
- Thymic lymphoid tissue → thymic fat: T cell educational architecture lost
- Skeletal muscle → myosteatosis: contractile function replaced by intramuscular fat infiltration
- Liver parenchyma → hepatic steatosis: metabolic processing capacity degraded
The common denominator is not lipid accumulation per se but the replacement of stem-cell-dependent regenerative capacity with adipose tissue that is often inhibitory to the function it displaces. Proposed shared drivers include age-related hormonal withdrawal, chronic low-grade inflammatory signalling, altered mesenchymal stem cell fate decisions, and reduced regenerative demand. This framing connects immunosenescence, osteopenia, marrow decline, sarcopenia, and metabolic disease under a single architectural concept — and positions structural modulation as a defence of regenerative tissue function across multiple organ systems simultaneously.
Ageing may be partly understood as the progressive replacement of regenerative infrastructure by storage infrastructure — a shift that degrades immune competence, skeletal integrity, metabolic flexibility, and tumour surveillance through a common upstream mechanism.
Pioglitazone — adipose remodelling
Pioglitazone at 7.5mg daily activates PPAR-γ, the master transcriptional regulator of adipocyte differentiation and lipid metabolism. Its primary structural effect is the redistribution of lipid away from visceral and hepatic depots toward subcutaneous storage — a shift that improves insulin sensitivity, reduces pro-inflammatory adipokine output, and reduces aromatase activity in adipose tissue. The remodelling is cumulative and occurs over months, not days. A single monitoring panel measurement will not capture it. The structural shift it produces — a different adipose architecture — determines the metabolic field conditions that all other protocol agents operate within.
Urolithin A — mitochondrial population renewal
Urolithin A drives targeted mitochondrial clearance through the PINK1–Parkin mediated mitophagy pathway. The effect is selective: mitochondria that have lost membrane potential are tagged for clearance; functional mitochondria are spared. The net result, accumulated across cycles, is a shift in the mitochondrial population toward a higher proportion of functional mitochondria — inferred from molecular signatures of improved mitochondrial and cellular health in human studies rather than directly demonstrated by population-level functional measurement. The downstream functional consequences — better metabolic efficiency, reduced ROS output, improved capacity for autophagy and cellular repair — are the logical consequence of that population shift, not yet independently confirmed endpoints at this stage.[5]
PHGG and fermented foods — microbiome architecture
Partially hydrolysed guar gum (PHGG) as a daily prebiotic and miso soup four times weekly provide the dietary substrate and microbial input that maintain gut microbiome architecture across the full cycle, including through the doxycycline alternating weeks that would otherwise disrupt the microbiome. The structural target here is not any single microbial species but the overall community composition — the balance of populations that regulate systemic inflammatory tone and the entero-hepatic cycling of oestrogen metabolites. Of the four structural targets, the microbiome carries the thinnest direct evidence chain: the causal path from PHGG plus fermented food to a specific microbiome composition to clinically meaningful prostate cancer consequences is less established than the corresponding chains for adipose remodelling, mitochondrial quality, or haematopoietic niche biology. The rationale is mechanistically grounded and the intervention is low-risk; the evidence gap is acknowledged.
Haematopoietic and lymphoid niche — immune surveillance infrastructure
The fourth structural target operates by a different logic from the first three: it is the architecture of the tissues that generate immune surveillance capacity, and its degradation is invisible to standard monitoring.
The immune system does not arrive at the tumour as a finished product. It is continuously generated, educated, and maintained by two tissue systems whose quality degrades with age in ways that are directly sensitive to metabolic and inflammatory terrain: the bone marrow haematopoietic niche, which produces the cellular raw material of immune competence, and the thymic architecture, which educates T cells in self/other discrimination. Both are part of the regenerative infrastructure replacement pattern described above.
Terrain is not only the environment the tumour grows in. It is also the environment the immune system operates from. Both are degraded by the same upstream conditions.
Active red marrow is progressively replaced by regulated marrow adipose tissue (rMAT) with age. Marrow adipocytes actively suppress haematopoietic stem cell function through local cytokine, fatty acid, and adipokine secretion, and bias mesenchymal progenitors toward further adipogenesis in a self-reinforcing loop. rMAT expansion is associated with impaired lymphopoiesis and may contribute to reduced generation or function of T cell and NK cell lineages.
T cell education occurs in the thymus through a selection process of considerable biological severity: roughly 95 to 98 percent of thymocytes do not survive it. Thymic involution — progressive fat replacement of lymphoid architecture from puberty onward — produces repertoire contraction: reduced capacity to recognise novel antigens, including tumour neoantigens. Volume and clonal diversity are not the same thing: even when thymic rebound produces a measurable increase in circulating naïve T cells, the breadth of the clonal repertoire remains constrained by age-related changes in both the thymic microenvironment and the upstream haematopoietic system.
NK cells develop and undergo KIR-based licensing in the bone marrow without thymic involvement, but their functional competence is continuously shaped by the peripheral stromal niche — secondary lymphoid organs, liver sinusoids, peripheral tissues — through IL-15, IL-12, and IL-18 signalling from macrophages and dendritic cells whose polarisation is directly sensitive to systemic metabolic and inflammatory conditions. In a chronically inflamed, metabolically dysregulated terrain, NK cell dysfunction is not primarily numerical. Although these relationships are supported mechanistically, direct clinical demonstration that correcting metabolic terrain variables restores tumour immune surveillance remains limited.
NK cell immune surveillance competence is not a fixed biological given. It is a dynamic output of niche quality — and niche quality is sensitive to the same metabolic and inflammatory variables that the rest of the structural layer addresses.
None of the protocol agents target NK cells, marrow niche quality, or thymic architecture directly. What this section establishes is that the terrain corrections already in the protocol — visceral adipose reduction through Retatrutide, inflammatory tone reduction through metabolic normalisation, and the exercise foundation — address the upstream conditions that drive rMAT expansion and NK niche degradation. Exercise is the clearest example: its evidence for maintaining NK cytotoxic activity operates through exactly this logic. In a prostate cancer context, ADT/BAT sequencing represents deliberate exploitation of the immune terrain window — the transient thymic rebound of Phase 1 — before Phase 2 terrain degradation overtakes the benefit. The full argument is developed in QB Paper 17 and the QB Research Note on Aged Immune Terrain (June 2026).
The evidence for rMAT-mediated lymphopoietic impairment as a clinically meaningful driver of immune surveillance failure in prostate cancer specifically is plausible from mechanism but not yet established in direct clinical evidence. The corrections are already being made for other reasons. That is the appropriate level of claim.
Outputs can be stabilised and signals can be modulated, but it is structure that strongly influences what the system will tend to return to. An intervention that does not reach the structural layer leaves the default state largely unchanged.
03The protocol mapped across three layers and time
The three-layer model becomes most powerful when combined with the temporal dimension. The protocol does not simply address all three layers simultaneously. It sequences its interventions so that each phase operates primarily on the layer most accessible in that window, while the other layers continue at their own pace.
Figure 2 maps the twelve-week cycle across the three layers, showing which agents are active in which phase and what biological work each phase is primarily doing. The continuous foundation column shows the agents that maintain the metabolic field and structural baseline throughout all phases of the cycle. The cycle is presented across two tables to preserve readability: Figure 2a covers the active block and continuous foundation; Figure 2b covers consolidation, the clear window, and continuous foundation.
Figure 2a: Weeks 1–8 (Active Block) and Continuous Foundation
Green = pressure/re-patterning phase. Unshaded blue = continuous foundation maintained throughout all phases.
Figure 2b: Weeks 9–12 (Consolidation and Clear Window) and Continuous Foundation
Amber = cleanup/consolidation. Rose = clear window/true system readout. Unshaded blue = continuous foundation.
Weeks 1–8: pressure and re-patterning
The active block operates primarily at the signalling layer, with output modulation providing the clean metabolic environment the signalling work requires. Rapamycin creates the mTOR oscillation window. Doxycycline applies the mitochondrial stress that gives the autophagy system specific targets to act on. Exercise drives AMPK activation and compounds the metabolic field improvement that retatrutide is maintaining continuously. The structural layer is moving slowly in the background — adipose remodelling continues under pioglitazone, microbiome architecture is supported through the doxycycline cycles by PHGG, and the systemic metabolic correction of Retatrutide is continuously attenuating the conditions that drive rMAT expansion.
This is the pressure and re-patterning phase. The signalling environment is being actively perturbed. The question the biology is being asked to answer is: which cellular components are robust enough to survive the combined stress of mTOR suppression and mitochondrial ribosome inhibition, and which are not?
Weeks 9–10: cleanup and consolidation
The washout's first half operates primarily at the structural layer, with signalling quietening but not yet silent. Urolithin A drives targeted mitochondrial clearance through the PINK1–Parkin pathway — a more specific mitophagy induction than rapamycin's broader autophagy activation.[5] Chinese skullcap's NF-κB suppression and macrophage repolarisation extend the anti-inflammatory pressure of the active block into the consolidation phase while the pharmacological burden reduces. The structural layer is doing its most active work: renewing the mitochondrial population that the stress phase has prepared for clearance.
Weeks 11–12: true system readout
The clear window is the most structurally revealing phase. All protocol compounds are withdrawn. The output layer returns to its true baseline, unperturbed by pharmacology. The signalling layer normalises. The structural layer expresses whatever the preceding ten weeks have actually achieved — the mitochondrial population that has been renewed, the adipose architecture that has been remodelled, the microbiome that has been maintained, the haematopoietic niche conditions that the metabolic corrections have been sustaining.
This is when PSA is measured, and when the full monitoring panel is most informative. The reading is not of any individual intervention. It is of the system as it now is — its structural baseline, expressed through its unperturbed outputs and signalling behaviour. The system readout tells you whether the accumulated cycles are shifting the structural default in the intended direction.
The monitoring panel at the clear window is not measuring the effects of any specific drug. It is measuring the structural state of the system — the baseline from which all signals emerge — after ten weeks of coordinated intervention across all three layers. That is a different and more meaningful measurement than any output captured during the active block.
04Why this model matters beyond the protocol
The three-layer framework has implications that extend well beyond the specific protocol it was developed to describe. It offers a way of categorising any therapeutic intervention — pharmacological, nutritional, or behavioural — by the depth at which it acts and the durability of effect that depth implies.
An antihypertensive drug lowers the blood pressure number. It does not modify the vascular stiffness, endothelial dysfunction, or chronic sympathetic activation that is generating the elevated pressure. A glucose-lowering agent reduces HbA1c. It does not change the insulin resistance, mitochondrial dysfunction, or ectopic lipid accumulation that is driving the elevated glucose. A GLP-1 agonist used in isolation acts primarily at the output and signalling layers: it reduces glucose and appetite and improves insulin dynamics. Used within a framework that also addresses the structural layer — mitochondrial quality, adipose architecture, microbiome composition, haematopoietic niche integrity — its effects are more durable because the structural default is simultaneously being modified.
The deepest formulation of the three-layer model is not mechanistic but temporal. The three layers are really three timescale regimes: the output layer operates on a timescale of hours to days, the signalling layer on days to weeks, and the structural layer on months to years. Durable change in chronic adaptive systems tends to require altering slower variables — a principle from dynamical systems theory with well-established analogues in ecology, climate science, and complex systems biology.[13,14] The therapeutic implication is that any intervention fast enough to be easily measured is likely too fast to alter the structural default. This temporal framing is arguably more fundamental than the mechanistic one: it explains why structure is hard to change, why monitoring must be designed differently for different layers, and why the clear window at Weeks 11–12 is a readout of the slowest variables in the system rather than of any specific intervention.
Biological systems contain variables operating at different characteristic timescales. For chronic adaptive systems, durable change occurs when slower variables are altered — not merely when faster ones are suppressed.
The temporal framing connects to a deeper systems-biology concept that the framework implicitly invokes throughout but does not yet name explicitly. The language used — structural default, return to baseline, what the system tends to return to — is the language of attractor states. In dynamical systems theory, an attractor is the state or set of states toward which a system tends to evolve from a broad range of starting conditions. The structural layer in this model is, in these terms, the set of slowly moving variables that define the attractor basin: the metabolic, regenerative, and immune architecture that determines where the system settles when faster perturbations resolve. Output-layer interventions move the system within the basin. Signalling-layer interventions shift the trajectories within it. Structural-layer interventions are attempts to move the basin itself — to alter the attractor state so that the system settles at a different default. This is why structural change is slow, hard to measure, and more durable than changes at the faster layers. And it is why the clear window at Weeks 11–12 is biologically significant: it is the moment when the fast variables have resolved and what remains visible is the attractor state as it currently stands.[13,14]
The model also clarifies why the sequence may matter as much as the stack. Running signalling-layer interventions into an output layer that has not been stabilised risks doing signalling work in a noisy metabolic environment. Running structural-layer interventions before the signalling work has done its selection may mean building on a substrate that has not yet been reformed by the selection pressure. The phases are not necessarily interchangeable — the logic is that each layer may provide conditions under which the next can operate more effectively. This is a reasoned hypothesis about sequencing rather than a demonstrated requirement; direct evidence for the specific ordering described here has not been tested in a clinical trial.
Most clinical intervention sits predominantly at the output layer, sometimes reaching into signalling. Structure is rarely addressed directly or systematically, partly because it is slow, partly because it is hard to measure, and partly because the clinical trial infrastructure — with its defined endpoints and fixed durations — is poorly suited to capturing changes that accumulate over months and express themselves in the behaviour of cellular populations rather than in the level of a single measurable marker.
The gap between what this framework achieves and what standard clinical approaches achieve is not primarily a gap in the agents used. It is a gap in the depth reached and the coordination maintained across that depth over time.
05Honest limitations
The three-layer model is a conceptual framework, not a validated clinical classification system. The assignment of specific agents to specific layers reflects their primary mechanisms of action, but all agents have effects across multiple layers — the model simplifies a complex reality in order to make it usable, not to describe it with complete accuracy.
The hierarchy implicit in the model — structure as the deepest and most determinative layer — is not universal. In developmental biology, endocrine transitions, and acute immune responses, signalling variables can dominate biological outcomes independently of structural state. The three-layer model describes the architecture of chronic adaptive conditions maintained by the system's own metabolic, inflammatory, and regenerative infrastructure. It does not claim that structure always dominates, only that in this domain, altering structural defaults tends to produce more durable outcomes than acting on faster variables alone.
The evidence base for the structural layer is the least developed of the three. Adipose remodelling under pioglitazone, mitochondrial population renewal under Urolithin A, microbiome architecture shifts from dietary intervention, and haematopoietic niche quality maintenance through metabolic correction are all supported by evidence, but the translation of these effects into durable clinical outcomes in the specific context of this protocol has not been directly studied. The structural argument is mechanistically grounded and clinically plausible. It is not yet clinically proven in this combination.
The word structural, as used in this document, is a QB framework designation rather than a precise biological category. The targets described are biologically heterogeneous — what they share is temporal depth and functional role, not biological class.
The claim that durability tends to increase with intervention depth in chronic adaptive systems is a logical proposition supported by dynamical systems principles and the mechanistic understanding of each layer. It is not yet the conclusion of a clinical trial designed to test it, and it applies to a specific domain: chronic adaptive conditions generated by the system's own architecture, not to all biological interventions.
The three-layer model — Output, Signalling, Structure — provides a framework for understanding why some interventions produce durable change and others do not. Output modulation is necessary and fast. Signalling modulation changes how outputs are generated. Structural modulation changes what the system tends to return to.
The structural layer now encompasses four targets: adipose distribution, mitochondrial population quality, gut microbiome architecture, and haematopoietic and lymphoid niche integrity. All four share the same property: they are slowly changing biological infrastructure that strongly influences the baseline from which all outputs and signals emerge. Their degradation with age follows a common pattern — the progressive replacement of regenerative capacity with adipose tissue — and their correction is addressed, directly or indirectly, by the protocol's existing agents and lifestyle architecture.
The quiet biology protocol is designed to span all three layers in a coordinated temporal sequence: stabilising the output environment first, perturbing and re-patterning the signalling layer during the active block, renewing the structural substrate during consolidation, and then reading the true system state during the clear window when all pharmacological noise has resolved.
The result is not a collection of interventions. It is an architecture of depth — one that treats the durability of biological change as a function of how far into the system's generative layers the intervention has reached, and how persistently the slower variables of the system have been shifted. In dynamical systems terms, it is an attempt to move the biological attractor state: to alter what the system tends to return to, not merely what it is doing at any given moment.
Outputs can be stabilised. Signals can be modulated. But it is the attractor state — the slowly changing structural default — that determines what the system will tend to return to.
Appendix: QB structural layer — four-target summary matrix
- Laplante M, Sabatini DM. mTOR signaling in growth control and disease. Cell. 2012;149(2):274–293. For the output-versus-system distinction in metabolic pharmacology: DeFronzo RA. Banting Lecture: from the triumvirate to the ominous octet: a new paradigm for the treatment of type 2 diabetes mellitus. Diabetes. 2009;58(4):773–795.
- Jastreboff AM, Kaplan LM, Frías JP, et al. Triple-hormone-receptor agonist retatrutide for obesity — a phase 2 trial. New England Journal of Medicine. 2023;389(6):514–526. Phase 3 confirmation: TRIUMPH-4 topline results, Eli Lilly, December 2025.
- Cantó C, Jiang LQ, Deshmukh AS, et al. Interdependence of AMPK and SIRT1 for metabolic adaptation to fasting and exercise in skeletal muscle. Cell Metabolism. 2010;11(3):213–219.
- Arriola Apelo SI, Neuman JC, Baar EL, et al. Alternative rapamycin treatment regimens mitigate the impact of rapamycin on glucose homeostasis and the immune system. Aging Cell. 2016;15(1):28–33. Also: Lamb R, Ozsvari B, Lisanti CL, et al. Antibiotics that target mitochondria effectively eradicate cancer stem cells. Oncotarget. 2015;6(7):4569–4584.
- Andreux PA, Blanco-Bose W, Ryu D, et al. The mitophagy activator urolithin A is safe and induces a molecular signature of improved mitochondrial and cellular health in humans. Nature Metabolism. 2019;1(6):595–603. Also: Ricote M, Li AC, Willson TM, Kelly CJ, Glass CK. The peroxisome proliferator-activated receptor-γ is a negative regulator of macrophage activation. Nature. 1998;391(6662):79–82.
- Ambrosi TH, Schulz TJ. Adipogenic and haematopoietic niche regulation by marrow adipose tissue. Nature Reviews Endocrinology. 2021;17(7):403–416.
- Marinelli Busilacchi E, Morsia E, Poloni A. Bone marrow adipose tissue. Cells. 2024;13(9):724. doi:10.3390/cells13090724.
- Zhang X, Tian L, Majumdar A, Scheller EL. Function and regulation of bone marrow adipose tissue in health and disease: state of the field and clinical considerations. Comprehensive Physiology. 2024;14(3):5521–5579. doi:10.1002/cphy.c230016.
- Kousa AI, Jahn L, Zhao K, et al. Age-related epithelial defects limit thymic function and regeneration. Nature Immunology. 2024;25(9):1593–1606. doi:10.1038/s41590-024-01915-9.
- Liu Z, Liang Q, Ren Y, et al. Immunosenescence: molecular mechanisms and diseases. Signal Transduction and Targeted Therapy. 2023;8(1):200. doi:10.1038/s41392-023-01451-2.
- Hou J, et al. Aged bone marrow macrophages drive systemic aging and age-related dysfunction via extracellular vesicle-mediated induction of paracrine senescence. Nature Aging. 2024;4:1562–1581. doi:10.1038/s43587-024-00694-0.
- Sutherland JS, Goldberg GL, Hammett MV, et al. Activation of thymic regeneration in mice and humans following androgen blockade. Journal of Immunology. 2005;175(4):2741–2753. doi:10.4049/jimmunol.175.4.2741.
- Scheffer M, Bascompte J, Brock WA, et al. Early-warning signals for critical transitions. Nature. 2009;461(7260):53–59. doi:10.1038/nature08227.
- Strogatz SH. Nonlinear Dynamics and Chaos: With Applications to Physics, Biology, Chemistry, and Engineering. 2nd ed. Westview Press; 2014.
- Laczkó D, Oláh J, Varga I, et al. Exercise and natural killer cell function: a systematic review. Journal of Physiology. 2020;598(6):1167–1185.
- Gruver AL, Hudson LL, Sempowski GD. Immunosenescence of ageing. Journal of Pathology. 2007;211(2):144–156.
- IndolentRead phase →
- InflectionRead phase →
- ActiveRead phase →