The science
One law, the architecture it runs on, and the experiment that would refute each part.
Components of this system have been known for decades in separate disciplines. What was missing was the ordered causal object that connects them, the mapping from a failed operation to a therapy, and a standard of proof that asks the body to do it a second time.
The law · the Reserve-Setting Cycle
Every biological response changes the machinery available for the next response.
Two people identical at rest and identical on the first response separate on the second. No clinical test measures the second.
Capacity is written by use. Every challenge leaves a system more able, less able, or wrongly able to meet the next one, and the balance carries forward. That is the whole claim, and it is a claim about direction as much as about loss. Training builds capacity. Sepsis spends it. Chronic infection misroutes it. Cancer takes it.
The cycle runs eight operations. Access, read, command, execute, compensate, terminate, clear, rebuild. Two of them have been mapped molecule by molecule for decades, because reading is the receptor and commanding is the switch. Drawing all eight as one closed loop is what makes the eighth operation visible, and the eighth is the one that decides what the next demand meets.
The same loop runs at both scales. A cell reads its interior. The body reads its blood. Inside a cell, rebuilding is mitochondrial biogenesis and receptor recycling. Across an organism, it is the reconstruction that has to finish before the next demand arrives.
The same load can build capacity or spend it. Trained with recovery, a system rebuilds above baseline. Pushed without it, below. The stimulus is identical. What differs is what the last response did to the machinery.
Disease is failure or misrouting inside the cycle. Aging is the running total of what was not rebuilt. Cancer is the cycle appropriated while the host's own capacity contracts.
Return to baseline is not recovery. Recovery is being able to do it again.
The missing variable · the Competent Receiver
The receptor tests present. The drug has nothing to work on.
An enzyme cuts the binding site off the receptor. It still reads as present, and more ligand cannot rebuild it.
The real target is the working, reusable receptor, not its name. Cutting is one of five ways the usable population disappears while the label says it is there.
Estrogen acts on two layers. Neurons read it as command, vessel support cells as maintenance. At menopause the ligand falls first, while receptors sit in a protease-rich field. Replacement works when the working receptor population is intact, so the receiver decides, not the ligand.
The receiver layer is now measurable. In 54 women imaged with estrogen-receptor PET across the menopause transition, 18 at each stage, receptor density climbs as estradiol falls, independently of age and of circulating estradiol, and classified every participant's stage correctly. Higher density in memory regions tracked worse delayed recall, and predicted mood and cognitive symptoms after menopause.1 More receptor is not more function.
Cleavage is not asserted receptor by receptor. We screen the human receptor complement computationally for protease vulnerability, and the same field that scores estrogen receptors scores the rest. The mechanism is already fixed in a real protein: gingipain, a protease shipped by Porphyromonas gingivalis, degrades the insulin receptor by direct proteolysis in hepatocytes, myocytes and adipocytes, and produces insulin resistance in mice.2 That result is peripheral tissue and mouse, and we cite it for the mechanism rather than for the brain.
A patient can score positive on every test and have almost none that still functions. More drug raises exposure. It does not rebuild the target.
The architecture · the circumventricular sentinel network
A handful of brain sites face the blood and the spinal fluid directly, read what they carry, set what the rest of the body does, and then read what comes back.
One distributed, high-gain control network, not a universal master, where the state of the blood and the spinal fluid becomes systemic command across hundreds of coupled axes, from metabolism and immunity to fluid balance, reproduction, temperature and the clock.
The vessels at these sites are deliberately fenestrated, so what circulates meets neural tissue directly. Arrangement, not molecular uniqueness, decides what shared machinery does.
Area postrema and dorsal vagal complex
Sickness, nausea, emesis, satiety liability, visceral autonomic and inflammatory output.
Median eminence, arcuate nucleus, pituitary stalk
Endocrine, metabolic, reproductive, growth, stress, appetite and allocation command.
Choroid plexus
Spinal fluid composition, immune trafficking, complement and thyroid handling, barrier and cargo distribution.
Subfornical organ, OVLT, neurohypophysis
Osmolality, sodium, thirst, vasopressin, pressure and volume, temperature, fluid and autonomic output.
Pineal
Phase and circadian output.
Tanycytes and perivascular layers
Transfer, transformation, relay, clearance and reconstruction of the interface itself.
Every output passes through eight classes, and the list is complete: endocrine, autonomic, spinal fluid, fluid balance, phase, sickness, barrier and behavioural output. Direct peripheral routes operate in parallel. A disease does not have to pass through a sentinel node to be real or actionable.
Reach comes first because the alternative invites a collapse the evidence does not support. The median eminence is one deeply resolved implementation of the architecture. It is not the architecture.
resolved axes across 38 atlases and seven interfaces. Confidence is stated per axis in the manuscripts, and unresolved axes are counted rather than dropped.
blood-vessel cell nuclei in the adult human median eminence, across 11 donors. Endothelium ranks highest in 977 of 1,000 donor-balanced resamples. Our analysis of the published human hypothalamus atlas.4
The same blood state produces different disease depending on which node and output branch becomes limiting.
A layer of the architecture · external effectors
The receiver, not the source name, decides what happens.
Pathogen, environmental and host-derived classes converge on the same fenestrated interfaces and the same clearance surfaces. Grouping them by where they came from hides the thing they share.
Bacterial vesicles, proteases, endotoxin and toxins. Virions and viral proteins. Secreted enzymes and tissue migration from parasites and fungi. Self-templating seeds and receptor-active mycotoxins.
Heavy metals that persist by ionic mimicry. Inhaled and carrier-bound particles that deposit in vasculature and brain. Receptor-active plasticisers and solvents. Pesticides arriving by gut, portal and blood.
Free heme after haemorrhage. Cytokines that cross barriers and amplify each other. Necrotic cargo and oxidative adducts that overload clearance. Injured-tissue stress signals that reallocate appetite and metabolism.
Porphyromonas gingivalis ships gingipain proteases that degrade the insulin receptor by direct proteolysis.2 SARS-CoV-2 spike protein alone, with no live virus, reduces ACE2 on the endothelial surface and impairs endothelial function.3 Two different attackers, one target class: the receiver.
External is always relative to the receiver. Insulin is external to the cell that reads it, a cytokine to the endothelium, a drug to whichever receiver it lands on.
Because these classes land on shared surfaces, they compete. A clearance route loaded by one input has less left for the next, which is why an exposure history that looks unremarkable item by item can still be limiting when read together.
This is also why source reduction and receiver restoration are different operations. Removing the source stops the loading. It does not rebuild what the loading removed.
Competition for a shared sink is the rule, not the exception. Source control is a companion operation, never a substitute for restoring the receiver.
A layer of the architecture · the sex-composed receiver
Men and women do not run the same receiver layer. Now it is measured.
A hypothalamus atlas at single-cell resolution, bulk tissue across four organs, and a population cohort agree with each other: receiver capacity at the interface is composed differently by sex.
Constitutional, in men
Per cell, male interface tissue runs higher senescence, antigen-presentation, complement and corticosteroid-receiver capacity. Every large split is male-high, present already at age 29, and independent of age.
Prenatal, in women
Female clearance architecture is laid down before birth. Female embryos build APOE-expressing tissue at 2.2 times the male rate. Estrogen holds that loop through reproductive life, and menopause removes the protection.
The twist bulk tissue cannot see
The same gene flips direction between tissues. Direction follows each tissue's cell composition, not the gene. A bulk measurement averages the flip away and reports nothing.
Both sexes clear, and they clear by different routes: male through myeloid cells, female through barrier and parenchyma. In our analysis of 24,727 participants in a national Alzheimer research cohort,6 the interaction between female sex and APOE e4 carries an odds ratio of 1.20, which reads as a second hit on the same prenatal clearance loop. In several million further cells the female age slope is steeper, with a crossover in midlife.
Stated carefully: this makes the female excess in Alzheimer disease a receiver measurement rather than an unexplained epidemiological fact. It does not yet make it a treatment.
These are our analyses of public human resources, not new tissue collection. 433,369 single nuclei across 11 donors in the hypothalamus atlas.4 2,704 bulk samples across four tissues from the human expression project.5 24,727 participants in the cohort.6 A 4.1 million cell embryonic screen.
Transcript maps the tissue layer. The first protein-level test, in a published spinal fluid proteome of 197 donors,7 lands in every predicted direction and none of it is decisive on its own. The composition lives at the blood-facing interface, not behind it.
Two women on one prescription can differ by measured receiver state.
A layer of the architecture · the first reserve
Capacity is constructed before birth.
Before the first demand arrives, a construction project runs. Maternal blood delivers. The placenta rations and converts, setting doses with its own hormones. The fetus grows vessel networks, places receptors and calibrates immune and repair systems. At birth the finished architecture meets breathing, feeding, cold and microbes for the first time.
CRHR1, the stress-phase receiver, carries enhancer wiring in the fetal atlas with no detectable RNA. The transcript count reports nothing. The chromatin reports a receiver being built.
Regulatory competence precedes detectable RNA, so a build programme can be read before any of it has been expressed. This is why counting transcripts in fetal tissue understates what is already committed.
Genes are starting architecture, not destiny. One construction constraint can present as several different diagnoses depending on which demand recruits the system first, which is a statement about arrangement rather than about any single gene.
enhancer to gene links across 203 fetal construction groups, resolved in our analysis of a published paired fetal RNA and chromatin atlas of 817,740 cells. The build programme is set before any of it appears as RNA.
cells in the embryonic screen behind the prenatal clearance result.
Birth size and resting output do not certify the reserve a life will have.
How the evidence is handled
A mechanism inherits its lowest unresolved layer.
The architecture is written down, with its ownership boundaries and its falsifiers. The three completed tests run on published data generated by other groups, which is the strongest position available before independent replication of our own work exists.
What does not exist yet: one decisive result in a living human system showing second-response prediction where resting and first-response measures fail, and independent replication of it. Those are the next two steps, in that order, and no claim on this site runs ahead of them.
Manuscripts covering the law, the receiver, the network, the input layer and the developmental construction of reserve are written and in preparation. None is released as a preprint yet. Researchers who want to read one before release can ask.
Falsifiers
Each discovery carries the experiment that would refute it.
These are published as open problems rather than held as internal milestones. A law that tells the field how to prove it wrong should be treated differently from a framework that asks to be believed.
| Discovery | The decisive test | What would refute it |
|---|---|---|
| The Reserve-Setting Cycle | The law predicts the second response from the recovery between responses. Resting measurements cannot do this. | Resting and first-response measures predict later function as well as the recovery between them, across several organ systems. |
| The Competent Receiver | Target capacity predicts response where every conventional target metric fails, and restoration survives withdrawal. | Transcript, total protein, occupancy or diagnosis predicts drug response as well as usable target capacity does. |
| The sentinel network | Perturb the gate, measure the routed output, exclude the parallel route. | Node-specific perturbation produces no routed output beyond what direct peripheral routes already explain. |
The framework does not depend on any single experiment, and it names the ones that would end it.
References
Sources for every result on this page that is not ours.
Two kinds of source appear here. Published results we cite as evidence, and public human resources we analysed. Both are listed, because a number is only as good as the reader's ability to go and check it.
- Mosconi L, Nerattini M, Matthews DC, et al. In vivo brain estrogen receptor density by neuroendocrine aging and relationships with cognition and symptomatology. Scientific Reports. 2024;14(1):12680. doi:10.1038/s41598-024-62820-7 ↩
- Liu F, Zhu B, An Y, et al. Gingipain from Porphyromonas gingivalis causes insulin resistance by degrading insulin receptors through direct proteolytic effects. International Journal of Oral Science. 2024;16(1):53. doi:10.1038/s41368-024-00313-z ↩
- Lei Y, Zhang J, Schiavon CR, et al. SARS-CoV-2 spike protein impairs endothelial function via downregulation of ACE 2. Circulation Research. 2021;128(9):1323-1326. doi:10.1161/CIRCRESAHA.121.318902 ↩
- Tadross JA, Steuernagel L, Dowsett GKC, et al. A comprehensive spatio-cellular map of the human hypothalamus. Nature. 2025;639(8055):708-716. doi:10.1038/s41586-024-08504-8 ↩
- GTEx Consortium. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science. 2020;369(6509):1318-1330. doi:10.1126/science.aaz1776 ↩
- Beekly DL, Ramos EM, Lee WW, et al. The National Alzheimer's Coordinating Center (NACC) database: the Uniform Data Set. Alzheimer Disease and Associated Disorders. 2007;21(3):249-258. doi:10.1097/WAD.0b013e318142774e ↩
- Bader JM, Geyer PE, Müller JB, et al. Proteome profiling in cerebrospinal fluid reveals novel biomarkers of Alzheimer's disease. Molecular Systems Biology. 2020;16(6):e9356. doi:10.15252/msb.20199356 ↩
The three completed tests, in cachexia, sudden infant death syndrome and the ponsegromab trial, are cited in full on the overview page.
Contact
Read it, then try to break it.
We are looking for groups willing to run the decisive tests, in their cohorts, on their terms.
fs@cirsentra.com for manuscripts, evidence boundaries and collaboration.
The programme for the translational and commercial position.