Null Subtyping - Absent vs. Blocked vs. Overwhelmed vs. Biased
Certainty: 0.20. The null subtyping framework extends the binary null logic of The Null Ladder - Why Null Is More Informative Than Positive and Confirmation Bias in Patient Self-Reported Cycle-Phase Drug Response with a pharmacological distinction that the existing framework acknowledges implicitly but never formalizes. The distinction between absent, blocked, overwhelmed, and biased-signalling receptor states is grounded in century-old receptor occupancy theory (Finlay, Duffull, and Glass 2020) and the well-established pharmacology of autoantibodies as endogenous GPCR modulators (Skiba and Kruse 2021), but the specific diagnostic probes proposed here have never been tested in ME/CFS. None of the probes described below are clinically available — this section is a conceptual framework for research design, not a clinical protocol. The framework is logically coherent and mechanistically specific — but zero empirical validation exists in this disease population.
The null ladder tells us a drug can eliminate multiple mechanisms in one trial. The null matrix formalizes the pattern of nulls as Bayesian evidence. But both treat null as binary: mechanism eliminated or not. Receptor pharmacology reveals this is a lossy compression. A drug failure can mean any of three distinct states (with biased signalling as a special case of BLOCKED — see line 38). In real tissue, receptor populations are mixtures; the categories are analytical distinctions for single-receptor molecules, and the observed response reflects the net output of the mixed population.
1 The Three Null Subtypes (plus One Variation)
Note: The subtypes below are defined as analytical categories for a single receptor molecule. In whole-tissue pharmacology, receptor populations are mixtures (some blocked, some functional, some internalized), and an observed null response reflects the net output of the mixed population. The probes diagnose the dominant state; they cannot guarantee uniform receptor status across every molecule in the tissue.
Scope: This taxonomy addresses receptor-level reasons a drug fails to produce its expected effect — assuming the drug reached the target at adequate concentration for adequate duration. It does not cover pharmacokinetic failure modes (non-adherence, malabsorption, CYP-mediated rapid clearance, drug interactions preventing target engagement) or trial-design failures (inadequate dose, insufficient duration, wrong outcome measure). These are covered separately in The Null Ladder - Why Null Is More Informative Than Positive (null failure modes). The framework layers on top of the existing null ladder, not in place of it.
The receptor, ion channel, enzyme, or transporter that the drug targets never existed in this patient’s tissue — or was lost irreversibly through cell death, tissue atrophy, or developmental failure. This is a true negative: the drug never had a target.
Distinguishing probe — Bypass agonist. Give an agent that bypasses the receptor entirely and acts downstream. Bypass success shows the downstream machinery is intact — consistent with ABSENT, BLOCKED, or OVERWHELMED (the receptor may be present but inaccessible; the probe only confirms the machinery beyond it works). Bypass failure → downstream machinery is also broken, but this says nothing about the receptor. The bypass probe cannot isolate ABSENT alone; it requires combining with the allosteric probe (to exclude BLOCKED) and the concentration-competition probe (to exclude OVERWHELMED).
Examples:
- Direct calcium ionophore for TRPM3 — if it produces calcium flux when LDN/pregnenolone cannot, the calcium-handling machinery is intact, consistent with TRPM3 being absent, blocked, or desensitized. (Note: LDN and pregnenolone’s action through TRPM3 is itself an inferential premise.) If the ionophore fails, the calcium machinery is defective at a broader level. Note: ionophores (ionomycin, A23187) are lab reagents, not clinically administrable in humans.
- L-DOPA for striatal dopamine — if L-DOPA produces motor or cognitive improvement when aripiprazole does not, the downstream postsynaptic machinery and dopamine synthesis machinery are intact. Note: L-DOPA is converted to dopamine and still acts through D2/D3 receptors — it does NOT bypass the receptor. The inference it supports is that presynaptic dopamine synthesis is intact, not that the receptor is absent. True receptor bypass would require a postsynaptic D2/D3-independent agent, which does not exist.
- Beta-3 adrenergic agonist (mirabegron) when β2-AR agonist produces null — β2 and β3 share downstream Gαs/cAMP machinery. If β3 works, the downstream signalling apparatus is intact, consistent with a β2-specific problem (absent, blocked, or biased). If β3 also fails, the Gαs pathway is broadly dysfunctional. Note: β3 response alone cannot distinguish ABSENT from BLOCKED. Note: human β3-AR expression is primarily adipose/bladder, limiting inference to non-CNS tissues.
Evidence basis: The logic follows from receptor pharmacology — if a downstream agonist bypasses the missing receptor and produces the expected intracellular response, the receptor was the missing component. No ME/CFS studies have tested this probe systematically. (Finlay, Duffull, and Glass 2020)
Certainty: 0.20. The probe logic is pharmacologically sound but untested in this clinical context.
The receptor is structurally present in the tissue but its orthosteric (primary) binding site is occupied — by an autoantibody, a competitive endogenous ligand, or a post-translational modification that locks the receptor in an inactive conformation. The drug cannot bind because the site is already taken.
Distinguishing probe — Allosteric ligand. Use a drug that binds a different site on the same receptor — an allosteric pocket. If the allosteric ligand produces a response when the orthosteric ligand does not, the orthosteric site is blocked but the receptor is structurally present and functionally capable of signaling through the allosteric pathway.
Examples:
- β2-AR autoantibody (AAb) occupying the orthosteric site → salbutamol and isoproterenol both fail (both are orthosteric agonists; both bind the same site). But a genuine allosteric modulator (such as Cmpd-6, a small-molecule PAM directed at the intracellular β2-AR–Gαs interface (Dodge-Kafka et al. 2026)) could elicit cAMP signalling even with the orthosteric site occupied. If the allosteric modulator produces response when orthosteric agonists fail → BLOCKED, not ABSENT. Note: this is a hypothetical probe — Cmpd-6 and related compounds are research tools, not clinically available.
- M2 muscarinic AAb as negative allosteric modulator (documented in POTS (Sunami, Sugaya, and Takahashi 2025)) → standard muscarinic agonist (pilocarpine) fails. But an allosteric agonist or positive allosteric modulator that binds away from the AAb site may recover function.
Caveat — β-arrestin-biased blockade. Some GPCR autoantibodies produce sustained biased signaling (β-arrestin over G-protein) and receptor internalization, making the receptor functionally absent for Gαs pathways while β-arrestin pathways remain active (Dodge-Kafka et al. 2026). This is a hybrid state: the receptor is structurally present but functionally absent for certain signaling modes. The probe in this case is a β-arrestin-biased ligand — if it works where a Gαs-biased ligand fails, chronic AAb-mediated desensitization is the mechanism. This biased-signalling variation is a sub-state of BLOCKED (the same AAb that blocks orthosteric access also biases the remaining signalling), not an independent subtype.
Evidence basis: Autoantibodies acting as orthosteric antagonists and allosteric modulators at GPCRs is well-established pharmacology (Skiba and Kruse 2021). GPCR autoantibodies are documented in POTS, Long COVID, and ME/CFS — though their existence in ME/CFS is contested ((Germain et al. 2025) null by REAP/Luminex; positive by CellTrend ELISA). No published study has tested an allosteric bypass probe in AAb-positive ME/CFS patients.
Certainty: 0.20. The probe logic is mechanistically sound. The existence of pathogenic orthosteric-site GPCR AAb in ME/CFS is the contested premise — if GPCR AAb in ME/CFS are non-pathogenic or measurement artefacts, this null subtype has no anchor in this disease.
The receptor is present, structurally intact, and capable of signaling — but it is outcompeted by a supraphysiological concentration of an endogenous antagonist. The drug fails not because the receptor is absent or blocked, but because the endogenous ligand occupies the site at concentrations the therapeutic dose cannot overcome.
Distinguishing probe — Concentration-competition. Escalate the drug dose or use a co-agonist that competes directly with the endogenous antagonist. For NMDA receptors with suspected KYNA elevation, the probe is glycine or D-serine (glycine-site co-agonists that compete with KYNA at the glycine modulatory site) rather than glutamate-site antagonists like ketamine (which acts at a different site and cannot test KYNA competition). If D-serine produces improvement → the glycine site was under-occupied, consistent with (but not uniquely diagnostic of) KYNA-mediated competition. If D-serine does nothing → the site is not responsive, the receptor may be absent or blocked at the glutamate site, or KYNA affinity exceeds the administered co-agonist dose. The probe identifies glycine-site NMDA dysfunction — not specifically KYNA elevation — and requires confirmatory measurement of KYNA and glycine levels for subtype attribution.
Example — KYNA (kynurenic acid) at NMDA receptors. KYNA is an endogenous NMDA antagonist that binds the glycine co-agonist site, not the glutamate site (Hazrati et al. 2024). The kynurenine pathway is dysregulated in ME/CFS (Kavyani et al. 2024) — though KYNA direction is inconsistent (elevated in some studies, depleted in others), meaning the overwhelmed null subtype only applies to patients with confirmed KYNA elevation. No published study has tested a concentration-competition probe in ME/CFS.
Caveat — confound. If glycine levels are also abnormal, the probe’s specificity degrades — D-serine may work because glycine was depleted (not because KYNA was competing). The probe requires confirmatory measurement of both KYNA and glycine levels for subtype attribution.
Evidence basis: KYNA as endogenous NMDA antagonist is well-established neuroscience (Schwarcz et al. 2012). Kynurenine pathway dysregulation in ME/CFS is documented (Kavyani et al. 2024), but KYNA direction is inconsistent — some studies show elevation, others depletion. The overwhelmed null subtype only applies to the elevated-KYNA patient subset.
Certainty: 0.15. The pharmacology is sound but depends on a specific biochemical state (KYNA elevation) whose prevalence in ME/CFS is unknown — and which may be the exception (depletion appears more common). The probe may apply to a small minority of patients.
2 Additional Dimension — Biased Signalling
Note: Biased signalling is NOT an independent null subtype — it is a special case of the BLOCKED state where the AAb drives biased signalling (β-arrestin over Gαs) rather than simple orthosteric blockade. It is documented here as an extension of the BLOCKED caveat (subtype 2, line 38), not as a fourth taxonomic category. The same AAb that blocks orthosteric access can simultaneously sustain β-arrestin signalling, making the receptor functionally absent for G-protein pathways while structurally present and actively signalling through another pathway. This distinction is pharmacologically interesting but currently impractical — biased ligands are research tools, and no clinical probe can distinguish BIASED from BLOCKED.
GPCRs signal through multiple pathways simultaneously — G-protein (Gαs, Gαi, Gαq) and β-arrestin — and these can be decoupled. Chronic AAb exposure produces sustained β-arrestin-biased signalling while Gαs-cAMP is desensitized (Dodge-Kafka et al. 2026). A drug that signals through Gαs (isoproterenol) produces null — the pathway is offline. But the receptor is structurally present, driven by AAb (satisfying the BLOCKED definition), and Gαs signalling is absent. The null is pathway-specific, not receptor-level.
Distinguishing probe — Biased ligand. Use a drug that selectively activates the opposite pathway. If a β-arrestin-biased ligand works when a Gαs-biased ligand fails, the receptor is signalling-mode-locked. This probe collapses to the BLOCKED allosteric probe in most cases — biased ligands often bind allosteric sites — and the diagnostic inference is the same: the receptor is present but the orthosteric agonist cannot access the relevant signalling pathway.
Certainty: 0.12. Biased GPCR signalling is well-established pharmacology. Biased ligands are research tools, not clinically available. The BIASED variation adds conceptual precision to the BLOCKED subtype rather than forming a distinct actionable category.
Consequence: The BIASED variation is unlikely to change clinical decisions in the near term. Its value is conceptual: it reminds researchers that “receptor signalling is intact” and “G-protein signalling is intact” are different claims, and a null response to a G-protein-biased drug does not rule out β-arrestin-mediated receptor function.
3 Diagnostic Implications and Integration
A null response under the existing binary framework means: “mechanism eliminated — this system is not rate-limiting.” Under the framework, the same null means one of three things depending on which probe resolves it (with biased signalling as a special case of BLOCKED):
- ABSENT → the mechanism IS eliminated — the target never existed. Move on.
- BLOCKED → the mechanism is NOT eliminated — the receptor is present but inaccessible. Treatment implication: if the block is confirmed, the mechanism can be accessed by removing the block rather than abandoned. Potential research-investigational approaches include immunoadsorption, plasmapheresis, or B-cell depletion for AAb-mediated blockade; or correcting the metabolic abnormality causing endogenous ligand accumulation. Note: B-cell depletion via rituximab failed a Phase III ME/CFS trial (n=151, (Fluge et al. 2019)), though the trial’s negative result may reflect long-lived plasma cell sanctuary — residual AAb production after CD20+ depletion — rather than refuting AAb-pathogenicity per se. These interventions are invasive, expensive, not guideline-supported for ME/CFS, and rest on the contested premise of pathogenic GPCR AAb. The null was a false elimination in principle; whether it can be corrected in practice is unproven.
- OVERWHELMED → the mechanism is NOT eliminated — the receptor works but needs a higher drug concentration or co-administration of an agent that lowers the endogenous antagonist. The null was a dosing failure, not a mechanism failure. This remains a research hypothesis — dose escalation in ME/CFS patients carries PEM risk and no validated protocol exists.
- BIASED (sub-state of BLOCKED) → the mechanism is NOT eliminated for all pathways — the receptor signals through one pathway but is desensitized for another. Treatment implication: exploit the intact pathway with a pathway-selective agonist. Biased ligands are research tools — no clinical action is currently possible.
The practical implication: when a drug fails, the null subtyping framework prevents premature elimination of mechanisms that are merely inaccessible rather than absent. A patient whose β2-AR is blocked by autoantibodies has a blocked receptor, not an eliminated adrenergic system — the framework would suggest investigating AAb removal before abandoning adrenergic mechanisms. A patient whose NMDA receptors are overwhelmed by elevated KYNA has saturated glycine-site occupancy, not absent glutamatergic signalling — the framework would suggest investigating KYNA-lowering strategies before concluding the system is offline. These suggestions are research hypotheses, not clinical recommendations — the probes are untested and the interventions carry risks without proven benefit in ME/CFS.
Evidence type: This is a logical extension of receptor pharmacology principles (Finlay, Duffull, and Glass 2020) to clinical pharmacodiagnostics. No empirical validation exists. The probes are proposed but untested in ME/CFS. The framework’s clinical value depends on the prevalence of BLOCKED and OVERWHELMED null subtypes in the ME/CFS population — which is currently unknown.
Consequence: The binary null inference — “drug didn’t work, mechanism eliminated” — may be systematically wrong for a subset of patients whose receptors are autoantibody-blocked or endogenously overwhelmed rather than absent. This subset is currently unmeasured and may be small, but identifying them would change treatment strategy from “abandon this mechanism class” to “restore access to this mechanism class.” The framework adds diagnostic precision at the cost of additional testing — not all patients need null subtyping, but patients with plausible AAb or metabolic profiles (elevated KYNA) may benefit from distinguishing which kind of null they have.
4 Relationship to Existing Frameworks
The null ladder (The Null Ladder - Why Null Is More Informative Than Positive) orders drug trials by coverage breadth — maximize mechanisms eliminated per trial. Null subtyping adds a second-pass analysis: after a drug fails, determine which kind of null it was before updating the differential. The null matrix (null matrix clinical) maps drugs to mechanisms; null subtyping adds a third dimension — null type (absent/blocked/overwhelmed) — so each cell is not binary but categorical.
The negative-control class (Protocol for Medication-Associated Permanent Decline) addresses a different logical layer: drugs whose mechanism targets a hypothesized causal pathway such that a null result constitutes mechanism-level falsification for the population. Null subtyping operates at the individual-patient level; negative-control inference operates at the population-mechanism level. When a negative-control drug fails at the individual-patient level, null subtyping determines whether the null means mechanism absent, blocked, or overwhelmed — preventing premature extension of population-level falsification to patients whose mechanism is merely inaccessible.
The origin-axis framework (What Pharmacodiagnostics Can and Cannot Rule Out: Origin Hypotheses) addresses a different question: what does a null response at one mechanism do to the probability of an origin category? Null subtyping operates one level below: it addresses what a null response means for the mechanism itself, independent of origin. The two frameworks are complementary: origin elimination translates mechanism-level nulls upward to origin constraints; null subtyping refines mechanism-level nulls inwardly into categorical types.
Limitation: Null subtyping requires additional testing (bypass agonist, allosteric ligand, dose escalation) that the standard null ladder does not. For most patients, binary null logic is sufficient — the marginal gain from distinguishing absent/blocked/overwhelmed may not justify the additional trial burden. The framework is proposed as a conceptual tool for interpreting ambiguous nulls, not as a mandatory addition to every pharmacodiagnostic protocol.
Consequence: For researchers designing pharmacodiagnostic protocols, null subtyping suggests that “drug failed” is underspecified — the protocol should distinguish whether the failure was due to target absence, target inaccessibility, target saturation, or pathway-level bias. For patients, the framework reframes treatment failures as potentially informative about how a mechanism is broken, not just whether it is broken.
5 Critical Limitations
The framework requires probes (bypass agonists, allosteric ligands, concentration-competition escalation) that are rarely available in clinical settings. For most ME/CFS-relevant drug targets — TRPM3, D2/D3, TLR4, opioid receptors, cholinergic receptors — no clinically available allosteric modulator, bypass agonist, or validated concentration-escalation protocol exists.
The gap is structural: the framework identifies a real pharmacological distinction (absent vs blocked vs overwhelmed) but provides no means to operationalize it for most of the targets that ME/CFS patients actually trial. A framework that is correct in principle but impossible to apply is not a clinical tool — it is a concept. This does not make the framework wrong, but it substantially limits its immediate utility.
Certainty: 0.30. The gap between available probes and needed probes is a factual observation, not speculation.
Consequence: For the foreseeable future, null subtyping is a research-design concept, not a clinical practice. Researchers designing pharmacodiagnostic trials should select targets for which distinguishing probes exist (e.g., β2-AR allosteric ligands, NMDA site-selective agonists) rather than retrofitting the framework onto targets where no probes are available. For patients and clinicians: the framework explains why nulls are ambiguous but does not yet resolve that ambiguity.
The null subtyping framework assumes a single disease-steady-state — that a receptor is either absent, blocked, or overwhelmed — and that this state is stable across time. Real pharmacology is messier:
- Mixed states: A receptor population is rarely uniform. Some receptors may be AAb-blocked while others are functional (partial blockade). Some are internalized (functionally absent) while surface receptors are blocked. The probe may detect the dominant state without revealing the mixture.
- Receptor reserve: Spare receptors mean near-maximal response at low fractional occupancy. A drug that appears to work may be activating only the unblocked fraction — the blocked fraction is masked by receptor reserve.
- State-dependence: Receptor populations cycle — diurnal expression rhythms, stress-induced trafficking, inflammatory cytokine-modulated expression. A drug that works in the morning may fail in the evening because receptor density changed, not because the receptor is absent/blocked/overwhelmed.
- Disease-progression confound: Some receptors may transition from overwhelmed → blocked → absent with disease duration (chronic AAb exposure → receptor internalization → eventual tissue-level loss). A single-timepoint measurement captures a potentially transitory state.
Certainty: 0.25. These are well-established pharmacological phenomena, not speculative novel hypotheses.
Consequence: A null subtyping result should be treated as provisional — valid for the measured receptor population at the measurement timepoint. It should not be treated as a permanent classification. Protocols should acknowledge that the state may change with time, treatment, and disease progression.
Null subtyping requires additional drug trials beyond the standard null ladder. After a drug fails, the framework proposes a second round of probing — allosteric ligand, bypass agonist, or dose escalation — to determine the null type. Each additional trial has costs: time, PEM risk, financial burden, emotional toll of another “failed” trial. The framework does not propose any PEM-risk mitigation for these additional trials (no pacing guidance, minimum washout intervals, or abort criteria).
The framework’s diagnostic yield is unknown. If 95% of nulls are true ABSENT and only 5% are BLOCKED or OVERWHELMED, null subtyping adds approximately 20 extra trials for every 1 informative result. The framework does not include a prior-probability estimate: which null types are most common in ME/CFS? How often does the probe actually change the treatment recommendation? Without these numbers, the framework advocates for additional testing without quantifying the expected benefit.
Certainty: 0.30. The burden-of-proof argument is logical, not empirical — the framework’s diagnostic yield is unknown because the framework has never been tested.
Consequence: Until prevalence data exists, null subtyping should be reserved for scenarios where the BLOCKED or OVERWHELMED prior is high — e.g., known GPCR AAb positivity, confirmed elevated KYNA — rather than applied indiscriminately to every null response. The framework’s clinical value scales with the prevalence of non-ABSENT nulls, which is currently a completely unknown parameter.
6 Supplemental and Low-Cost Probes
One null subtype — ABSENT — can sometimes be predicted before a drug trial begins, though its prevalence relative to the other subtypes in ME/CFS is unknown (no systematic data exist). Pharmacogenetic variants that reduce receptor expression or function may pre-identify patients for whom a drug targeting that receptor will produce null — regardless of disease state.
Receptor-level pharmacogenetic screening for ME/CFS-relevant drug targets could pre-assess the likelihood of ABSENT null for a subset of targets without requiring a drug trial:
- D2 receptor variants (Taq1A, DRD2 C957T) → reduced striatal D2 expression; may produce functionally absent D2 signalling
- OPRM1 A118G → altered μ-opioid receptor signalling and LDN response variability (partial loss, not ablation — provides prior probability, not definitive assignment)
- β2-AR polymorphisms (Gly16Arg, Gln27Glu) → altered receptor expression and desensitization kinetics
Note on CYP2D6: Poor-metabolizer status is a pharmacokinetic null (drug never reaches target), not a receptor ABSENT null — it belongs to the existing null-failure-modes in The Null Ladder - Why Null Is More Informative Than Positive (dose-access failure), not to null subtyping. CYP2D6 genotyping is separately actionable today through standard pharmacogenetic testing with CPIC clinical guidelines.
If a patient carries a known loss-of-function variant, that receptor’s ABSENT null subtype is established before a single pill is taken — saving time, PEM risk, and emotional cost. For variants with partial loss (e.g., OPRM1 A118G reduces μ-opioid signalling), the pre-screening provides a prior probability of ABSENT rather than a definitive assignment — the drug trial is still needed to confirm but the expectation of null is higher.
Evidence basis: Pharmacogenetic prediction of drug response is well-established in oncology and psychiatry, but never applied to ME/CFS pharmacodiagnostics. The variants exist; the structured application to pre-trial null subtyping is novel.
Certainty: 0.20. The variant-disease associations are established in general populations. ME/CFS-specific variant frequencies are unknown. No ME/CFS study has used pharmacogenetic pre-screening to guide pharmacodiagnostic protocols.
Falsifiable prediction: In a prospective ME/CFS cohort (N ≥ 100), genotyping for receptor-level loss-of-function variants (OPRM1 A118G, D2 Taq1A/C957T, β2-AR Gly16Arg/Gln27Glu) before drug trialling will produce significantly fewer null-result drug trials for drugs targeting those specific receptors, compared to drugs targeting non-genotyped receptors within the same patients. If receptor variants show zero correlation with trial outcomes, the ABSENT pre-screening claim is falsified. Note: CYP2D6 is excluded — it is a pharmacokinetic variant handled separately.
Consequence: A pragmatic implementation would be a ~20-variant pharmacogenetic panel covering the most commonly trialled ME/CFS drug targets with receptor-level loss-of-function markers. Receptor-variant panels require validation; the individual components are well-studied in population genetics, and CYP2D6 and other PK-relevant variants are separately actionable through existing CPIC-guideline-supported testing orderable today.