Reelin-ApoE receptor axis disruption

Description

Reelin is the large secreted glycoprotein that instructs migrating neurons where to stop during development, laying down the layered cortex and hippocampal formation whose orderly ruin is the anatomy of dementia. It does not retire when construction ends: in the adult brain it signals through two lipoprotein receptors — ApoER2 and VLDLR, the same receptors that bind apolipoprotein E — to the cytoplasmic adaptor Disabled-1 (Dab1), and from there it restrains tau phosphorylation (by suppressing GSK-3β), sharpens NMDA-receptor function, and enhances the long-term potentiation on which memory is built. In the adult, the architect becomes a guardian.

Which cells make it — and why the answer matters. Reelin has two source populations separated by a lifetime. In development it is secreted by the transient Cajal-Retzius neurons of the cortical surface, the positioning signal whose loss inverts the layers in the reeler mouse (D'Arcangelo 1995). In the adult brain, after those cells are gone, two populations carry it. (1) A defined subset of GABAergic interneurons — the reelin-positive cells are the somatostatin- and NPY-expressing ones, and never the parvalbumin-expressing ones; morphologically the layer I horizontal cells, the layer II–V bitufted cells, and the deep-layer Martinotti cells. They secrete reelin into the perineuronal net, where it acts extrasynaptically, and Dab1 — the receiving adaptor — sits predominantly in pyramidal neurons (Pesold 1998, 1999). (2) The glutamatergic projection neurons of entorhinal layer II — the stellate cells of medial and fan cells of lateral entorhinal cortex, the origin of the perforant path. Chin's paper calls these "pyramidal cells" following its own laboratory's usage; the entorhinal-anatomy convention distinguishes them from the calbindin-positive layer II pyramids that project to CA1 (the two namings are reconciled corpus-wide as C-021 in research/CORRECTIONS.md).

The identity is load-bearing, not taxonomic. The source cells are the unnetted ones and the recipient cells are the netted ones, and every population at the front of the disease — somatostatin interneuron, entorhinal layer II, locus coeruleus — sits on the unnetted side, two of the three being reelin sources. The cells that fail first supply the brake to everybody else and keep none of it for themselves (developed in The Crowded Cleft, research/crowded-cleft/). And the entorhinal source is doubly implicated: reelin-expressing entorhinal neurons are selectively depleted in AD while the reelin-expressing GABAergic interneurons are spared (Chin 2007), and the reelin-immunoreactive layer II neurons are the population in which intracellular amyloid-β first appears — at the pre-plaque stage in a transgenic rat, and in human tissue with early AD-related change (Kobro-Flatmoen 2016).

Verdict: guardian, not trigger. The corpus frames reelin as an axis of resilience — a homeostatic guardian of the synapse and of tau whose failing signal permits the disease, and whose reinforcement (in the one human experiment available) held the disease at bay — rather than as an initiating cause. The strong "reelin loss initiates sporadic AD" claim is rejected as stated: reelin's decline is partly downstream of amyloid, and its bulk protein level rises rather than falls.

The ligand-versus-signal distinction (reelin resistance). A naïve "reelin goes down" model is false. Total reelin protein and its fragments increase ~40% in the AD cortex and CSF, even as reelin-dependent Dab1 phosphorylation — the actual readout of a working signal — falls. Amyloid-β both induces reelin expression and traps the secreted protein in aggregates, so the tissue ends up with more ligand and less signal: a state precisely analogous to insulin resistance, best named reelin resistance. The rising-reelin biomarker studies and the failing-signal mechanistic studies are the same pathology seen from two instruments.

The two human resilience cases and the heparan-sulfate convergence. The pathway's causal weight rests on two carriers of the fully-penetrant PSEN1-E280A (Paisa) mutation who each escaped dementia for three decades with heavy amyloid but spared entorhinal tau: the APOE3-Christchurch homozygote (R136S) and the RELN-COLBOS heterozygote (H3447R, a gain-of-function variant that activates Dab1 more strongly and lowers human tau phosphorylation in a knock-in mouse). These are not separate curiosities filed under "ApoE" and "reelin": both variants dial the same heparan-sulfate-dependent lipoprotein-receptor node — Christchurch loosening ApoE's harmful grip on heparan sulfate, COLBOS tightening reelin's protective grip on the same required co-receptor. This convergence is why reelin belongs at the centre of the resilience conversation, and it reframes amyloid: in both brains amyloid ran unchecked while the protective variants stopped the tau downstream. The same ApoE–heparan-sulfate bond is the subject of Quintero's glycosaminoglycan work, which reads AD as a disorder of aberrant matrix sulfation and names ApoE–heparin binding as the controllable step — ApoE4 binds heparan sulfate most tightly and Christchurch least — so that loosening the bond is itself protective.

The receptor-side reading. The guardian signal can also be lost from the receptor end. On the lipid-peroxidation model (Ramsden), peroxidized ApoE4-borne PUFA cargo generates reactive aldehydes that pyrrole-crosslink the ApoE–ApoER2 complex, arresting receptor recycling and abrogating the Reelin–ApoER2–Dab1 cascade — the same guardian signal, silenced from the receptor rather than the ligand. Because ApoER2 is most densely expressed in entorhinal layer II and the locus coeruleus, the first regions to tangle, the model doubles as an account of selective vulnerability, and it is the mechanistic mirror of the resilience genetics.

The perineuronal compartment — where the signal is staged. Reelin is secreted into the perineuronal net by a subset of GABAergic interneurons (Pesold 1998, 1999), and its signal requires N-sulfated heparan sulfate as an obligate co-receptor (Pan 2025). The polymer distinction matters: the net is built from chondroitin sulfate, while reelin's co-receptor and tau's route of entry are both heparan sulfate — a general constituent of the neuronal surface, not a net component (Fawcett 2022; Holmes 2013). So (a) staging the reelin tau-brake and (b) gating tau uptake are one chemistry, and the net's shield is a different chemistry beside it. Reelin does not require a net in order to signal. Net-bearing neurons carry low tau (Morawski 2010; de Vries 2024), which on this reading is co-location rather than identity. Reelin's resilience axis and the PNN axis are therefore two functions of one guarded compartment — developed in full in the monograph The Architect's Scaffold (research/reelin-pnn/). Honest seam: reelin-secreting, net-bearing, and reelin-responding cells are often distinct, so the unity is of the matrix compartment, not a single cell.

Convergence Nodes

• APOE4 Hub — reelin and ApoE compete for ApoER2/VLDLR; ApoE4 is predicted to degrade the reelin signal at the shared receptor (graded plausible, the dedicated human demonstration is still lacking)
• Endosomal Nexus — ApoER2 recycling and trafficking; pyrrole-crosslinking of the ApoE–ApoER2 complex traps the receptor and abrogates reelin signaling
• Perineuronal Nets — the sulfated compartment into which reelin is secreted; its chondroitin-sulfate net shields the neuron, while the heparan sulfate beside it stages the reelin signal and gates tau uptake (the shared sulfation node — two polymers, one place)

Prize Entrants

• Ramsden — the receptor-side reading: peroxidized-ApoE aldehyde crosslinking of the ApoE–ApoER2 interface arrests receptor recycling and abrogates the Reelin–ApoER2–Dab1 cascade
• Quintero — the heparan-sulfate reading: aberrant glycosaminoglycan sulfation and ApoE–heparin binding govern the Aβ-to-tau step; the APOE3-Christchurch variant's reduced ApoE–heparin affinity is protective

External Scientists

• Yakeel Quiroz — co-author on both resilience reports (APOE3-Christchurch, 2019; Reelin-COLBOS, 2023) in the Colombian E280A kindred
• Joachim Herz (not yet a KB node) — worked out reelin's receptor biology (VLDLR/ApoER2 → Dab1) and its "synaptic guardian" framing of amyloid-β
• Eduardo Soriano (not yet a KB node) — reelin's amyloid antagonism and delay of fibril formation

Key Open Questions

• Does entorhinal reelin loss precede or follow local tau onset? (cross-sectional data are causally ambiguous)
• Does ApoE4 promote AD in part by degrading reelin signaling at the shared receptor? (mechanistically coherent, human demonstration lacking)
• Can a reelin-pathway agonist, active fragment, or heparan-sulfate modulator reproduce the COLBOS entorhinal-tau sparing — and must it be given in the early (preclinical) window before the amyloid ceiling is breached?
• What is the right biomarker: reelin signalling (Dab1 / receptor-cleavage fragments), not total reelin, which rises misleadingly?
• Does the loss of the interneuron source population measurably lower reelin availability in the cortex it innervates — and does that fall precede the degradation of the recipient's net? (the source/recipient map predicts source-first; unmeasured)

See Also

• Perineuronal Nets — the sulfated-matrix twin of this axis
• Monograph: The Architect's Reprieve — Reelin in Alzheimer's Disease (research/reelin-alzheimers/) — the resilience case, graded in a validity ledger
• Monograph: The Architect's Scaffold — Reelin and the Perineuronal Net (research/reelin-pnn/) — the shared sulfation node uniting reelin's tau-brake with the PNN