The reelin protein plays a starring role in neural development, and new findings suggest the neuromodulator does an important job in older animals, too. Researchers led by Joachim Herz at the University of Texas Southwestern Medical Center in Dallas generated mice that lack reelin in the brain when they are adults. As reported July 7 in Science Signaling, these mice appeared normal—unless crossed with a mouse strain that overproduces Aβ. In the face of amyloid toxicity, reelin-deficient mice had accelerated learning and memory problems. The findings suggest that reelin protects neurons from the scourge of AD pathology. Because the AD risk factor ApoE4 thwarts reelin signaling, the study also supports the idea that disease risk and severity in ApoE4 carriers relate to poor reelin signaling.

“In this elegant set of experiments, Herz and colleagues have continued to shed light on the role of reelin in synaptic function and provide further evidence for the protective nature of reelin against amyloid-associated synaptic plasticity and memory dysfunction,” commented Edwin Weeber of the University of South Florida in Tampa.

Reelin is a large secreted protein first discovered for crafting central nervous system (CNS) architecture during development. “Reeler” mice, which lack the reelin gene, suffer a slew of neurodevelopmental problems, including inverted cortical layering and wonky hippocampal and cerebellar structures (see D’Arcangelo et al., 1995). They have severe movement problems and die young. Animals expressing only a single copy of reelin have learning and memory problems despite seemingly normal brain structure (see Tueting et al., 1999; Qiu et al., 2006).

In adults, reelin plays a role in synaptic signaling and plasticity by latching onto Apoer2, a receptor that binds none other than ApoE. Herz and colleagues previously reported that reelin’s engagement of Apoer2 activates the cytosolic receptor disabled-1 (Dab1), leading to tau phosphorylation, as well as the phosphorylation of the NR2 subunit of NMDA receptors (see Hiesberger et al., 1999Aug 2005 news). Phosphorylated NR2 elevates NMDAR levels at the synaptic surface and boosts calcium influx when these receptors are turned on, raising long-term potentiation (LTP), a hallmark of synaptic plasticity. On the other hand, ApoE4 (but not ApoE2 or E3) promotes cell uptake of Apoer2, limiting reelin effects, including LTP (see Weeber et al., 2002May 2012 conference news). Aβ also takes a toll on synaptic plasticity, and some studies suggest it may block Apoer2 function as well.

It has been difficult for researchers to tease out reelin’s synaptic effects from its developmental ones, and thus to home in on its potential role in AD. First author Courtney Lane-Donovan and colleagues addressed this issue by generating conditional knockout (cKO) mice that lose reelin expression in the brain after treatment with tamoxifen. The researchers allowed the mice to develop normally, and then turned off reelin expression. In all experiments they treated the animals with tamoxifen when they reached 2 months of age.

One month after switching off reelin, the researchers noticed that the mice had more Dab1 in their brains than did controls. This made sense because reelin is known to phosphorylate Dab1 and target it for degradation. Surprisingly, this elevation of Dab1 did not result in any of the downstream changes previously seen in reelin heterozygous or knockout mice: Glutamate receptor levels were normal, as was the phosphorylation of several key signaling molecules, including tau. CNS architecture did not change. The researchers attributed the more drastic effects of germline reelin removal to developmental deficits that did not occur in the cKO mice.

Despite their normal glutamate receptor levels, hippocampal slices taken from cKO mice at approximately 7 months of age displayed elevated LTP. Oddly, previous studies indicated that mice deficient in reelin had a reduction in LTP from the get-go, while injecting reelin directly into the brain boosted it (see Weeber et al., 2002). Herz speculated that the elevated LTP in the reelin cKO mice represented a compensatory mechanism unique to the loss of reelin in adulthood.

Turning reelin off for one month had little behavioral effect. Compared to untreated mice, the 3-month-old tamoxifen-fed cKO mice had a slight reduction in anxiety: They were less apt to shy away from open spaces in one test, yet displayed a normal propensity to explore the exposed arms of a maze. The treated mice learned the location of submerged platforms in a water maze just as well as untreated mice, and responded by freezing in place when put into a context similar to one in which they had previously received a shock. The researchers concluded that unlike mice deficient in reelin throughout development, loss of reelin in adulthood had few to no behavioral or learning consequences.

The authors next tested the effects of reelin loss in adult mice that overexpress mutant human amyloid precursor protein. They crossed the reelin cKO mice to Tg2576 mice, which start to accumulate Aβ in the brain at around 4 months of age but don’t develop plaques until around 11 months of age. The researchers found no significant differences in the concentrations of soluble or insoluble Aβ between Tg2576 mice with or without reelin expression at 7 months, when both strains had elevated Aβ in the brain compared to wild-type mice. Though neither strain harbored detectable plaques in the cortex, the tamoxifen-treated Tg2576/reelin cKO mice performed poorly in the Morris water maze, while the Tg2576 mice still performed just as well as normal mice. Seven-month-old reelin cKO mice also performed normally on the test, indicating that the combination of mutant APP overexpression and reelin deficiency caused learning difficulties. Hippocampal slices taken from the Tg2576/reelin cKO mice at approximately 7 months of age displayed none of the LTP boost that reelin cKO mice did.

Some commentators pointed out that it is unclear whether Aβ itself caused the effects seen in Tg2576/reelin cKO mice. “Without a reversal by Aβ immunization, for example, it is not possible to conclude that the memory deficits seen here reflect a cooperative effect of reelin loss and Aβ,” wrote Steven Barger of the University of Arkansas in Little Rock. Barger added that APP has been reported to reduce Dab1 signaling, so elevated levels of APP in the Tg2576 mice could have effected reelin signaling independently of Aβ. In the paper, the authors acknowledged that Dab1 could indeed be doused by APP or other binding partners. However, Herz told Alzforum that preliminary results from his lab indicate that removal of Aβ abolishes the early learning and memory problems in the Tg2576/reelin cKO mice.

Herz hypothesized that to make up for lack of reelin signaling in the synapse, neurons produce more AMPA receptors, which would explain the elevated LTP in reelin cKO mice. His lab is investigating this. While additional AMPA receptors may maintain synaptic signaling when everything is running smoothly, Herz proposed that Aβ oligomers, which have been reported to appear by 7 months of age in Tg2576 mice, derail the system by reducing glutamate receptors at the synapse (see Westerman et al., 2002). 

Interestingly, ApoE4 knock-in mice, which respond sluggishly to reelin due to reduced Apoer2 surface expression, also have elevated LTP (see Korwek et al., 2009). This compensatory mechanism could explain why ApoE4 carriers are cognitively normal at first, Herz said, but once Aβ comes into play, the compensation is rapidly thrown off.

Herz also speculated that boosting reelin signaling could protect against cognitive decline in the face of AD pathology in some people. One genome-wide association study reported that certain variants of reelin were enriched in cognitively normal people with a high burden of AD pathology, hinting that those variants may be protective. Similarly, some normal older people with a high level of neurofibrillary tangles had high reelin levels (see Kramer et al., 2011). Herz speculated that protective variants might increase reelin levels.—Jessica Shugart


  1. Joachim Herz and his lab have accomplished a milestone in research on the biology of reelin, a protein that has complex biochemistry and genetic contributions to brain development, as well as synaptic plasticity in the adult. Because it is so critical for the correct patterning of cortical layers during neural development, reelin’s roles at adult synapses have not been amenable to investigation via germline-knockout animals. Lane-Donovan et al. now report the results of a conditional knockout (cKO) that ablates reelin only in the forebrain and only after the cortical layers have been established. Their results are very important, and not without a few surprises.

    Reelin is perhaps most relevant to Alzheimer’s disease because its most important effects are mediated by a pair of lipoprotein receptors that are also binding sites for ApoE.  Although several members of the LDL-receptor family have biological effects that indicate roles other than lipid transport, apoER2 and VLDLR have very acute, well-characterized impacts on signal transduction via an intracellular accessory protein called Dab1. Agonism of apoER2 or VLDLR, apparently via the formation of dimers or other multimers induced by multimeric ligands, leads to tyrosine phosphorylation of Dab1, and this propagates signals to important post-synaptic proteins. A key downstream event is accentuation of NMDA receptor responses, thus the reelin→apoER2→Dab1 pathway can make important contributions to long-term potentiation (LTP) and other types of synaptic plasticity dependent on NMDA receptors. One might imagine that interruptions in this pathway could therefore underlie the memory deficits apparent in dementia.

    Prior studies have indicated that ApoE opposes the actions of reelin in some respects. Dr. Herz has contributed to that body of work with demonstrations that the antagonism involves ApoE binding to apoER2 and evoking ligand-stimulated receptor internalization, temporarily taking the apoER2 “out of commission” regarding reelin signaling. ApoE4 was reported to hold the receptor inside the cell longer than did ApoE3; the latter allowed quicker recycling of the receptor to the cell surface. Thus, one potential mechanism by which inheritance of an ε4 allele of ApoE fosters Alzheimer’s disease is by inhibiting reelin’s ability to enhance LTP and, thus, memory. Because Aβ has been demonstrated to bind other LDL-receptor family members, it is also conceivable that Aβ could antagonize reelin actions.

    Together, these prior studies on reelin—conducted either in reelin heterozygotes or with anti-reelin antibodies—made some of the results obtained here by Lane-Donovan et al. surprising. The reelin-cKO mice they generated reportedly had no deficits in memory and actually showed enhancements in their capacity for LTP. Given the diminutions of LTP when reelin has been lowered or inhibited in other paradigms, this is counterintuitive. Perhaps the complete (vis à vis the heterozygote paradigm) and subchronic (vis à vis acute application of an antibody) removal of reelin results in overcompensation by some reactive or redundant mechanism. 

    Another of the surprises among the reelin cKO results is that this manipulation did not lead to dispersion of the dentate granule cells of the hippocampal formation. Adult neurogenesis produces continuous renewal of dentate granule cells from the subgranular zone. Healthy hippocampal circuitry requires proper migration of the newborn cells to the narrow granule cell layer, and prior studies have found that disruption of any member of the reelin→apoER2/VLDLR→Dab1 pathway leads to inappropriate migration. The aberrant migrations cause a disruption of normal circuitry and produce epileptiform activity. Indeed, temporal-lobe seizures—either experimentally induced or those that naturally occur in human patients—are correlated with diminished levels of reelin. The emerging recognition of subtle epileptiform activity in AD thus makes reelin’s role in these events another point of intrigue for AD researchers. However, Lane-Donovan et al. report that the removal of reelin in their mice for a month had no adverse effects on dentate organization. The authors discuss this in relation to one other experimental paradigm that implicated reelin in granule cell dispersion, and they suggest that their results may indicate a role for other reelin receptors that exert a “dominant-negative interference” when only apoER2/VLDLR-mediated events are lost. But this seems a bit of a challenge to Occam’s Razor. It seems more likely that whatever compensatory response was responsible for the reversal of the expected LTP drop also created compensation regarding granule cell migration.

    The primary take-home message Lane-Donovan et al. would leave us with is that loss of reelin renders enhanced vulnerability to the Aβ-dependent behavioral deficits seen in APPsw (Tg2576) mice. This line typically performs well in the Morris water maze until 9 to 10 months of age, but in combination with cKO of reelin, deficits were apparent in mice at just 7 months of age; this, despite reelin cKO having no detected effects on accumulation of soluble or insoluble Aβ. These results are subject to a considerable caveat, however: As Lane-Donovan et al. remind us, Dab1 can be modulated by several other membrane proteins, most notably for readers here, APP itself. Evidence suggests that APP can act as a decoy to remove Dab1 from the transduction of reelin signaling. The APP transgene is expressed at five- to 10-fold that of normal endogenous levels in Tg2576 mice, and it may therefore exert the sort of “dominant-negative interference” Lane-Donovan et al. invoke for other phenomena. Without a reversal by Aβ immunization or the like, it is not possible to conclude definitively that the memory deficits seen here reflect a cooperative effect of reelin loss and Aβ production. 

    Nevertheless, it certainly is possible that an Aβ-dependent effect is at work here. It might be worth considering that the Aβ present in Tg2576 mice differs from that present when reelin is simply removed from the cKO mice. The authors point out that such a removal is less prone to artifacts than the complementary overexpression studies, where reelin may be expressed at inappropriate sites and certainly at supraphysiological levels. However, the one thing that such a simple KO (even the conditional sort) cannot tell us is how reelin might functionally interact with human Aβ. The sequence of the mouse peptide is famously resistant to aggregation and may not be capable of exerting pathogenic actions when relieved of a reelin-mediated antagonism. If reelin has no dramatically important roles in adults other than Aβ antagonism—which Lane-Donovan et al. would seem to suggest—knocking it down or out in mice will not be particularly informative unless the Aβ is humanized. So, as with so many other fascinating questions in the AD field, the best data may yet come from studies performed in the context of an APP knock-in mouse, such as those created by Cephalon researchers, by Christoph Köhler, or by Takaomi Saido.

  2. In an elegant set of experiments, Herz and colleagues have continued to shed light on the role of reelin in synaptic function and provide further evidence for the protective nature of reelin against amyloid-associated synaptic plasticity and memory dysfunction using a novel mouse model; the inducible reelin knockout (reelin cKO) mouse. The reelin cKO model reveals that adult knockdown of reelin expression results in no discernable differences in learning and memory, but in contrast causes enhanced hippocampal late-phase LTP. These mice show an increase in the expression of the intracellular adaptor protein disabled-1, without significant alterations in other downstream effectors, glutamate receptors, or tau phosphorylation. They also show normal cerebellar, hippocampal, and cortical lamination and development and no differences in dendritic spine morphology. These findings set the stage for a cross of the reelin cKO with the Tg2576 (APP overexpressing) AD mouse model. Interestingly, these crosses did not show increased Aβ pathology; however, they were more susceptible to the detrimental effects of Aβ on spatial learning and memory and the enhanced late-phase LTP phenotype was lost. These results suggest that endogenous reelin can be protective against memory loss associated with AD pathology and raises the interesting possibility that reduced reelin expression in the early stages of AD may underlie the initial symptoms associated with mild cognitive impairment. Furthermore, in the normal aged brain, reduced reelin expression may not be as detrimental to cognitive ability as what was predicted from earlier mouse models.

  3. Loss of reelin in embryonic development leads to profound consequences in brain development (such as disrupted cortical layering), making it hard to delineate what impact reelin deficiency has in the adult brain. Therefore, the reelin conditional knockout (cKO) generated by the authors is an important tool for researchers who want to study reelin’s impact on adult function.

    The authors demonstrate that adult reelin cKO mice have normal brain structure, normal cognitive function, and a moderate enhancement of long-term potentiation (LTP) compared to controls. After characterizing this animal model, the authors crossed the cKO mice with a Tg2576 AD model. Although loss of reelin did not accelerate Ab plaque load at early stages in the Tg2576 background, it accelerated cognitive deficits. This study complements a study that reported that reelin supplementation promotes cognitive and synaptic function in wild-type mice (Rogers et al., 2011). Taken together, these findings strengthen the case for reelin being a viable therapeutic target to address the cognitive impairments observed in AD. It would be interesting to address whether loss of reelin in human AD patients correlates with cognitive deficits independently of plaque load.


    . Reelin supplementation enhances cognitive ability, synaptic plasticity, and dendritic spine density. Learn Mem. 2011 Sep;18(9):558-64. Print 2011 Sep PubMed.

Make a Comment

To make a comment you must login or register.


Alzpedia Citations

  1. APOE

News Citations

  1. Are You Reelin in the Years? Not without Alternative Splicing
  2. Keystone: ApoE Receptors and Ligands in Memory and AD

Research Models Citations

  1. Tg2576

Paper Citations

  1. . A protein related to extracellular matrix proteins deleted in the mouse mutant reeler. Nature. 1995 Apr 20;374(6524):719-23. PubMed.
  2. . The phenotypic characteristics of heterozygous reeler mouse. Neuroreport. 1999 Apr 26;10(6):1329-34. PubMed.
  3. . Cognitive disruption and altered hippocampus synaptic function in Reelin haploinsufficient mice. Neurobiol Learn Mem. 2006 May;85(3):228-42. Epub 2005 Dec 20 PubMed.
  4. . Direct binding of Reelin to VLDL receptor and ApoE receptor 2 induces tyrosine phosphorylation of disabled-1 and modulates tau phosphorylation. Neuron. 1999 Oct;24(2):481-9. PubMed.
  5. . Reelin and ApoE receptors cooperate to enhance hippocampal synaptic plasticity and learning. J Biol Chem. 2002 Oct 18;277(42):39944-52. PubMed.
  6. . The relationship between Abeta and memory in the Tg2576 mouse model of Alzheimer's disease. J Neurosci. 2002 Mar 1;22(5):1858-67. PubMed.
  7. . ApoE isoform-dependent changes in hippocampal synaptic function. Mol Neurodegener. 2009 May 27;4:21. PubMed.
  8. . Alzheimer disease pathology in cognitively healthy elderly: a genome-wide study. Neurobiol Aging. 2011 Dec;32(12):2113-22. PubMed.

External Citations

  1. Weeber et al., 2002

Further Reading

Primary Papers

  1. . Reelin protects against amyloid β toxicity in vivo. Sci Signal. 2015 Jul 7;8(384):ra67. PubMed.