In Alzheimer’s disease, seizures in the hippocampus contribute to cognitive decline. Scientists have known for some time that this correlates with persistent loss of the calcium-binding protein calbindin-D28k, a regulator of synaptic plasticity, learning, and memory. Now, researchers report that, in mouse models of AD and epilepsy, accumulation of the long-lived transcription factor ΔFosB suppresses calbindin-D28k expression by altering the structure of its promoter. In the October 16 Nature Medicine, Jeannie Chin and colleagues at Baylor College of Medicine in Houston also describe how interfering with ΔFosB activity, or restoring calbindin-D28k expression, improves the mice’s performance in a test of spatial memory. In autopsy samples from people with AD or MCI, ΔFosB levels were high when calbindin-D28k was low, and the changes correlated with scores on the Mini Mental State Exam. The results reveal a missing link between seizures and cognitive decline, and a possible explanation for how infrequent bursts of hippocampal hyperactivity can cause stable changes in gene expression and cognition.

  • Seizures induce the ΔFosB transcription factor, leading to memory deficits.
  • ΔFosB causes suppression of the calbindin-D28k gene.
  • Calbindin-D28k plays a role in learning and memory, and is down in epilepsy and AD.

“This paper details a mechanism that relates seizures to Alzheimer's disease, and strengthens the hypothesis that these two disorders are intimately related,” said Andrew Cole, Massachusetts General Hospital in Boston.

“There’s been a long-standing interest in the calbindin-D28k story in both the epilepsy world and the Alzheimer’s world. The demonstration that ΔFosB can regulate calbindin-D28k levels seems solid and very interesting,” Cole told Alzforum.

Jeffrey Noebels, who also works at Baylor, was excited by the paper, calling ΔFosB a missing link. “This is a welcome discovery revealing an epigenetic target for protection against seizure-induced malfunction of hippocampal circuitry, even in the face of other structural cellular pathology,” he wrote (see full comment below). 

Seizures in people with AD have long been overlooked. They can start early in the disease, and have few outward signs, but are a harbinger of cognitive decline (Vossel et al., 2016Jul 2013 news). A decade ago, Lennart Mucke and colleagues at University of California, San Francisco, discovered that mouse models of AD also manifest hippocampal seizures at a young age, long before the appearance of amyloid plaques or neuronal loss (Sep 2007 news). The group later blamed Aβ oligomers for the hippocampal hyperexcitablity (May 2012 news), and treated mice with the anticonvulsant levetiracetam to improve their memory deficits (Aug 2012 news). About the same time, researchers led by Michela Gallagher at Johns Hopkins University, Baltimore, reported that the drug reduces hyperactivity in the hippocampus and improves cognition in people with amnestic mild cognitive impairment (May 2012 news). A Phase 3 trial of a reformulated version of levetiracetam, called AGB-101, will begin in patients with MCI in early 2018 (Mar 2015 news, and see company website). Despite the growing interest in the field, how seizures disrupt cognition remains a mystery.

One clue came in a report from the Chin lab earlier this year. They found that seizure activity in mice induced the transcription factor ΔFosB, a long-lived truncated version of FosB that has been extensively studied for its role in epigenetic gene regulation in the nucleus accumbens in response to drugs of addiction (Corbett et al., 2017). This protein, with a half-life of eight days, mediated a long-term downregulation of the c-fos gene after seizures in AD mice, they showed. That work indicated that ΔFosB could contribute to persistent changes in gene expression even during seizure-free periods, but what were its targets?

In the new study, first author Jason You focused on a prime suspect in the hippocampus, calbindin-D28k. This protein, the product of the CALB1 gene, has been used as a marker for cognitive decline. When calbindin-D28k levels go down, whether from aging, AD, or epilepsy, cognitive function wanes. The degree of calbindin-D28k loss matches the severity of cognitive defects in mice and in people (Jul 2003 newsKarádi et al., 2012). 

Ups and Downs.

Highly expressed in the hippocampi of normal non-transgenic (NTG) mice (left), calbindin levels plummet in APP mice (bottom). For ΔFosB (right), the opposite is true. [Courtesy of You et al., Nature Medicine 2017.]

You and colleagues began by recording brain activity in J20 AD mice, using implanted electrodes. At two to four months of age, the mice don’t have plaques yet, but they do have memory deficits, and some showed spontaneous recurrent seizures, consistent with what Mucke and colleagues saw. The investigators found that, as expected, calbindin-D28k levels dipped in mice with seizure activity. At the same time, ΔFosB increased. Mice with the highest levels of ΔFosB had the lowest calbindin-D28k.

Does ΔFosB directly regulate calbindin-D28K expression? Chromatin immunoprecipitation indicated more ΔFosB bound to the calbindin gene promoter in J20 mice than in non-transgenics. Once at the promoter, ΔFosB triggered chromatin modifications linked to gene repression, including a decrease in acetylation of histone H4 and an increase in its methylation. These changes were not unique to AD mice—induction of seizures in normal mice by administration of pilocarpine led to similar outcomes. 

To prove ΔFosB curbs calbindin-D28k, the investigators forced expression of the transcription factor in the dentate gyri of wild-type mice using an adenovirus vector. Overexpressed ΔFosB bound to the calbindin-D28k promoter, decreased H4 acetylation, halved calbindin-D28k mRNA, and cut its protein level by a third.

Fos Fluctuations.

As calbindin (red) in normal mice (top) fades from dentate gyrus granule cells in APP mice with mild and severe seizures, ΔFosB (green) accumulates. [Courtesy of You et al., Nature Medicine 2017.]

The decrease in dentate gyrus calbindin-D28k undermined the animals’ spatial memory. Both J20 and wild-type mice expressing adenovirus ΔFosB became more forgetful than controls in an object-location test where they had to remember which of two flasks were moved in an experimental enclosure. On the other hand, blocking ΔFosB by expressing a dominant-negative partner protein reversed this memory defect. For direct proof that calbindin-D28k affects memory, the investigators expressed the protein in the J20 mice, and found that the animals better recognized the shifted flask. 

Does the same pathway operate in people? The researchers looked at postmortem tissue from six healthy controls, 10 people diagnosed with MCI, and 10 with AD. On average, ΔFosB was threefold higher and calbindin-D28k about 50 percent lower in MCI and AD brains compared to control tissue. Across all subjects, levels of calbindin-D28k negatively correlated with ΔFosB. In addition, declining MMSE scores in people with MCI tracked with higher ΔFosB and lower calbindin-D28k, though this relationship broke down in AD patients. Overall, the yin-yang of ΔFosB/calbindin-D28k seems to be a general feature of epileptic activity, as it occurs in resected tissue from temporal lobe epilepsy patients, as well.

The authors suggest that dampening down ΔFosB, or boosting calbindin-D28k, could help cognition in people with AD, and even epilepsy. At the same time, they acknowledge that ΔFosB induction and calbindin-D28k reduction in granule cells also appear to protect neurons against excitotoxicity (Mattson et al., 1991Nagerl et al., 2000; Lopez-Meraz et al., 2010). Chin pointed out that they saw no adverse effects from inhibiting ΔFosB, although those studies only lasted a month. “That was long enough to get improvement in cognition, but it’s possible that if we block ΔFosB for longer that could be detrimental,” she acknowledged. “In this paper, we focused on calbindin-D28k, but clearly ΔFosB has many other target genes. We’d like to determine exactly which are detrimental for memory and selectively block the interaction of ΔFosB with those,” Chin said.

The question of whether all the regulation occurs via this epigenetic mechanism or there are other molecular regulators needs to be addressed, said Cole. Chin agreed, and said her group is looking for other regulators now.

Alice Lam, Massachusetts General Hospital, said she would like to see more direct evidence for a causal link between seizures and calbindin-D28k reduction. “The data show a correlation between number of seizures and the reduction in calbindin-D28k. One possible interpretation is that the number of seizures indicates the severity of AD. If treating the mice with levetiracetam would reverse the loss of calbindin-D28k, that would have been really convincing,” she said. More clinical information on the human cases studied might also strengthen the link between seizures and calbindin-D28k expression, Lam said. “Early onset and familial AD patients have a higher risk of developing seizures, so it would be interesting to know age of onset in the patients, and if that relates to calbindin-D28k levels,” she said.

Defining the disease stage in which targeting this pathway might work is also important. “The [authors] mainly look at young two- to four-month-old mice, but we don’t know if old mice will respond, as well,” Lam noted.—Pat McCaffrey


  1. Chronic hippocampal seizures impair cognition, an all-too-common biological comorbidity of temporal lobe epilepsy. What can be done? In the face of irreversible hippocampal cell loss followed by extensive structural rearrangements due to axonal sprouting, dendritic spine loss, and neosynaptogenesis, the prospect of restoring native memory networks has long seemed remote. At the molecular remodeling level, the relative plasticity of heteromeric GABAergic receptor subunit stoichiometry, altered transmembrane chloride gradients, and reduced sodium ion channel current density seem to offer a chance, however slim, of selectively reversing impaired synaptic strength, but these targets have not yet led to effective treatments.

    Now the search for reversible pathophysiology has taken an important new turn. From within the riptides of transcriptional dysregulation long known to occur in dentate granule cells following seizures, Chin and collaborators have isolated a critical pathway centering on the newfound ability of the transcription factor ΔFosB to modulate intracellular calcium buffering by suppressing calbindin-D28, an epigenetic switch that may offer a tractable target for restoring membrane excitability by regulating channel kinetics and other steps in neurotransmitter release.

    Ironically, as one of the members of the immediate early gene family long used as a nuclear biomarker for tracing neuronal hyperactivity, ΔFosB has been hiding in plain sight. Chin showed in a convulsant model of seizures, and in the genetic AD mouse model where behaviorally silent hippocampal epilepsy was first observed, that when seizures stimulate excess calcium entry into granule cells, ΔFosB and ΔFosB alone is sufficient to suppress Calb1 transcription by interfering with promotor-level transcriptional control, resulting in neuronal hyperexcitability. Clever use of a dominant negative interaction with JunD protein to prevent chromatin modification at the promoter region preserved calbindin levels and prevented induced spatial memory deficits.

    This is a welcome discovery revealing an epigenetic target for protection against seizure-induced malfunction of hippocampal circuitry, even in the face of other structural cellular pathology. Not all patterns of epilepsy trigger immediate early gene expression, even in the presence of axonal sprouting, and leave calbindin levels undisturbed, which may account for relative sparing of cognition (Nahm et al., 1998), however, a single strong hippocampal seizure may lead to transient amnesia that might reflect a brief suppression of neuronal calbindin levels. In addition, as the authors suggest, this mechanism could well extend to other clinical etiologies of learning and memory impairment that are below the threshold for classically measurable hypersynchronous discharges.


    . Nonobligate role of early or sustained expression of immediate-early gene proteins c-fos, c-jun, and Zif/268 in hippocampal mossy fiber sprouting. J Neurosci. 1998 Nov 15;18(22):9245-55. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Epilepsy in Alzheimer’s Can Be Early and Subtle
  2. Do "Silent" Seizures Cause Network Dysfunction in AD?
  3. Soluble Aβ Takes Blame for Hyperactive Neurons in Mouse Brain
  4. Anticonvulsants Reverse AD-like Symptoms in Transgenic Mice
  5. Epilepsy Drug Calms the Hippocampus, Aids Memory
  6. More Evidence That Epilepsy Drug Calms Neurons and Boosts Memory
  7. Calbindin Study: Is Calcium the Molecular Handle on Dysfunction in AD?

Therapeutics Citations

  1. Levetiracetam

Paper Citations

  1. . Incidence and impact of subclinical epileptiform activity in Alzheimer's disease. Ann Neurol. 2016 Dec;80(6):858-870. Epub 2016 Nov 7 PubMed.
  2. . ΔFosB Regulates Gene Expression and Cognitive Dysfunction in a Mouse Model of Alzheimer's Disease. Cell Rep. 2017 Jul 11;20(2):344-355. PubMed.
  3. . Correlation between calbindin expression in granule cells of the resected hippocampal dentate gyrus and verbal memory in temporal lobe epilepsy. Epilepsy Behav. 2012 Sep;25(1):110-9. Epub 2012 Jul 15 PubMed.
  4. . Evidence for calcium-reducing and excito-protective roles for the calcium-binding protein calbindin-D28k in cultured hippocampal neurons. Neuron. 1991 Jan;6(1):41-51. PubMed.
  5. . Surviving granule cells of the sclerotic human hippocampus have reduced Ca(2+) influx because of a loss of calbindin-D(28k) in temporal lobe epilepsy. J Neurosci. 2000 Mar 1;20(5):1831-6. PubMed.
  6. . Vulnerability of postnatal hippocampal neurons to seizures varies regionally with their maturational stage. Neurobiol Dis. 2010 Feb;37(2):394-402. Epub 2009 Oct 29 PubMed.

External Citations

  1. company website

Further Reading

Primary Papers

  1. . Epigenetic suppression of hippocampal calbindin-D28k by ΔFosB drives seizure-related cognitive deficits. Nat Med. 2017 Nov;23(11):1377-1383. Epub 2017 Oct 16 PubMed.