By Cleaving GABA(A) Receptors, BACE1 Lifts Lid on Hyperactivity

BACE1, aka β-secretase, is infamous for its fateful snip of amyloid precursor protein that leads to the production of Aβ peptides. Yet this might not be the only way BACE1 eggs on Alzheimer’s pathogenesis. According to a study published February 26 in Neuron, the enzyme also cleaves protein subunits off GABAAR, a receptor that transmits inhibitory currents responsible for reining in neuronal activity. Led by Danlei Bi, Feng Gao, and Yong Shen of the University of Science and Technology of China in Hefei, the scientists report that in one mouse model of amyloidosis, BACE1 cleavage of the receptor’s essential subunits hobbles inhibitory signaling and unleashes neuronal hyperexcitability. BACE1 inhibition, or expression of a BACE1-resistant GABAAR subunit, counteracted this effect, and even lowered Aβ deposition and spatial memory loss in mice. Fragments released by BACE1’s cleavage of GABAAR were reportedly spotted in AD brain, suggesting this mechanism contributes to the unbridled neuronal firing frequently observed in people with AD.

  • In an amyloidosis mouse, BACE1 splits subunits off GABAA receptor.
  • This reduces GABAA inhibitory currents, unleashing neuronal hyperactivity.
  • Protecting the receptors curbed hyperactivity, Aβ plaques, and memory loss.

“These findings provide direct evidence that BACE1 can promote circuit dysfunction and promote Aβ aggregation,” wrote Nicolai Franzmeier and Sebastian Römer of the Ludwig Maximilians University in Munich. “More broadly, this work reinforces the emerging view that neuronal hyperexcitability is not merely a bystander but an active contributor to AD progression.”

Aβ aggregates have long been blamed for stoking neuronal hyperexcitability, including seizures (Sep 2007 news; May 2012 news). More recently, scientists have tied this to the propagation of tau pathology (Dec 2023 news; Jan 2025 news).

Does BACE1 Beckon Hyperactivity? In healthy brain, inhibitory signaling through GABAA receptors keeps neuronal activity in check (left). In AD brain, the hypothesis goes, more BACE1 results in GABAAR cleavage, triggering hyperactivity and spatial memory loss (right). [Courtesy of Bi et al., Neuron, 2025.]

Normally, inhibitory neurons release GABA neurotransmitter to keep excitatory activity in check, and promoting GABAergic signaling fixes circuit dysfunction in amyloid-ridden mice (Busche et al., 2015). Years ago, scientists reported that BACE1 inhibitors quiet overactive neurons in an AD mouse model, but they credited Aβ reduction (Oct 2016 conference news).

For this study, first author Danlei Bi and colleagues wondered if BACE1 might also affect neuronal activity directly. They used the APP23 transgenic mouse model, which develops plaques in the hippocampus and cortex between 6 to 9 months of age. They inserted probes into the hippocampi of anesthetized mice to measure neuronal activity. The scientists recorded revved-up excitatory activity in 3- and 12-month-old APP23 mice relative to wild-type, and squelched inhibitory currents. This suggested to them that even before plaques form, the balance of excitatory and inhibitory neurotransmission shifts, such that inhibitory GABAAR currents weaken while excitatory action potentials fire off with less restraint.

The GABAAR receptor comprises five subunits: two α and two β subunits, and one γ or δ subunit, each of which come in different isoforms. Bi and colleagues investigated their expression and function. In APP23 hippocampi, they found levels of β subunits 1, 2, and 3 to be reduced, while α, γ, and δ subunits were normal. Meanwhile, N- and C-terminal fragments of β subunits were more abundant than normal. Curiously, BACE1 levels were high, too. The scientists report finding the same pattern in cortical samples from 10 people with AD and nine controls: low β1, 2, 3 expression, while levels of β1, 2, 3 fragments and BACE1 were up. Seven of these came from the brain bank at the University of Science and Technology of China, and 12 from the University of Pennsylvania, Philadelphia.

Suspecting BACE1 might be cleaving these GABAAR subunits, Bi and colleagues deployed BACE inhibitors and BACE1 knockdown strategies in primary neuronal cultures of wild-type mice. They report that indeed, BACE1 does snip GABAAR β subunits, pinpointing the cleavage sites to Leu234/Ser235 in β2, and Leu235/Ser236 in β1 and β3. In brain extracts from APP23 mice, BACE1 was found mingling with these receptor subunits in endosomal compartments, where it is most active.

Next, the scientists toggled BACE1 levels in mice—either knocking it out or overexpressing it. They found that brain extracts from BACE1 knockouts had fewer β1/2/3 CTFs and NTFs, while those from BACE1 overexpressers had more. In hippocampal slice cultures, GABAergic inhibitory currents rose in BACE1 knockouts and plummeted in those overexpressing the protease.

Notably, deleting BACE1 did not extinguish cleavage of the β1/2/3 entirely, suggesting other metalloproteases also have their way with the GABAAR subunits.

Finally, the scientists tested whether sliding a BACE1-resistant β3 subunit into the receptor might rebalance inhibitory versus excitatory signaling in APP23 mice. They injected an adeno-associated virus expressing a BACE1-non-cleavable or wild-type β3 subunit into the dentate gyri of 11-month-old APP23 or wild-type mice, and one month later recorded neuronal activity electrophysiologically. Compared to the normal β3, the BACE1-resistant subunit bolstered expression of GABAAR receptors, and reduced neuronal activity to wild-type levels. Moreover, the treated APP23 mice had fewer plaques and staved off spatial memory loss, the scientists claim (image below).

More GABA, Less Plaque. APP23 mice (left) accumulate amyloid plaques (red) in the hippocampus (dashed oval) and somatosensory cortex (dashed quadrilateral). Mice accumulated fewer plaques in the hippocampus (right) when they expressed GABA 3 subunits resistant to BACE cleavage there (green). [Courtesy of Bi et al., Neuron, 2025.]

Why would facilitating GABA transmission reduce plaque load? The idea implies a sinister feedback loop afoot in AD, whereby more BACE1 expression spurs both Aβ production and neuronal hyperexcitability, which amplify each other (Sep 2002 news; Li et al., 2004). “In turn, increased Aβ deposition has been shown to upregulate BACE1 expression, thereby potentially sustaining and amplifying this pathogenic cycle,” Bi, Shen, and Gao wrote to Alzforum.

Could low-dose BACE1 inhibitors break this cycle? The authors are trying to find out if such a sweet spot exists.

On the flip side, Stefan Lichtenthaler of the German Center for Neurodegenerative Diseases in Munich wondered whether BACE1 inhibitors, which notoriously worsened cognition among participants in clinical trials, might have taken neural inhibition too far (Sep 2019 news; Dec 2020 news). “We need to consider whether GABAAR-related functional changes contribute to the cognitive side effects seen for the high doses of BACE inhibitors used in previous clinical trials,” he wrote (comment below).

The study adds to the evidence implicating GABAergic signaling in AD, commented Jefferson Kinney of the University of Las Vegas in Nevada, who studies GABAB receptors expressed on both neurons and glia. Kinney and others reported that, in mouse models of AD, these receptors plummet on microglia, exacerbating amyloidosis (Salazar et al., 2021; Osse et al., 2023).

“This new study brings in another piece about how changes in GABA signaling influence AD pathogenesis,” Kinney told Alzforum. He noted that beyond inhibiting neuronal firing, GABA has many other roles, including immune function, which come to the fore in AD. “There’s a larger story shaping up about the role of GABA in AD—we just don’t know what it is yet,” he said.—Jessica Shugart

Comments

  1. This is a very interesting story, adding GABAAR subunits to the growing list of BACE1 substrates. In some of the experiments, changes in BACE1 activity had surprisingly strong effects on the cleavage and activity of the receptor/subunits. I would love to see whether clinically used BACE1-targeted inhibitors show effects similar to those of the knockout. If yes, we need to consider whether GABAAR-related functional changes contribute to the cognitive side effects seen for the high doses of BACE inhibitors used in previous clinical trials.

    View all comments by Stefan Lichtenthaler

  2. I was not surprised by the findings. We know that the GABAergic system is altered in AD, and this includes at least alteration/loss of interneurons and alteration of GABA receptors. We have been working quite a bit, and published some articles, on how GABAB receptors are downregulated in animal models of AD, using histological and ultrastructural techniques. We have also been working for the last three years on the potential alteration of GABAA receptors and found that the alpha1 subunit of GABAA receptors is downregulated in GABAergic synapses and have some evidence that β subunits can be altered. This information is not ready for publishing because we are now concentrated in glutamatergic receptors and do not have enough time and people to complete the experiments.

    This article by Bi and colleagues confirms what I had in mind. The role of GABAA receptors is to inhibit neurons. It is known that AD causes hyperexcitation, which may occur by increasing the number or activation of glutamate receptors, by decreasing the number or activation of GABAA receptors, or by both at the same time. This paper demonstrates that BACE1 cleaves GABAA β1/2/3 subunits, resulting in decreased GABAAR-mediated inhibitory currents. Therefore, this cleavage promotes neural hyperexcitability, which is involved in the progression of cognitive impairments.

    The article is fantastic and elegantly demonstrates their findings, thus opening a new avenue for therapeutic intervention to prevent cleavage of β subunits of GABAA and slow AD progression.

    For sure, we will also have to take into consideration glutamate receptors and how to prevent their hyperactivity, but at least finding this new target is a good step to understand the pathological mechanisms taking place in AD and why it is such a complex neurodegenerative disease.

    View all comments by Rafael Luján Miras

  3. In this insightful study, Bi et al. identify a novel mechanism linking Aβ pathology to neuronal hyperexcitability via BACE1-mediated impairment of inhibitory GABAergic signaling. Best known for its role in Aβ generation, the BACE1 enzyme is shown here to aberrantly cleave the β1-3 subunits of GABAA receptors in an amyloid mouse model and in postmortem samples from AD patients, thereby diminishing inhibitory currents and promoting neuronal hyperactivity. Strikingly, the authors demonstrate that preventing this cleavage through a non-cleavable GABAA β3 subunit variant can restore inhibitory function. This intervention alleviated neuronal hyperactivity, led to reduced Aβ plaque burden, and improved memory performance in an APP transgenic mouse model. These findings provide direct evidence that BACE1 can promote circuit dysfunction and promote Aβ aggregation.

    More broadly, this work reinforces the emerging view that neuronal hyperexcitability is not merely a bystander but an active contributor to AD progression (Roemer-Cassiano et al., 2025; Targa Dias Anastacio et al., 2022; Busche and Konnerth, 2015). Excessive neuronal firing has been observed in AD mouse models (often in proximity to Aβ plaques; Busche et al., 2008) and such aberrant activity has been linked to exacerbation of tau pathology (Wu et al., 2016). For example, hyperactive neurons can release greater amounts of pathological tau, potentially seeding its spread to connected brain areas (Roemer-Cassiano et al., 2025; Wu et al., 2016). 

     Consistent with this, our recent human imaging study found that Aβ deposition triggers an increase in functional connectivity (i.e., indicative of network hyperactivity) thereby driving accelerated spread of tau pathology across those hyperconnected regions (Roemer-Cassiano et al., 2025). In other words, Aβ-related neural hyperactivity appears to promote the spread of tau pathology, effectively linking the two hallmark pathologies.

    Bi et al.’s findings add an important mechanistic dimension to this picture, illustrating one way that Aβ may trigger such detrimental neuronal hyperactivity. Their data suggest a vicious cycle: BACE1 levels become elevated in the AD brain (perhaps in response to early Aβ accumulation), leading to excessive cleavage of both APP and GABAA receptor subunits. The latter weakens inhibitory control and amplifies neuronal firing. This hyperexcitability, in turn, feeds back to promote increased Aβ production, ultimately accelerating AD progression. What the current study does not show, however, is how this BACE1-related upregulation of neuronal excitability and activity may contribute to tau secretion and spread, which should be tested in future studies.

    The study also carries important therapeutic implications. By pinpointing a specific molecular cascade that links Aβ to network dysfunction, it highlights new opportunities to break this cycle. One approach is to directly dampen neuronal hyperactivity in at-risk patients—indeed, clinical trials are testing anti-epileptic drugs and neuromodulation (e.g., transcranial magnetic stimulation, focused ultrasound) to reduce aberrant brain activity in AD. Another approach would be to modulate BACE1 activity more selectively. The disappointing past trials of broad BACE1 inhibitors might be revisited with renewed hope if inhibitors can be tailored to spare key substrates like GABAA receptors while still curbing Aβ production. Overall, by elucidating how Aβ pathology undermines inhibitory circuits to promote AD progression, Bi et al. provide a compelling case that stabilizing neural networks—through restoring inhibition or fine-tuning BACE1’s actions—could help treat neuronal hyperexcitability and ultimately AD progression.

    References:

    Roemer-Cassiano SN, Wagner F, Evangelista L, Rauchmann BS, Dehsarvi A, Steward A, Dewenter A, Biel D, Zhu Z, Pescoller J, Gross M, Perneczky R, Malpetti M, Ewers M, Schöll M, Dichgans M, Höglinger GU, Brendel M, Jäkel S, Franzmeier N. Amyloid-associated hyperconnectivity drives tau spread across connected brain regions in Alzheimer's disease. Sci Transl Med. 2025 Jan 22;17(782):eadp2564. Epub 2025 Jan 22 PubMed.

    Targa Dias Anastacio H, Matosin N, Ooi L. Neuronal hyperexcitability in Alzheimer's disease: what are the drivers behind this aberrant phenotype?. Transl Psychiatry. 2022 Jun 22;12(1):257. PubMed.

    Busche MA, Konnerth A. Neuronal hyperactivity--A key defect in Alzheimer's disease?. Bioessays. 2015 Jun;37(6):624-32. Epub 2015 Mar 14 PubMed.

    Busche MA, Eichhoff G, Adelsberger H, Abramowski D, Wiederhold KH, Haass C, Staufenbiel M, Konnerth A, Garaschuk O. Clusters of hyperactive neurons near amyloid plaques in a mouse model of Alzheimer's disease. Science. 2008 Sep 19;321(5896):1686-9. PubMed.

    Wu JW, Hussaini SA, Bastille IM, Rodriguez GA, Mrejeru A, Rilett K, Sanders DW, Cook C, Fu H, Boonen RA, Herman M, Nahmani E, Emrani S, Figueroa YH, Diamond MI, Clelland CL, Wray S, Duff KE. Neuronal activity enhances tau propagation and tau pathology in vivo. Nat Neurosci. 2016 Aug;19(8):1085-92. Epub 2016 Jun 20 PubMed.

    View all comments by Nicolai Franzmeier View all comments by Sebastian N. Roemer-Cassiano

  4. In his seminal publication of 1911 titled “Über eigenartige Krankheitsfälle des späteren Alters“ (On peculiar diseases of older age), Alois Alzheimer already noted that epileptiform activity is a common symptom of “his” disease, but it took about a century to unravel the underlying mechanisms. Previous work from the Kovacs lab at Harvard and the Palop/Mucke lab at University of California, San Francisco, showed that the activity of GABAergic interneurons is reduced in Alzheimer´s disease (AD) models due to a BACE1-mediated decline in the surface expression of Nav1.1. This subtype of voltage-gated Na+ channel is essential to drive action potential firing of GABAergic interneurons. The new study by Bi et al. examined what happens after the interneurons have fired and released GABA. The paper is very interesting, revealing an additional BACE1-dependent mechanism that impairs GABAergic inhibition, this time by clipping GABAA receptors, which serve as major effectors of synaptic inhibition in the postsynaptic neuron.

    The authors make a compelling case for a fatal, disease-propelling interaction between BACE1 and postsynaptic GABAA receptor. Like any paper that opens a new venue of research, it gives also rise to several intriguing questions. For example, does the BACE1-induced loss of functional GABAA receptors on the postsynaptic site engender changes in spontaneous release properties on the presynaptic site, as suggested by the reduced frequency of miniature inhibitory postsynaptic currents (mIPSCs)? Are synaptic and extrasynaptic GABAA receptors that mediate phasic and tonic inhibition, respectively, both cleaved by BACE1, thus equally reducing the two functionally distinct components of inhibition? In a similar vein, is the delta-subunit of GABAA receptors, which is essential to target GABAA receptors to extrasynaptic sites in hippocampal granule cells, a substrate of BACE1? How can we explain the apparent paradox that BACE1-deficient mice exhibit an epileptic phenotype, too, rather than being over-inhibited?

    The final question relates to the conspicuous coincidence that, in the healthy adult brain, the axons of granule cells, the so-called mossy fibers, express high levels of BACE1 concomitant with the presence of presynaptic GABAA receptors, which regulate transmitter release from mossy fiber terminals. Assuming that the presynaptic GABAA receptors are also substrates of BACE1, would the high secretase level not predict a functionally relevant interaction with presynaptic GABAA receptors already under physiological conditions? 

    View all comments by Christian Alzheimer

  5. A while back, we sought to characterize the electrophysiological phenotypes in one of the earliest presenilin transgenic mouse models (Zaman et al., 2000). LTP was enhanced in hippocampi of these mice, and this phenotype could be normalized with the use of the benzodiazepine diazepam (a GABAA positive modulator). One hypothesis of this GABAA receptor effect was that it was a consequence of the “build-up” of hyperexcitability, resulting in a compensatory response as seen in the brain slices.

    Bi and colleagues’ paper provides a possible explanation of a mechanism for this change. Benzodiazepines bind to γ and α subunits and do not require β subunits for action on the receptor.

    The epidemiological clinical data on the use of benzodiazepines, however, has been inconclusive in terms of benzodiazepines delaying cognitive decline, probably due to the adverse effects of these drugs especially in the elderly. Although there is an ongoing clinical trial using the non-GABAergic drug levetiracetam to test the notion of hyperexcitability of seizures being detrimental to disease progression, it would be important to establish at what stage of the disease this type of intervention would be useful.

    References:

    Zaman SH, Parent A, Laskey A, Lee MK, Borchelt DR, Sisodia SS, Malinow R. Enhanced synaptic potentiation in transgenic mice expressing presenilin 1 familial Alzheimer's disease mutation is normalized with a benzodiazepine. Neurobiol Dis. 2000 Feb;7(1):54-63. PubMed.

    View all comments by Shahid Zaman

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Do "Silent" Seizures Cause Network Dysfunction in AD?
  2. Soluble Aβ Takes Blame for Hyperactive Neurons in Mouse Brain
  3. Plaques Kick Neocortical Neurons into Overdrive, Entangling Tau
  4. Plaques Spur Spread of Tangles by Sending Synapses into Overdrive
  5. Does BACE Drive Neurites into Dystrophy, Shorting Circuits?
  6. BACE Above Base in Alzheimer’s Patients
  7. End of the BACE Inhibitors? Elenbecestat Trials Halted Amid Safety Concerns
  8. New Data from Past BACE Inhibitor Trials Shed Light on Side Effects

Research Models Citations

  1. APP23

Paper Citations

  1. Busche MA, Kekuš M, Adelsberger H, Noda T, Förstl H, Nelken I, Konnerth A. Rescue of long-range circuit dysfunction in Alzheimer's disease models. Nat Neurosci. 2015 Nov;18(11):1623-30. Epub 2015 Oct 12 PubMed.
  2. Li R, Lindholm K, Yang LB, Yue X, Citron M, Yan R, Beach T, Sue L, Sabbagh M, Cai H, Wong P, Price D, Shen Y. Amyloid beta peptide load is correlated with increased beta-secretase activity in sporadic Alzheimer's disease patients. Proc Natl Acad Sci U S A. 2004 Mar 9;101(10):3632-7. PubMed.
  3. Salazar AM, Leisgang AM, Ortiz AA, Murtishaw AS, Kinney JW. Alterations of GABA B receptors in the APP/PS1 mouse model of Alzheimer's disease. Neurobiol Aging. 2021 Jan;97:129-143. Epub 2020 Oct 23 PubMed.
  4. Osse AM, Pandey RS, Wirt RA, Ortiz AA, Salazar A, Kimmich M, Toledano Strom EN, Oblak A, Lamb B, Hyman JM, Carter GW, Kinney J. Reduction in GABAB on glia induce Alzheimer's disease related changes. Brain Behav Immun. 2023 May;110:260-275. Epub 2023 Mar 9 PubMed.

Further Reading

Papers

  1. Rodriguez GA, Barrett GM, Duff KE, Hussaini SA. Chemogenetic attenuation of neuronal activity in the entorhinal cortex reduces Aβ and tau pathology in the hippocampus. PLoS Biol. 2020 Aug;18(8):e3000851. Epub 2020 Aug 21 PubMed.

Primary Papers

  1. Bi D, Bao H, Yang X, Wu Z, Yang X, Xu G, Liu X, Wan Z, Liu J, He J, Wen L, Jing Y, Zhu R, Long Z, Rong Y, Wang D, Wang X, Xiong W, Huang G, Gao F, Shen Y. BACE1-dependent cleavage of GABAA receptor contributes to neural hyperexcitability and disease progression in Alzheimer's disease. Neuron. 2025 Feb 25; Epub 2025 Feb 25 PubMed.

Annotate

To make an annotation you must Login or Register.