Transcriptional modulation of β-Secretase ameliorates amyloid pathology and cognitive deficits in Alzheimer’s disease models

Dear Editor,

Excessive amyloid‑β (Aβ) production, driven by β‑site APP cleaving enzyme‑1 (BACE1), is central to the pathogenesis of Alzheimer’s disease (AD), yet active‑site BACE1 inhibitors have failed in trials because of efficacy and safety limitations. Targeting BACE1 transcription offers an orthogonal strategy that may enable partial, context‑dependent suppression compatible with neuronal function.¹

We developed a ratiometric BACE1 promoter reporter (RFP under the human BACE1 promoter; GFP under CMV for internal normalization) and screened 1200 clinically used molecules (Prestwick). Chlorhexidine (CHX) emerged as the strongest suppressor of BACE1 promoter-linked reporter activity in this screen, with a clear dose–response, no detectable cytotoxicity, and no significant change in the action potential threshold at effective concentrations (≤1 µM) (Fig. 1a). Because this reporter system was used as a screening tool rather than as stand-alone proof of promoter-specific regulation, candidate hits were further evaluated using endogenous BACE1 expression and APP processing readouts. In SH-swAPP cells, CHX reduced BACE1 levels and decreased the C99/C83 ratio, indicating reduced β-secretase-dependent APP processing. Together, these findings support the suppression of BACE1 expression and amyloidogenic processing under AD-relevant conditions.

Fig. 1.

Chlorhexidine (CHX) suppresses BACE1 expression and is associated with reduced amyloid pathology and improved behavioral performance in APP/PS1 mice. a Representative fluorescence images of DAPI, cell body, and BACE1 staining (scale bar = 30 μm).b Behavioral analysis of vehicle- or CHX-treated APP/PS1 mice and WT littermate controls. Behavioral performance was assessed using the Morris water maze. c Amyloid pathology and cortical BACE1 in vehicle- or CHX-treated APP/PS1 mice. Representative Aβ immunohistochemistry images in the hippocampal and cortical regions detected by DAB staining (scale bar = 90 μm). d Schematic of the human BACE1 promoter 5′-deletion constructs showing the putative STAT (GAS) elements and their sequences with luciferase reporter activity (Firefly/Renilla) for vehicle and CHX (1 µM) across constructs (left). Representative immunoblots of BACE1, SOCS3, phospho-STAT3 and STAT3 under control and SOCS3-knockdown conditions with or without CHX (right). Data are presented as mean ± SEM. Statistical analyses were performed using one-way ANOVA followed by Tukey’s post hoc test, unless otherwise indicated. Luciferase reporter data were compared between vehicle and CHX groups for each construct using unpaired two-tailed t-tests

To assess whether the BACE1-suppressing effect of chlorhexidine observed in vitro was associated with improvement in vivo, APP/PS1 transgenic mice and wild-type (WT) littermate controls were orally administered 10 mg/kg chlorhexidine per day from 9 months of age for 8 weeks. Throughout dosing, CHX-treated and control mice showed no significant differences in body weight or in locomotor activity during behavioral testing, with no overt signs of toxicity. CHX-treated APP/PS1 mice showed improved performance in the Morris water maze and passive avoidance tests (Fig. 1b). In parallel, CHX was associated with reduced plaque burden in the cortex and hippocampus (Fig. 1c). CHX treatment also reduced BACE1 expression in the cortex of APP/PS1 mice and decreased the C99/C83 ratio, consistent with reduced amyloidogenic processing in vivo.

To identify upstream mediators of BACE1 repression, we profiled hESC-derived neurons after CHX exposure. RNA-seq and pathway enrichment analysis revealed a JAK1–STAT3–SOCS3-related response. Consistent with these findings, CHX no longer decreased phospho-STAT3 or BACE1 under SOCS3 knockdown, suggesting that the suppression of BACE1 by CHX involves a SOCS3-dependent pathway (Fig. 1d). Using BACE1 promoter reporters with serial 5′-deletions, the full-length construct (-1942 form), which contains three putative STAT elements, had the highest basal activity and was suppressed by CHX (to approximately 73% of the control at 1 µM), whereas deletion of the two distal elements reduced basal promoter activity and abolished the CHX response. Together with earlier reports that JAK/STAT signaling can act on the BACE1 promoter,^2,3 these data fit a working model in which CHX drives a SOCS3-dependent, JAK/STAT-responsive state that decreases STAT-driven BACE1 transcription.

This work identifies CHX as a previously unrecognized chemical modulator emerging from a BACE1 regulatory screen and shows that CHX treatment is associated with reduced amyloid pathology and improved behavioral performance in APP/PS1 mice; these findings support the involvement of a SOCS3-dependent pathway in BACE1 suppression in hESC-derived neurons. In contrast to catalytic BACE1 inhibition, the modulation of BACE1 expression may provide an alternative route for partial suppression associated with pathological benefit.^1,4 Given the established clinical use of CHX as an antiseptic and its pleiotropic pharmacology, CHX should be regarded here as a nonselective pharmacological probe rather than a repurposing-ready AD therapeutic. That said, the effect does not appear to be simply broad, nonspecific suppression. Components of the γ-secretase complex were unchanged by CHX, and the reduction in gliosis and neuroinflammatory markers is more readily explained as a downstream consequence of the lower amyloid burden than as a separate off-target effect.

This study has several limitations. Our analysis focused on Aβ-associated pathology and did not extend to Tau phosphorylation, Tau aggregation, or Aβ–Tau interactions. A preliminary tau-BiFC cell assay showed a modest trend toward lower signal after CHX treatment, but this was not quantified statistically and should be regarded as exploratory. Dedicated tau-relevant models will be needed to test this directly. We did not measure CHX pharmacokinetics or BBB penetration, and although we assessed several specificity-related readouts, a comprehensive off-target and selectivity profile of CHX, including effects on synaptic transmission, was not performed. Direct pharmacokinetic measurements, longer-term safety assessments, and optimization of brain-penetrant derivatives are needed before any translational conclusions can be drawn.

Nonetheless, these findings support the broader proof-of-concept that pharmacological suppression of BACE1 expression can have pathological and functional relevance in AD models and may inform future mechanism-anchored optimization targeting BACE1 regulation.⁵ Whether this approach can be combined with other strategies, such as anti-Aβ immunotherapy or tau-directed interventions, remains to be explored.

Data availability

The RNA-seq data generated in this study have been deposited in the European Nucleotide Archive (ENA) under accession number PRJEB107683. All other data supporting the findings of this study are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Ethical approval

All in vivo procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Sungkyunkwan University (SKKUIACUC20150041).

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41392-026-02957-1.