Synaptic Toxicity of OGA Inhibitors and the Failure of Ceperognastat

Abstract

O-GlcNAcase inhibitors (OGAi) have been proposed as therapeutics for Alzheimer's disease due to their ability to increase O-GlcNAcylation of tau and reduce its aggregation. However, the recent failure of ceperognastat in a Phase II trial-marked by accelerated cognitive decline in the treatment arm-has raised concerns about the safety of this therapeutic class. Here, we evaluated the acute synaptic effects of three structurally distinct OGAi (ceperognastat, ASN90, and MK8719) in mouse hippocampal slices. Electrophysiological recordings revealed that all three compounds impaired both short- and long-term synaptic plasticity, as evidenced by reduced paired-pulse facilitation/depression and suppressed long-term potentiation. Immunohistochemistry showed altered synaptic protein levels, with increased PSD-95 and reduced Synaptophysin 1 in neurons, alongside a biphasic shift in Tau phosphorylation. These findings indicate that OGAi produce rapid and convergent synaptotoxic effects across pre- and postsynaptic compartments, likely reflecting a class-wide mechanism. We argue that electrophysiological screening should be standard in CNS drug development and caution against targeting essential synaptic processes in chronic neurodegenerative conditions.

Figure 1.. OGA inhibitors alter synaptic plasticity.

A, sample traces of whole-cell patch clamp recording of hippocampal CA1 pyramidal neurons during the first episode of paired-pulse stimulation (A1) and aligned first response from every episode (A2). Scale bar, 20ms (horizontal) and 0.5pA (vertical). The color coding for DMSO (black), ASN90 (i.e., ASN, red), LY3372689 (i.e., LY, green), and MK8719 (i.e., MK, blue) is used in all plots. B, plots of EPSC ratio for the responses during the first episode (B1) and the first responses of five episodes (B2). C, LTP measured as relative EPSP slopes after the treatments of OGAi or DMSO. Left insets are sample traces pre- and post 100-Hz 1-s tetanus stimulation. Right inset is a violin plot of the average values during the last 30-s recording (i.e., gray box in the plot). Welch’s ANOVA test, W (DFn, DFd) = 10.61 (3.000, 17.50), p = 0.0003; Kolmogorov-Smimov test, p = 0.0063 (ASN vs. DMSO), 0.0336 (LY vs. DMSO), and 0.0063 (MK vs. DMSO). N = 3 mice and n = 3 slices (i.e., 9 recordings in total for every treatment).

Figure 2.. OGA inhibitors alter synaptic protein levels and Tau phosphorylation.

A, sample images of fluorescence immunocytochemistry for PSD95, Tuj1, GFAP, and DAPI. Scale bar, 0.5 mm. Images on the left are zoom-in images of PSD95 immunofluorescence in the cell-body layer (indicated by the white box) for all four treatments. Scale bar, 50 μm B, sample images of fluorescence immunocytochemistry for Synaptophysin 1, Tau, pTau, and DAPI. Images on the left are zoom-in images of Synaptophysin 1 immunofluorescence in the cell-body layer (indicated by the white box) for all four treatments. C&D, violin plots of PSD95 (C) and Synaptophysin 1 (D) immunofluorescence intensities. E, violin plots of pTau vs. Tau ratios from the same ROIs of Synaptophysin 1. Welch’s ANOVA test for PSD95, W (DFn, DFd) = 90.96 (3.000, 2168), p < 0.0001; for Synaptophysin 1, W (DFn, DFd) = 153.30 (3.000, 2187), p < 0.0001; for pTau/Tau, W (DFn, DFd) = 128.40 (3.000, 1918), p < 0.0001. Kolmogorov-Smimov test for all three, p < 0.0001 (ASN vs. DMSO, LY vs. DMSO, and MK vs. DMSO). N = 9 slices from 3 mice (3 ea), n = 999 ROIs (111 randomly selected from every slice).