Targeting the PSD-95–GluN2B Protein–Protein Interaction in Excitotoxic Signaling: Structural Mechanisms, Disruptor Design, and Neuroprotective Potential

Disruption of disease-relevant protein–protein interactions (PPIs) has emerged as a strategy to modulate pathological N-methyl-D-aspartate receptor (NMDAR) signaling while preserving essential physiological functions. A leading example is the interaction between postsynaptic density protein-95 (PSD-95; also known as SAP90, encoded by DLG4) and the C-terminus of the GluN2B subunit (encoded by GRIN2B), a PDZ-dependent coupling that organizes signaling nanocomplexes at the postsynaptic density and links NMDAR activation to downstream neurotoxic pathways. This narrative review synthesizes the molecular, structural, preclinical and clinical evidence for disrupting this interface, and applies an explicit evidence-weighting framework so that mechanistic, animal and human data are not treated as interchangeable. We summarize the architecture and dynamic regulation of the complex; the structural basis of PDZ-mediated recognition, affinity, specificity and multivalency; and the pathophysiological consequences of PSD-95–GluN2B coupling, including neuronal nitric oxide synthase (nNOS)-linked nitric oxide signaling, NADPH oxidase-dependent superoxide production and downstream death pathways. We then distinguish three agent classes that are frequently conflated: direct PSD-95–GluN2B interface disruptors derived from the GluN2B PDZ ligand, agents directed at the downstream PSD-95–nNOS interface, and compounds that modify PSD-95 by other mechanisms. Clinically, the most advanced agent is nerinetide (Tat-NR2B9c; formerly NA-1). ESCAPE-NA1 and ESCAPE-NEXT did not meet their primary efficacy endpoints, and FRONTIER was neutral in its overall modified intention-to-treat population; the favourable results reported subsequently derive from subgroup and post hoc pooled analyses and are therefore exploratory and hypothesis-generating. We conclude that the PSD-95–GluN2B interface is well validated at the molecular and preclinical level, but that its clinical value remains unproven: no agent of this class is approved by any regulatory authority, and a prospective, adequately powered trial in an early-window, reperfusion-selected population is still required.