Short-term spatial training is associated with reduced amyloid pathology and modulated HMGB1-GSK3β-IL-18BP signaling in APP/PS1 mice

Amyloid-β (Aβ) pathology and neuroinflammation are central features of Alzheimer’s disease (AD). The Morris water maze (MWM) is widely used to assess spatial learning and memory, but its use as a brief training protocol may also influence subsequent pathological measurements. Here, 280-day-old female APP/PS1 mice underwent 5 days of hidden-platform training followed by a probe test on day 6, with swim-only and naïve control groups for comparison. Trained mice spent more time in the target quadrant than swim-only mice, with no significant difference in swimming speed. Cortical 6E10-positive Aβ plaque density was lower in trained mice than in both control groups; hippocampal plaque density was lower than in swim-only mice but did not differ significantly from naïve controls. Size-stratified analyses showed lower small-plaque density (< 20 μm), with no significant differences in the larger size categories. Aβ40- and Aβ42-positive plaque densities were also lower in trained mice than in naïve controls. Training was associated with reduced microglial clustering and increased plaque-associated microglia per plaque in the cortex. Western blot analysis showed lower HMGB1 and total GSK3β and higher IL-18BP protein levels, whereas RAGE did not change significantly. These expression and morphological findings identify candidate inflammatory changes associated with MWM training, without establishing pathway causality, enhanced phagocytosis, or plaque clearance. The absence of swim-only comparisons in several downstream analyses limits attribution of those findings specifically to spatial learning. These results highlight the need to consider MWM exposure when interpreting subsequent Aβ-related pathological endpoints.