Landmark-based but not vestibular-based orientation elicits mossy fiber synaptogenesis in the mouse hippocampus

This study tries to shed light on the paradoxical finding that two inbred strains of mice C57BL/6 (C57) and DBA/2 (DBA), with differences in hippocampal function, perform similarly in the water maze (WM). Mice from both strains were trained on WM protocols permitting or preventing the use of vestibular signals. Hippocampal involvement in performance was then assessed by estimation of post-training mossy fiber (MF) synaptogenesis. We found that C57 and DBA mice performed similarly when both visual and vestibular information were available but only C57 mice exhibited new MF synapses. Disruption of vestibular inputs impaired performance in DBA mice but not in C57 mice which still exhibited a post-training increase of hippocampal MF synaptic terminals. This strain-specific dissociation indicates that DBA mice can navigate successfully by relying on vestibular signals without engaging their hippocampus. In contrast, vestibular signals are irrelevant for C57 mice since their suppression neither disrupts their behavior nor prevents the formation of new hippocampal synapses. These findings suggest some caution is required in considering performance on standard WM protocols as an index of hippocampus-based learning. Estimating the extent of post-training mossy fiber synaptogenesis would be helpful in solving this issue.