Extended Data Fig. 3: Chronic IL-1β impacts neither anxiety, activity or motility but causes long-lasting memory impairment despite clearance of the cytokine from the serum.

C57BL/6 J male mice were subcutaneously implanted with osmotic mini-pumps containing saline or IL-1β diluted in saline solution (5 µg/kg/day) for 35 days. (a) Body mass was monitored post-implantation (Impact of time (p < 0.0001), treatment (p = 0.03) and the interaction of both (p < 0.0001) analyzed by mixed effects model analysis (REML), *p < 0.05, ***p < 0.001 by Bonferroni post-hoc tests, pooled results from three independent experiments. Three (3) weeks post-implantation, mice were tested in the EPM (b-c), followed by the Barnes maze (d-f) the object location task (g). A Barnes maze was also performed 10 weeks post implantation (h-k). (b-c) In the EPM, all mice behaved similarly in terms of distance run during the first 5 min (b) or anxiety levels as shown by the percentages of time spent within the open arms (c) (Pool from 2 independent cohorts, results analyzed by Student’s t-tests with Welch correction). (d-f) In the Barnes maze performed when IL-1β concentrations are elevated in the serum (see Fig. 3a–f), the velocities of mice were assessed during training (e) and the probe trial (f) and the total distance run was measured during the probe trial (d). (g) In the object location task (as shown in Fig. 3h, during the first exploration of the arena, the distance run by the mice was measured. (h) IL-1β concentrations in the serum were assessed by ELISA 3 days after the Barnes maze probe test, that is, 75 days post-minipump implantation. Results from one experiment (unpaired Student’s t test). Each dot represents the value for one mouse. (i-k) In the Barnes maze, learning curves show mean daily distance run (i), or mean number of errors of hole visits (j) prior to finding the target hole connected to a hidden exit box. (k) Number of visits of the original exit hole (target) vs. other holes in other quadrants (other) during the probe trial. Data pooled from 2 independent cohorts. Effect of time (p < 0.0001) was analyzed by repeated measures mixed-effects models in e, i and j. ns, non-significant by Dunn’s tests. Results were analyzed by paired t-tests within group between target and other (k), and by Student’s t-tests compared to the saline group as indicated by brackets (d, f, g & h). Results pooled from 2 independent cohorts. ns: non-significant. All statistical tests were two-sided. Each dot represents an individual mouse, and n are indicated in each bar of the graphs, or in parentheses. Graphs show mean ± s.e.m.