Fig. 9. Ablation of CXCR3 alleviates astrocytic TDP-43–linked memory deficits.
(A) Experimental design. Transgenic Cxcr3-WT or Cxcr3-KO male mice on a Aldh1l1-Cre background were injected with AAV PHP.eB-hGfaABC1D-DIO–hTDP-43–ΔNLS (ΔNLS) or saline (Con) at 4 to 9 months of age and tested in the Morris water maze at 7 to 12 months of age. Control AAV injections in NTG mice are shown in Fig. 3 and fig. S4. (B) hTDP-43 (green) and the astrocytic marker GFAP (red). DAPI (blue) was used to visualize nuclei. Yellow indicates overlay of green and red channels. Insets i to iv show magnified views. Scale bar, 300 μm. (C) Distance traveled during hidden platform training. Mixed-effects model: F(1,70) = 17.84, P < 0.0001 for ΔNLS; F(1,70) = 8.93, P = 0.0039 for Cxcr3-KO; F(1,70) = 0.12, P = 0.73 for ΔNLS–Cxcr3-KO interaction. (D to G) Probe trial 24 hours after training. Durations in target and nontarget (other) quadrants. Two-way ANOVA (target): F(1,69) = 8.82, P = 0.004 for ΔNLS; F(1,69) = 5.94, P = 0.017 for Cxcr3-KO (D); and F(1,69) = 16.3, P = 0.0001 for ΔNLS (E). Bonferroni’s test: ##P < 0.01 versus control/Cxcr3-WT and #P < 0.05 versus ΔNLS/Cxcr3-WT. Student’s t test with Welch’s correction: ***P < 0.001, **P < 0.01, and *P < 0.05 versus other. (F) Swim paths during the probe trial; insets show magnified views of the target platform area. (G) Swim speeds in the 24-hour probe trial.