Fig 2: TFEB expression significantly reduces the number of lipofuscin puncta in both the cortex and hippocampus of the P301S model of tauopathy. A, Coronal brain sections from the three genotypes of mice as shown were stained with 1% SBB for 8 min and then counterstained with 0.1% Nuclear Fast Red for 10 min. The quantitation of lipofuscin density in the frontal cortical neurons revealed an average of ∼10.9 granules/neuron in the P301S tauopathy model, which was significantly reduced to 4.6 granules/neuron, a reduction of 57%, in the double-transgenic mice. WT neurons barely showed any lipofuscin granules. B, Brain sections were similarly processed as in A, and lipofuscin granules were quantified in the CA1 region of the hippocampus, which showed an average of 11 granules in the P301S mice and 4.9 in the double-transgenic mice, a reduction of 55% due to TFEB overexpression. Statistical analysis by one-way ANOVA followed by Student–Newman–Keuls post hoc test revealed significant differences. ***p < 0.001, compared to WT or P301S mice, as indicated. The data are reported as the mean ± SEM. n = 5 per genotype.

Fig 3: TEM demonstration of increased autophagy markers in the flag-TFEB transgenic mice. The ultramicrotome sections were stained for TEM analysis. Unbiased quantitation also showed increases in single-membrane autolysosomes (AL) by 339% per 100 µm2 area and in autophagosome fusion with lysosomes (AF) by 144% per 100 µm2 area in the cortices of flag-TFEB mice compared with those of WT control mice. For each of the autophagy markers, the boxed structures in red are shown at higher magnification for clarity. Data are reported as the mean ± SEM. n = 4 mice per genotype. **p < 0.01 by Student’s t-test.

Fig 4: TFEB expression in the P301S mice significantly attenuates learning and memory deficits. A, In the Barnes maze test, on day 1 of training, time spent in the target quadrant was 33% less by P301S mice and only 18% less by P301S/flag-TFEB double-transgenic mice compared with WT controls. On day 2, P301S mice spent 32% less time, but double-transgenic mice spent only 5% less time. On day 4, the difference was 38% for P301S mice and only 6% for the double-transgenic mice. During the probe test, the difference was from 50% to 17%. B, In the T maze paradigm, on day 3 of testing, the correct responses were 58% for WT mice, 28% for P301S mice, and 57% for double-transgenic mice. On day 4 of testing, the correct responses were 73% for WT mice, 43% for P301S mice, and 68% for double-transgenic mice. On day 5, the correct responses were 87% for WT mice, 55% for P301S mice and 80% for the double-transgenic mice. During the probe test, the correct responses were 82% for WT mice, 32% for P301S mice, and 68% for the double-transgenic mice. The latency also differed. On day 3, the average latencies for WT, P301S, and double-transgenic mice were 29, 45, and 36 s, respectively. On day 5, the average latencies for WT, P301S, and double-transgenic mice were 17, 32, and 19 s, respectively. On the day of the probe test, the latencies for WT, P301S, and double-transgenic mice were 14, 30, and 18 s, respectively. *p < 0.05, **p < 0.01, and ***p < 0.001, compared with WT controls. $$p < 0.01, $$$p < 0.001, compared with P301S mice. The data are reported as the mean ± SEM. n = 6 per genotype.

Fig 5: Immunohistochemical demonstration of widespread expression of flag-TFEB in the cortex by confocal microscopy in the transgenic line 7903. A, Cortical brain sections from WT controls showed no signals for flag-TFEB (red, M2 antibody) but showed only DAPI-stained nuclei (blue). B, Brain sections from the transgenic line 7903 shows the expression of flag-TFEB (red). The merged images showed the expression of flag-TFEB both in the nuclei and the cytoplasm.