The majority of TUNEL positive cells in neocortex of Tsc1^GFAPCKO mice are neurons. Double-labeling demonstrated that the majority of TUNEL positive cells in neocortex of Tsc1^GFAPCKO mice were also positive for Neu-N, a neuronal marker. Examples of double-labeled cells marked by arrows. Calibration bars = 50 μm for all panels.

Low frequency glutamatergic synaptic transmission is normal in CA1 neurons in hippocampal slices from Tsc1^GFAPCKO mice. Representative traces show excitatory and inhibitory postsynaptic currents mediated by non-NMDA, NMDA and GABA receptors in CA1 neurons in response to low frequency (0.1 Hz) stimulation of Schaffer collaterals in control and Tsc1^GFAPCKO hippocampal slices. See Table 1 for quantitative analysis.

Neuronal death occurs in the neocortex and hippocampus of Tsc1^GFAPCKO mice. (A,B) Representative examples of TUNEL-positive cells in the CA1 region of hippocampus (B) of a 3 month old Tsc1^GFAPCKO mouse, compared to absent TUNEL staining in a 3 month old control mouse (A). (C,D) Representative examples of Fluoro-Jade B positive cells in CA1 region of hippocampus (D) in the same Tsc1^GFAPCKO mouse as in B, compared to absent Fluoro-Jade B staining in a 3 month old control mouse (C). (E,F) Representative examples of Caspase-3 staining in neocortex (E) and hippocampus (F) in a 1 month old Tsc1^GFAPCKO mouse. Calibration bars = 50 μm for all panels.

Spatial and temporal distribution of TUNEL and Caspase-3 positive cells in Tsc1^GFAPCKO and control mice. Numbers of TUNEL (A) and Caspase-3 (B) positive cells in neocortical or hippocampal regions per section were counted in a blinded fashion in 1 and 3 month old control and Tsc1^GFAPCKO mice (n = 3 sections per mouse, 4-5 mice per group). There was a significant difference between control and Tsc1^GFAPCKO mice in TUNEL and Caspase-3 staining in neocortex, CA1, and CA3 regions (p < 0.05 by ANOVA for overall comparison between all 4 groups; p <0.05 for post-tests between control and Tsc1^GFAPCKO mice at both 1 and 3 months). Despite an apparent trend towards increased cell death between 1 and 3 month old Tsc1^GFAPCKO mice in CA1 and CA3 regions of hippocampus, there were no statistically-significant differences in TUNEL or Caspase-3 staining between 1 and 3 month old Tsc1^GFAPCKO mice.

Long-term potentiation induced by high-frequency synaptic stimulation is impaired in hippocampal slices from Tsc1^GFAPCKO mice. (A) Representative examples of field EPSPs in the CA1 region before and after tetanic (100 Hz) stimulation to the Schaffer collaterals to activate LTP. (B) Normalized initial EPSP slope from Tsc1^GFAPCKO mice demonstrate a significant decrease in LTP following the tetani (arrow) compared to control mice. Application of 0.5 μM D-APV during the period of tetanic stimulation (10 minutes before and after tetanic stimulation) partially reversed the decrease in LTP in Tsc1^GFAPCKO mice. (C) Average potentiation ratio (% of baseline field EPSP slope) with LTP in control slices, untreated Tsc1^GFAPCKO slices, and Tsc1^GFAPCKO slices treated with 0.5 μM D-APV was significantly reduced in untreated Tsc1^GFAPCKO slices (n = 6-8 slices per group; *p<0.05 by ANOVA).

Extracellular glutamate is elevated in the hippocampus of Tsc1^GFAPCKO mice in vivo (A) Extracellular glutamate concentrations were measured from control and Tsc1^GFAPCKO mice with microdialysis at different flow rates. Fitted polynomial curves show the dependence of the measured dialysate glutamate concentration on flow rate for each group and allow extrapolation of the absolute ECF glutamate concentration by the extrapolated zero flow method (See Methods). (B) Average in vivo ECF glutamate concentration of control and Tsc1^GFAPCKO mice was determined based on the extrapolated ECF glutamate concentration calculated individually for each mouse. Tsc1^GFAPCKO mice had significantly elevated ECF glutamate concentrations compared to controls (n=6 mice per group; *p<0.05). (C) Recovery percentage of glutamate at various flow rates was not significantly different between control and KO mice, indicating that the microdialysis technique and other biological factors were consistent between the two groups. (D) Representative microdialysis probe site in the hippocampus of 4 week old mice.

Tsc1^GFAPCKO mice exhibit deficits in contextual fear conditioning and spatial learning. (A) Although Tsc1^GFAPCKO mice tended to freeze more often than controls when the mice were first placed in the training chamber (baseline; Day 1), differences were not significant. However, control mice froze significantly (p = 0.035) more often following the tone (T)/shock(S) pairings with significant (*) differences being observed following the third T/S pairing, suggesting that Tsc1^GFAPCKO mice may have had short-term context deficits. †p = 0.024; *p = 0.007. (B) Tsc1^GFAPCKO mice showed profound deficits in contextual fear conditioning (Day 2) when tested 24 h after T/S training with differences being greatest during the first 4 min. of the testing (*p < 0.003; †p = 0.024). In addition, control mice showed habituation of the freezing response after peaking at 2 min (min 2 vs min 8; **p = 0.001), while Tsc1^GFAPCKO showed a potentiation of freezing over most of the session [min 1 vs min 7 (peak); ††p = 0.025]. (C) Tsc1^GFAPCKO and control mice showed similar levels of freezing when first introduced into the altered environment (baseline, Day 3) of the new chamber used to test auditory cue conditioning. However, control mice showed significantly greater levels of freezing following the onset of the tone compared to the Tsc1^GFAPCKO group (*p < 0.0005; †p = 0.009). Again, control mice showed significant habituation of the freezing response (Block 1 vs Block 8: **p = 0.006), while the Tsc1^GFAPCKO mice exhibited a significant potentiation of freezing across most of the test session [Block 1 vs Block 7 (peak freezing: ††p = 0.044). (D) Control and Tsc1^GFAPCKO mice did not differ in the shock levels which elicited various behaviors (flinching, running, vocalizing, jumping) indicating that Tsc1^GFAPCKO mice did not have alterations in shock sensitivity which could account for differences in conditioning. (E) No differences between groups were observed in terms of path length during cued trials in the water maze. (F) However, Tsc1^GFAPCKO showed a significant deficit in spatial learning acquisition during place trials in the water maze where controls had significantly shorter path lengths to the escape platform across the blocks of trials with differences being greatest during the second (*p = 0.004) and eighth blocks (†p = 0.049) of trials. Also, control mice showed significant improvement between the first and last block of trials (**p = 0.024) suggesting learning had occurred while Tsc1^GFAPCKO mice showed no such improvement. (G) Tsc1^GFAPCKO swam significantly slower than controls during the place trials, thus making path length to the submerged platform a more appropriate dependent variable than latency to assess acquisition performance during the place trials. Differences between groups were greatest during the sixth and tenth blocks (*p < 0.009; †p < 0.029). (H) Retention data collected during the probe trial suggested that control mice, but not Tsc1^GFAPCKO mice, displayed spatial memory by spending more time in the target quadrant, which had contained the submerged platform, than in the quadrants to the left (†p = 0.044), right (*p = 0.001), or opposite (†p = 0.022) of the target quadrant. (I) The pole test was the only measure out of all activity and sensorimotor variables where significant differences were found between Tsc1^GFAPCKO and control mice. Specifically, Tsc1^GFAPCKO mice were found to take significantly longer to climb down the pole (p = 0.002).