Reduced juvenile long-term depression in tuberous sclerosis complex is mitigated in adults by compensatory recruitment of mGluR5 and Erk signaling

Abstract

Tuberous sclerosis complex (TSC) is a multisystem genetic disease that manifests with mental retardation, tumor formation, autism, and epilepsy. Heightened signaling through the mammalian target of rapamycin (mTOR) pathway is involved in TSC pathology, however it remains unclear how other signaling pathways are perturbed and contribute to disease symptoms. Reduced long-term depression (LTD) was recently reported in TSC mutant mice. We find that although reduced LTD is a feature of the juvenile mutant hippocampus, heightened expression of metabotropic glutamate receptor 5 and constitutively activated Erk signaling in the adult hippocampus drives wild-type levels of LTD. Increased mGluR5 and Erk results in a novel mTOR-independent LTD in CA1 hippocampus of adult mice, and contributes to the development of epileptiform bursting activity in the TSC2(+/-) CA3 region of the hippocampus. Inhibition of mGluR5 or Erk signaling restores appropriate mTOR-dependence to LTD, and significantly reduces epileptiform bursting in TSC2(+/-) hippocampal slices. We also report that adult TSC2(+/-) mice exhibit a subtle perseverative behavioral phenotype that is eliminated by mGluR5 antagonism. These findings highlight the potential of modulating the mGluR5-Erk pathway in a developmental stage-specific manner to treat TSC.

Figure 1. TSC2^+/− adult hippocampal slices are unaffected by mTORC1 inhibition.

(a) Juvenile (21 d) TSC2^+/− hippocampal slices display reduced LTD [79.6±3.1%, n = 16(5)] [two-way ANOVA; F(1, 26) = 5.72, p = 0.024] compared to WT [70.0±4.4%, n = 9(4)]. Rapamycin (20 nM) increased LTD magnitude in TSC2^+/− slices to WT levels [68.0±3.9%, n = 14(5)]. (b) Rapamycin does not impact juvenile WT LTD [n = 9(3)]. (c–g) Slices from mature TSC2^+/− (8–12 wk). (c) Mature TSC2^+/− [n = 12(4)] show no difference in LTD magnitude from littermate WT [n = 16(5)] controls. (d) Rapamycin (20 nM) reduces mGluR-LTD in mature WT littermates [84.4±5.4%, n = 8(4)] [two-way ANOVA; F(1, 18) = 17.4, p = 0.0006]. (e) Rapamycin (20 nM) does not affect mGluR-LTD in TSC2^+/− hippocampus [70.8±5.4%, n = 9(5)]. (f) Metformin (5 µM) has no effect on mGluR-LTD in TSC2^+/− hippocampus [63.4±8.0%, n = 7(3)]. (g) Metformin (5 µM) significantly reduces mTOR-dependent mGluR-LTD in WT slices (metformin, 86.6±4.6%; control, 51.2±5.2%) [two-way ANOVA; F(1, 13) = 15.65, p = 0.0016], an effect that is reversed with the AMPK inhibitor Ara-A (100 µM) [66.4±6.6%, n = 8(4)]. (h) Inhibition of protein synthesis with anisomycin eliminates mGluR-LTD in TSC2^+/− hippocampus [96.5±2.5%, n = 6(3)] [two-way ANOVA; F(1, 13) = 37.62, p<0.0001]. Representative traces: solid line is 4 min before DHPG ‘1’, dashed line is at the end of the recording ‘2’.

Figure 2. Rapamycin impacts mTORC1 assembly and signaling.

(a, b) Rapamycin treatment significantly reduces RAPTOR-mTORC1 association in DHPG-treated WT and TSC2^+/− slices (Student t test; WT n = 4, p = 0.030; TSC2^+/− n = 4, p = 0.0043). (c, d) Rapamycin significantly reduces p-S6 levels in WT and TSC2^+/− hippocampal slices (Student t test; WT n = 22, p = 0.031; TSC2^+/− n = 18, p = 0.016). Representative traces: solid line is 4 min before DHPG ‘1’, dashed line is at the end of the recording ‘2’. WCE-whole cell extract. *p<0.05, **p<0.005.

Figure 3. mGluR5 and Erk1/2 signaling is heightened in adult TSC2^+/− mice.

(a) Acutely harvested hippocampal tissue displays increased mGluR5 expression in TSC2^+/− adult mice compared to adult WT mice (one-way ANOVA, F(1,70) = 5.536, p = 0.0018). There is enhanced Erk1/2 (T202/Y204) phosphorylation in adult TSC2^+/− hippocampus compared to WT (one-way ANOVA, F(1,74) = 5.859, p = 0.0012). There is no observable difference in phosphorylated Akt (T308) between genotypes, however there was a highly significant reduction with age (one-way ANOVA, F(1,41) = 6.873, p = 0.0007). Western quantifications are displayed in (b). (c) MPEP preincubation (40 µM, present −40 to −20 min), followed by washout does not impact mGluR-LTD magnitude in TSC2^+/− hippocampus [MPEP, 55.6±5.7%, n = 6(3)]. (d) MPEP preincubation restores rapamycin (20 nM, −10 to 20 min) sensitivity [MPEP+rapamycin, 83.5±5.4%, n = 8(3); two-way ANOVA; F(1,15) = 21.06, p = 0.0004] and metformin sensitivity of mGluR-LTD [metformin+rapamycin, 81.4±4.5%, n = 7(3); two-way ANOVA; F(1, 14) = 17.15, p = 0.0010]. (e) MPEP (40 µM) incubation during LTD induction (present −20 to 20 min) eliminates mGluR-LTD in TSC2^+/− hippocampus [99.7±2.7%, n = 6(3)]. (f) In WT and TSC2^+/− slices, DHPG (50 µM, 10 min) significantly increases phospho-Erk1/2 (one-way ANOVA, F(1,55) = 13.1, p<0.0001). In WT slices, DHPG increased phospho-Erk1/2 levels (one-way ANOVA, F(1,30) = 10.28, p = 0.0004). MPEP preincubation, followed by DHPG treatment, causes a significant reduction in p-Akt (one-way ANOVA, F(1,60) = 9.402, p<0.005), and p-Erk1/2 (one-way ANOVA, F(1,55) = 13.1, p<0.0001) levels in TSC2^+/− hippocampus. (h) 40 µM MPEP (20 min) alone does not impact downstream signaling to Akt, Erk, or rpS6 in WT or TSC2^+/− hippocampus. Representative traces: solid line is 4 min before DHPG ‘1’, dashed line is at the end of the recording ‘2’. *p<0.05, **p<0.005.

Figure 4. Inhibition of Mek-Erk1/2 signaling, but not PI3K-Akt, reduces mGluR-LTD in TSC2^+/− hippocampus.

(a) Incubation with the PI3K antagonists LY294002 (10 µM) or wortmannin (500 nM) reduce mGluR-LTD in WT slices [LY294002, 89.9±2.4%, n = 9(3); wortmannin, 79.5±3.6%, n = 13(3); control, n = 16(4)]. (b) LY294002 has no effect in TSC2^+/− hippocampus [49.6±5.3%, n = 6(3)], nor does it restore rapamycin sensitivity [67.2±1.4%, n = 6(3)]. (c) Likewise, wortmannin (500 nM) does not impact mGluR-LTD in TSC2^+/− hippocampus [70.4±4.3%, n = 7(3)], nor does it restore rapamycin sensitivity [57.4±3.5%, n = 7(3)]. (d) Analysis of p-Erk levels in response to a U0126 concentration series indicates that 200 nM U0126 reduces p-Erk1/2 to untreated WT levels. (e) Application of low-dose (200 nM) U0126 does not impact LTD magnitude in WT slices [61.8±5.2%, n = 7(3)]. (f) Subtle reduction of Mek-Erk signaling with 200 nM U0126 (−20 to 20 min) does not impact mGluR-LTD magnitude [57.3±4.3%, n = 8(3)], yet does restore rapamycin sensitivity [80.4±3.9%, n = 7(3), p = 0.0054). (g) 200 nM U0126 (20 min) reduces Erk signaling, but does not impact signaling to Akt or rpS6 in WT or TSC2^+/− hippocampus. (h) In the presence of DHPG, U0126 blocked activation of rpS6 (Student t test; n = 18, p = 0.016). Rapamycin plus U0126 produced an even greater reduction in p-S6 (Student t test; n = 15, p = 0.0005) (untreated and DHPG-treated quantified data are reproduced from Figure 1). Representative traces: solid line is 4 min before DHPG ‘1’, dashed line is at the end of the recording ‘2’.

Figure 5. Antagonism of mGluR5 or Erk signaling reduces induction of epileptiform activity in TSC2^+/− CA3 hippocampus.

(a) Representative traces of WT and TSC2^+/− slices following DHPG treatment (50 µM, 30 min) and in the presence and absence of drug. (b) Ictal burst duration is greater in TSC2^+/− hippocampus compared to WT (Student t-test, p = 0.0020) and is significantly reduced with MPEP (p<0.005) and U0126 (p<0.0005). (c) TSC2^+/− slices develop significantly more long-duration bursts in CA3 neurons than WT slices (Chi-square test for trend, p<0.0001). (d) Antagonism of mGluR5 (40 µM MPEP) or Mek-Erk (20 µM U0126) signaling during DHPG treatment significantly reduces epileptiform bursting in TSC2^+/− slices (TSC2 versus TSC2+MPEP Chi-square, p = 0.027; TSC2 versus TSC2+U0126 Chi-square, p = 0.041). WT bursting was not significantly different with MPEP (Chi-square, p = 0.871). (e) Co-incubation of MPEP and DHPG in WT slices produced an insignificant trend toward less bursting in WT slices (Chi-square for trend, p = 0.182). (f) Rapamycin (20 nM) did not impact DHPG-induced bursting in WT slices. (g) MPEP (40 µM) treatment during DHPG incubation significantly reduces CA3 bursting in TSC2^+/− slices (Chi-square for trend, p<0.0001). (h) U0126 (20 µM) significantly reduces CA3 bursting profile in TSC2^+/− slices (Chi-square for trend, p<0.0001). (i) Rapamycin (20 nM) did not impact the bursting profile in TSC2^+/− slices (Chi-square for trend, p = 0.211). (j) Cumulative probability of developing long duration ictal activity is greatest in TSC2^+/− slices and decreases with MPEP and even more so with U0126. *p<0.05, **p<0.005, ***p<0.0005.

Figure 6. MPEP eliminates a perseverative behavior in TSC2^+/− mice.

(a) During reversal training, TSC2^+/− mice make significantly more errors toward the conditioned arm in the RAWM (two-way ANOVA; WT n = 21; TSC2 n = 23; F(3, 43) = 9.160, p = 0.0034). Differences between individual groups were assessed with paired Mann Whitney t test: Trial 1, WT versus TSC2; p = 0.049. Trial 2, WT versus TSC2, p = 0.010; TSC2 versus TSC2+MPEP; p = 0.0466. (b) There were no differences for open field performance between genotypes (WT n = 14; TSC2 n = 12; TSC2+MPEP n = 14). There were no differences for open pool performance between genotypes. (WT n = 13; TSC2 n = 12; TSC+MPEP n = 14). NS, not significant, *p<0.05, **p<0.005.