An adaptive role of TNFα in the regulation of striatal synapses

Abstract

Elevation of inflammatory cytokines in the striatum precedes symptoms in a number of motor dysfunctions, but it is unclear whether this is part of the disease process or an adaptive response to the pathology. In pyramidal cells, TNFα drives the insertion of AMPA-type glutamate receptors into synapses, and contributes to the homeostatic regulation of circuit activity in the developing neocortex. Here we demonstrate that in the mouse dorsolateral striatum, TNFα drives the internalization of AMPARs and reduces corticostriatal synaptic strength, dephosphorylates DARPP-32 and GluA1, and results in a preferential removal of Ca(2+)-permeable AMPARs. Striatal TNFα signaling appears to be adaptive in nature, as TNFα is upregulated in response to the prolonged blockade of D2 dopamine receptors and is necessary to reduce the expression of extrapyramidal symptoms induced by chronic haloperidol treatment. These data indicate that TNFα is a regulator of glutamatergic synaptic strength in the adult striatum in a manner distinct from its regulation of synapses on pyramidal cells and mediates an adaptive response during pathological conditions.

Figure 1.

TNFα regulates AMPA receptor content at striatal synapses. A, Representative traces from striatal (top) medium spiny neurons and from CA1 pyramidal neurons (bottom) with the cell held at −70 mV and +40 mV from control slices (left), and TNFα-treated slices (100 ng/ml; 1–2 h) (right). B, Group data showing that TNFα significantly reduced the ratio of AMPA to NMDA current amplitudes in striatal cells (n = 13–15; p = 0.0002), whereas in CA1 pyramidal neurons TNFα significantly enhanced the ratio of AMPA to NMDA current amplitudes (n = 10; p = 0.035). AMPA/NMDA ratios were calculated using the peak amplitude at −70 mV for AMPA and the amplitude at +40 mV taken 40 ms after the peak at −70 mV. C, Representative recording of sEPSCs from control slices and slices treated with 100 ng/ml TNFα. D, Mean sEPSC amplitude is significantly decreased by TNFα treatment (17.7 ± 0.7 pA vs 15.7 ± 0.5 pA; n = 14–15; p = 0.03). E, Cumulative distributions of sEPSC amplitudes from cells shown in C. F, Sample traces of isolated AMPAR EPSCs at +40 and −70 mV. G, Group data demonstrating that synapses from TNFα-treated slices have significantly less rectification than untreated slices. The RI is expressed as EPSC amplitude_{+40 mV}/EPSC amplitude_{−70 mV}, with a larger ratio indicating less rectification (n = 10–12; p = 0.002). H, Sample traces of isolated AMPAR EPSCs at −70 mV before and after IEM-1460 perfusion. I, Group data demonstrating that EPSC ratio is significantly reduced in control slices after IEM-1460 compared with TNFα-treated slices (n = 6). p < 0.05. Data are mean ± SEM. *p < 0.05. **p < 0.01. ***p < 0.001.

Figure 2.

TNFα treatment significantly reduces surface AMPAR levels in the striatum. Striatal slices were treated with TNFα 100 ng/ml for 1 h and analyzed by surface biotinylation for (A) surface GluA1 (n = 10 mice), surface GluA2 (n = 7 mice) and surface GluN1 (n = 8 mice) and were compared with the total corresponding protein level. The protein loading for surface exceeds the loading for total in some cases. Only the decrease in AMPAR subunit surface expression was significant (GluA1: p = 0.002; GluA2: p = 0.02; GluN1: p = 0.96). B, TNFα treatment significantly decreased the phosphorylation of DARPP-32 T34 (n = 7 mice), and both GluA1–S831 (n = 7 mice) and GluA1–S845 (n = 8 mice) (p-T34: p = 0.02; p-S831: p = 0.0011; p-S845: p = 0.001), with no change in total protein expression of either DARPP-32 or GluA1 (GluA1: p = 0.96; DARPP-32: p = 0.17). Data are mean ± SEM. *p < 0.05. **p < 0.01.

Figure 3.

Blocking TNFα signaling increases AMPA receptor content at striatal synapses. A, Representative traces from striatal medium spiny neurons with the cell held at −70 mV (bottom) and +40 mV (top) in slices from WT mice (left), TNFα^−/− mice (TNFKO) (middle), and WT slices treated with a dominant-negative form of TNFα (DN-TNF) (right). B, Group data showing that, in TNFα^−/− mice and WT slices treated with DN-TNF, there is a significantly higher ratio of AMPA to NMDA current amplitudes. AMPA/NMDA ratios were calculated using the peak amplitude at −70 mV for AMPA and the amplitude at +40 mV taken 40 ms after the peak at −70 mV. A one-way ANOVA demonstrated a significant effect of the treatment (n = 15–17 per group; F_(2,48) = 3.9, p < 0.026). C, Representative recordings of sEPSCs from control slices and slices treated with DN-TNF. D, Mean sEPSC amplitude is significantly increased by DN-TNF treatment (18.2 ± 0.6 pA vs 15.3 ± 0.4 pA; n = 14; p = 0.001). E, Cumulative distributions of sEPSC amplitudes. F, Sample traces of isolated AMPAR EPSCs at +40 and −70 mV. G, Group data demonstrating that synapses from DN-TNF-treated slices have increased rectification relative to untreated slices. The RI is expressed as EPSC amplitude_{+40 mV}/EPSC amplitude_{−70 mV}, with a larger ratio indicating less rectification (n = 10–12). p = 0.005. Data are mean ± SEM. *p < 0.05. **p < 0.01.

Figure 4.

Loss of TNFα signaling exacerbates the extrapyramidal symptoms caused by chronic haloperidol. A, Effects of chronic haloperidol on the development of VCMs in TNFα^−/− and WT mice. VCMs were generated by a single administration of long-acting haloperidol decanoate (50 mg/kg, i.m., equivalent of 2.5 mg/kg/d) and evaluated weekly (Ethier et al., 2004). TNFα^−/− mice had significantly more VCMs than WT mice (n = 8 per group). Two-way repeated-measures ANOVA: genotype, F_(1,14) = 1.14, p = 0.001; days, F_(4,11) = 14.9, p < 0.0001; genotype × days, F_(4,11) = 2.54, p = 0.0047. A small group of mice were monitored for an extended period and checked for VCM at 55 and 69 d after haloperidol injection. TNFα^−/− mice still had significant levels of VCMs, whereas WT mice had returned to baseline levels (n = 4 per group). Two-way repeated-measures ANOVA: genotype, F_(1,6) = 6.36, p = 0.0008; days, F_(1,6) = 0.12, p = 0.42; genotype × days, F_(1,6) = 0.0007, p = 0.95. B, Effects of acute haloperidol on catalepsy. WT or TNFα^−/− mice received an intraperitoneal injection of haloperidol (0.1 mg/kg) and were tested 15, 30, 60, and 90 min later. No significant differences in cataleptic behavior were observed between the different strains. Two-way repeated-measures ANOVA: genotype, F_(1,16) = 0.005, p = 0.76; time, F_(3,14) = 7.68, p < 0.0001; genotype × time, F_(3,14) = 0.09, p = 0.73. C, Effects of chronic haloperidol on VCMs in WT mice treated with DN-TNF. Top, Schematic of the experimental design. Mice were injected with DN-TNF (30 mg/kg initial dose; 15 mg/kg every 2 d thereafter) before baseline VCM monitoring and haloperidol injection. Bottom, DN-TNF treatment significantly increased VCMs. Two-way repeated-measures ANOVA: treatment, F_(1,14) = 0.83, p = 0.004; days, F_(2,13) = 5.33, p < 0.001; treatment × days, F_(2,13) = 1.11, p = 0.007. D, Effect of acute blockade of TNFα signaling on a previously established dyskinesia. Top, Schematic of the experimental design. Mice were injected with haloperidol and evaluated 14 d later for VCMs; mice were then injected with DN-TNF (30 mg/kg) and reevaluated 24 h later. Bottom, Acute DN-TNF treatment significantly increased VCMs. Paired t test (p = 0.0003; n = 5). Each data point represents VCM scores, mean ± SEM. *p < 0.05. **p < 0.01. ***p < 0.001.

Figure 5.

TNFα treatment reduces DARPP-32 phosphorylation in the striatum after haloperidol treatment. A, Striatal slices were treated with TNFα 100 ng/ml for 1 h and analyzed by surface biotinylation for dopamine D2 receptors (n = 7 mice). No significant difference was observed (p = 0.35). B, Striatal slices from TNFα^−/− mice were compared with WT mice for the surface expression of D2 receptors (n = 7 mice per group). No significant difference was observed (p = 0.68). C, Mice were treated with haloperidol for 2 weeks and analyzed for total DARPP-32 and phospho-DARPP-32 T34 in the striatum. Two weeks of haloperidol significantly reduced striatal Thr34 phosphorylation compared with untreated control striatum; however, 24 h treatment with DN-TNF in haloperidol-treated mice restored Thr34 phosphorylation to control levels (n = 4 mice per group). Data are mean ± SEM. A, B, *p < 0.05, versus control (Mann–Whitney test). C, *p < 0.05, versus control (one-way ANOVA with a post hoc Kruskal–Wallis test).

Figure 6.

TNFα signaling limits the abnormal synaptic and behavioral alterations induced by chronic haloperidol. A, Representative traces from striatal medium spiny neurons held at −70 mV (bottom) and +40 mV (top) 14 d after haloperidol injection in WT mice or TNFα^−/− mice. B, Group data showing that haloperidol significantly increased the ratio of AMPA to NMDA current amplitudes in TNFα^−/− compared with naive mice, but this effect was not significant in WT (n = 15 or 16 per group). A one-way ANOVA indicates a significant difference between the groups (F_(3,61) = 9.52, p < 0.0001). C, Group data demonstrating that synapses from slices obtained from mice after 14 d of haloperidol have a significantly higher rectification. The RI is expressed as EPSC amplitude_{+40 mV}/EPSC amplitude_{−70 mV}, with a larger ratio indicating less rectification (n = 9 haloperidol, 13 control; p = 0.015). D, Effects of acute blockade of CP-AMPAR on VCMs at day 14 after haloperidol. WT mice were injected with haloperidol. After 14 d VCMs were counted and mice were injected with IEM-1460 (3 mg/kg) or saline. At 20 min later, VCMs were counted again. IEM-1460 significantly reduced VCM levels (paired t test; p = 0.02; n = 8). E, In contrast, mice treated with saline had significantly higher levels of VCMs after injection (paired t test; p = 0.04; n = 7). F, Striatal slices from TNFα^−/− mice chronically treated with haloperidol for 14 d were treated ex vivo with 100 ng/ml TNFα for 1–2 h. TNFα treatment significantly reduced the AMPA/NMDA ratio (n = 11–13; p = 0.0017). G, Effects of chronic haloperidol on VCMs after stopping treatment with dominant-negative TNFα. WT mice were administered with DN-TNF either for 2 weeks or just for the first week and compared with untreated mice. Stopping the treatment of DN-TNF reduced the level of VCMs to that of control mice (n = 9–11 mice per group). Two-way repeated-measures ANOVA: treatment, F_(2,26) = 0.05, p = 0.0048; days, F_(1,26) = 0.17, p = 0.042; treatment × days, F_(2,26) = 0.28, p = 0.038. Data are mean ± SEM. *p < 0.05. **p < 0.01.