Mitf Links Neuronal Activity and Long-Term Homeostatic Intrinsic Plasticity

Abstract

Neuroplasticity forms the basis for neuronal circuit complexity and differences between otherwise similar circuits. We show that the microphthalmia-associated transcription factor (Mitf) plays a central role in intrinsic plasticity of olfactory bulb (OB) projection neurons. Mitral and tufted (M/T) neurons from Mitf mutant mice are hyperexcitable, have a reduced A-type potassium current (I_A) and exhibit reduced expression of Kcnd3, which encodes a potassium voltage-gated channel subunit (Kv4.3) important for generating the I_A Furthermore, expression of the Mitf and Kcnd3 genes is activity dependent in OB projection neurons and the MITF protein activates expression from Kcnd3 regulatory elements. Moreover, Mitf mutant mice have changes in olfactory habituation and have increased habituation for an odorant following long-term exposure, indicating that regulation of Kcnd3 is pivotal for long-term olfactory adaptation. Our findings show that Mitf acts as a direct regulator of intrinsic homeostatic feedback and links neuronal activity, transcriptional changes and neuronal function.

Keywords: Kcnd3; genetics; hyperactivity; intrinsic plasticity; potassium channel; transcription.

Figure 1.

Mitf mutant mice have an increase in Tbr+ neurons in the glomerular layer. A, RNA in situ hybridization of Mitf (red) and Tbr2 (green) in glomeruli and MC/GCL of wild-type and Mitf^{mi-vga9/mi-vga9} mice. Scale bars: 50 μm. B, Mitf count per Tbr2-positive cells below glomeruli and MCL. N = 9 per genotype. C, Mitf mRNA expression, determined by RT-qPCR. N = 6 per genotype. D, Immunofluorescent staining of Tbr2/Eomes (Tbr2+) neurons (green) in the OBs of wild-type, Mitf^mi-vga9/+, and Mitf^{mi-vga9/mi-vga9} mice. E, Cell count of Tbr2/Eomes (Tbr2+) cells in the glomeruli and MCL of wild-type, Mitf^mi-vga9/+, and Mitf^{mi-vga9/mi-vga9} mice. N = 6 per genotype. The values on the graphs are mean ± SEM DAPI nuclear staining is shown in blue. Scale bars: 100 μm; p values were calculated using two-way ANOVA (B, E) or two-tailed unpaired Student’s t test (C); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 2.

Mitf^{mi-vga9/mi-vga9} mice have normal OBs. A, Representative images of the H&E histologic analysis of coronal sections of the medial OBs of the indicated genotypes. B, Immunofluorescent staining of TH showing the localization of TH in the glomeruli of the indicated genotypes. C, Cell count of TH+ cells of the OB of wild-type, Mitf^mi-vga9/+, and Mitf^{mi-vga9/mi-vga9} mice. N = 6 per genotype. The values on the graphs are mean ± SEM DAPI nuclear staining is shown in blue. Scale bar: 100 μm; p values were calculated using one-way ANOVA (C).

Figure 3.

Increase in apoptosis in OBs of Mitf^{mi-vga9/mi-vga9} mice. A, TUNEL staining of sections of OBs of the indicated genotypes, showing localization of apoptotic cells. B, The number of apoptotic cells of OB of wild-type and Mitf^{mi-vga9/mi-vga9} mice. N = 9 per genotype. The values on the graphs are mean ± SEM DAPI nuclear staining is shown in blue. Scale bars: 50 μm; p values were calculated using two-tailed unpaired Student’s t test; *p < 0.05.

Figure 4.

Mitf mutant mice have a decrease in I_A and a concomitant increase in MC activity. A, Representative images of voltage clamp recordings. B, Representative images of voltage clamp recordings, where the I_A was inactivated by a –40-mV prepulse. C, The isolated I_A recorded from wild-type and Mitf^{mi-vga9/mi-vga9} M/T neurons. D, I_{A max} and I_{DR max} in wild type and Mitf^{mi-vga9/mi-vga9}. N = 12 per genotype. E, The relation between current density (pA/pF) of isolated I_A currents and membrane voltage (mV) of wild-type and Mitf^{mi-vga9/mi-vga9} M/T neurons. N = 12 per genotype. F, Representative recordings of spontaneous action potentials observed in wild-type (N = 5) and Mitf^{mi-vga9/mi-vga9} M/T neurons (N = 3) by current clamp with examples from both recorded traces enlarged on the right. G, RNA in situ hybridization of c-Fos (green) and Mitf (red) in wild-type and Mitf^{mi-vga9/mi-vga9} OBs. H, c-Fos dots per cell in the ETC, MC, and GC. N = 4 per genotype. The values on the graphs are mean ± SEM DAPI nuclear staining is shown in blue. Scale bars: 50 μm; p values were calculated using two-tailed unpaired Student’s t test (D), nonlinear regression and one-way ANOVA (F) and two-way ANOVA (H) and *p < 0.05.

Figure 5.

Gating kinetics of the A-type current (I_A). A, Representative recordings obtained to assess the gating kinetics of the I_A current mediating channels. The projection neurons were first voltage clamped at –30 mV and subsequently depolarized to 20 mV using a prehyperpolarizing pulse in between to recover the I_A current from inactivation. This was done by increasing the time, Δt, spent in the hyperpolarized state, as indicated by the voltage trace on the left. To gauge its activation kinetics, the neurons were held at –90 mV, as shown in the voltage trace on the right, and subsequently depolarized to 20 mV, with a predepolarization pulse to –30 mV applied in between. The A-type current gets inactivated by increasing the duration of the prepulse, Δt. B, Normalized mean currents (±SEM) plotted separately during recovery of the I_A current from inactivation and during deactivation, in cells from wild-type and Mitf^{mi-vga9/mi-vga9} mice. The functions fitted to the data yield half-activation and half-recovery times of 15.2 ± 4.7 and 42.6 ± 7.2 ms for the wild-type (N = 5, p = 0.53) and 18.2 ± 1.5 and 51.3 ± 9.3 ms for the Mitf^{mi-vga9/mi-vga9} (N = 6, p = 0.49) M/T.

Figure 6.

Kcnd3 and Mitf expression are activity dependent in the OB. A, RNA in situ hybridization of Tbr2 (green), Mitf (red), and Kcnd3 (magenta) performed on wild-type and Mitf^{mi-vga9/mi-vga9}OBs. B, Kcnd3 count per cell. N = 5 per genotype. C, Schematic representation of the amyl acetate (AA) experiment where mice were habituated to an odorless cage for 60 min. Upon exposure to AA for 30 min (red block), they were sacrificed at different time points (black vertical lines). D, Mitf and c-Fos mRNA expression in wild-type OB following AA, determined by RT-qPCR. N = 6 per time point. E, c-Fos dots per cell in wild-type OB mice following AA. N = 6 per time point. F, Mitf dots per cell in wild-type OB following AA. N = 6 per time point. G, Kcnd3 dots per cell in wild-type OB following AA. N = 6 per time point. The values on the graphs are mean ± SEM DAPI nuclear staining is shown in blue. Scale bars: 50 μm; p values were calculated using two-way ANOVA (B, D, F, F, I); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. PG = periglomerular.

Figure 7.

Kcnd2 expression is increased in the Mitf mutant OB. A, RNA in situ hybridization of Tbr2 (green), Mitf (red), and Kcnd2 (magenta) performed on wild-type and Mitf^{mi-vga9/mi-vga9}OBs. B, Kcnd3 count per cell. N = 4 per genotype. The values on the graphs are mean ± SEM DAPI nuclear staining is shown in blue. Scale bars: 50 μm; p values were calculated using two-way ANOVA (B); *p < 0.05.

Figure 8.

c-Fos and Mitf expression is increased upon activity induction in the OB. RNA in situ hybridization of c-Fos (red) and Mitf (green) performed on wild-type OBs following AA. DAPI nuclear staining is shown in blue. Scale bars: 50 μm.

Figure 9.

Kcnd3 expression is increased on activity induction in the OB. RNA in situ hybridization of Mitf (red) and Kcnd3 (green) performed on wild-type OBs following AA. DAPI nuclear staining is shown in blue. Scale bars: 50 μm.

Figure 10.

Activity-dependent increase in Kcnd3 expression requires MITF enhancer activity. A, RNA in situ hybridization of c-Fos (red) and Kcnd3 (green) performed in Mitf^{mi-vga9/mi-vga9} OB following AA. B, Kcnd3 dots per cell in Mitf^{mi-vga9/mi-vga9} OB following AA. N = 3–5 per time point. C, ChIPseq peaks of MITF, H3K27ac and H3K4me1 on KCND3 gene in 501mel cells. The MITF peaks are labeled A (orange) and B (yellow). Yellow shows overlapping peak of MITF binding in Strub and Laurette datasets (B), whereas orange indicates MITF peaks that overlap with H3K27ac and H3K4me1 peaks (A). D, Sequence of the basic region of wild-type (MITF-M) and transcriptionally inactive MITF with four argenines mutated to alanines (MITF-M B4RA). E, Transcription activation assays performed in HEK293T cells co-transfected with constructs containing the hTYR, hKCND3_pA, and hKCND3_pB regulatory regions fused to luciferase, together with empty vector or wild-type or transcriptionally inactive (B4RA) MITF constructs. N = 3. F, Transcription activation assays performed in N2A cells co-transfected with constructs containing the hTYR, hKCND3_pA, and hKCND3_pB regulatory regions fused to luciferase, together with empty vector or wild-type or transcriptionally inactive (B4RA) MITF constructs. N = 3. The values on the graphs are mean ± SEM DAPI nuclear staining is shown in blue. Scale bars: 50 μm; p values were calculated using two-way ANOVA (B, E, F); *p < 0.05, ***p < 0.001, ****p < 0.0001.

Figure 11.

Mitf regulates olfactory discriminatory ability and odor fatigue recovery. A, Quantification of the results of the hidden cereal odor detection-assay. N = 8 per genotype. B, Avoidance assay where a cage was sectioned in three equal areas and mice were exposed to water or propionic acid (PA) in section 1. Quantification shows time spent in each section (Table 2). N = 6–8 per genotype. C, Schematic overview of the dishabituation-habituation assay, where mice are exposed to odor A (blue) for 30 s six times with 5-min intervals, after which they are exposed to odor B (red) for 30 s. D, Dishabituation-habituation, mice were exposed to vanilla as odor A and almond as odor B. N = 8 per genotype. E, Dishabituation-habituation, where mice were exposed to almond as odor A and vanilla as odor B. N = 6–8 per genotype. F, Dishabituation-habituation, mice were exposed to almond as odor A and lime as odor B. N = 8 per genotype. G, Dishabituation-habituation assay where the mice were exposed to vanilla for 2 h, followed by almond as odor A and vanilla as odor B. N = 15 per genotype. The values on the graphs are mean ± SEM; p values were calculated using one-way ANOVA (A) or two-way ANOVA (B, D–G); *p < 0.05, **p < 0.01, ****p < 0.0001.

Figure 12.

Model of MITF as a regulator of homeostatic intrinsic plasticity. Upon neuronal activity in the OB, MITF is required for an increase in Kcnd3 expression. More Kv4.3 protein results in an increase in the I_A potassium current and reduced likelihood of action potentials, subsequently leading to decreased activity of M/T cells.