Dopamine neuron-derived IGF-1 controls dopamine neuron firing, skill learning, and exploration

Abstract

Midbrain dopamine neurons, which can be regulated by neuropeptides and hormones, play a fundamental role in controlling cognitive processes, reward mechanisms, and motor functions. The hormonal actions of insulin-like growth factor 1 (IGF-1) produced by the liver have been well described, but the role of neuronally derived IGF-1 remains largely unexplored. We discovered that dopamine neurons secrete IGF-1 from the cell bodies following depolarization, and that IGF-1 controls release of dopamine in the ventral midbrain. In addition, conditional deletion of dopamine neuron-derived IGF-1 in adult mice leads to decrease of dopamine content in the striatum and deficits in dopamine neuron firing and causes reduced spontaneous locomotion and impairments in explorative and learning behaviors. These data identify that dopamine neuron-derived IGF-1 acts as a regulator of dopamine neurons and regulates dopamine-mediated behaviors.

Keywords: IGF-1; behavior; dopamine; firing; somatodendritic.

Fig. 1.

IGF-1 and IGF-1R are expressed in the SNc and VTA. (A) Igf1 transcripts, detected by ISH, are detected in TH-positive neurons in the SNc and VTA. Merge shows the overlap of immunofluorescence for TH and bright-field imaging for Igf1 ISH. White arrowheads point at TH-positive neurons positive for Igf1 ISH probe. Percentage of TH-positive neurons expressing Igf1 transcripts is higher in the SNc than in the VTA (n = 3 mice, P = 0.0017, t = 7.502 df = 4; two-tailed unpaired Student’s t test). (B) IGF-1 protein immunoreactivity, detected by immunofluorescence, is detectable in TH-positive neurons in the SNc and VTA. Merge shows the overlap of immunofluorescence for TH and IGF-1. White arrowheads point at neurons double-positive for TH and IGF-1. Percentage of TH-positive neurons expressing IGF-1 protein is higher in the SNc than in the VTA (n = 3 mice, P = 0.0446, t = 2.888 df = 4; two-tailed unpaired Student’s t test). (C) IGF-1R protein immunoreactivity, detected by immunofluorescence, is present in TH-positive neurons in the SNc and VTA. Merge shows the overlap of immunofluorescence for TH and IGF-1R. White arrowheads point at neurons double-positive for TH and IGF-1R, open arrowheads point at TH positive neurons, negative for IGF-1R immunoreactivity. Percentage of TH-positive neurons with IGF-1R protein is the same between SNc and VTA (n = 3 mice, P = 0.4282, t = 0.8807 df = 4; two-tailed unpaired Student’s t test). Dotted squares in the merge pictures are magnified and split for clarity. Crossed arrows show image orientation. D, dorsal; L, lateral; M, medial; V, ventral. Graphs show mean ± SEM, together with individual values. *P < 0.05. (Scale bars: 20 μm.)

Fig. 2.

IGF-1 is released in an activity-dependent manner. Images in A and B show representative immunofluorescence of mDA neurons (TH−, MAP2-positive) expressing IGF-1, following different treatments. IGF-1 intracellular fluorescence intensity in the cell bodies (dotted ROIs in the images) was quantified. (A) Graph shows mean ± SD, together with distribution of individual neuron following 10 μM sulpiride, 1 μM TTX, or 10 μM sulpiride + 1 μM TTX [CTR and 10 μM sulpiride n = 6 cultures, 1 μM TTX and 10 μM sulpiride + 1 μM TTX n = 3 cultures. P CTR vs. 10 μM sulpiride = <0.0001, P 10 μM sulpiride vs. 10 μM sulpiride + 1 μM TTX = <0.0001, P CTR vs. 1 μM TTX = 0.1520; F_{(3, 624)} = 28.04; one-way ANOVA with Tukey’s correction for multiple comparisons]. (B) Graph shows mean ± SD, together with distribution of individual neuron following 15 mM NaCl or 15 mM KCl treatments (n = 3 cultures, P ≤ 0.0001, t = 5.086 df = 219; two-tailed unpaired Student’s t test). *P < 0.05. (Scale bars: 20 μm.)

Fig. 3.

IGF-1 exposure to the ventral midbrain reduces DA release from the cell bodies of mDA neurons. FFN102 release from the cell bodies of mDA neurons was quantified over time, upon pharmacological stimulation. Panels on the left show representative FFN102 fluorescence at different time points. Dotted ROIs show examples of mDA neurons, tracked over time. Graphs show normalized FFN102 fluorescence intensity over time. (A) Fifty nanograms per milliliter of des(1-3)IGF-1 (IGF-1 receptor agonist), applied at T20 min, slows down DA release rate from the cell bodies [n = 3, P CTR vs. des(1-3)IGF-1 (T30 min, T40 min, T50 min) = <0.0001; F_{(2, 356)} = 3.006; two-way ANOVA with Sidak’s correction for multiple comparisons]. (B) One micromolar AG1024 (IGF-1 receptor antagonist), applied at T10 min, minimally increases DA release rate from the cell bodies [n = 3, P CTR vs. AG1024 (T30 min) = 0.0028; F_{(3, 858)} = 2.117; two-way ANOVA with Sidak’s correction for multiple comparisons]. Time points show mean ± 95% CI. (Scale bars: 20 μm.)

Fig. 4.

Igf1 cKO mice are characterized by reduced IGF-1 signaling in DA neurons in the SNc. Upstream (IGF-1R) and downstream (S6 ribosomal protein) effectors of IGF-1 signaling were analyzed. (A) Images show representative immunofluorescence for TH and phospho IGF-1R in the SNc of CTR and Igf1 cKO mice. Top shows phospho IGF-1R only for clarity. Phospho IGF-1R fluorescence intensity was reduced in TH-positive SNc neurons in Igf1 cKO mice, compared with CTRs (n = 4, P = 0.0331, t = 2.755 df = 6; two-tailed unpaired Student’s t test). (B) Images show representative immunofluorescence for TH and phospho S6 in the SNc of CTR and Igf1 cKO mice. Top shows phospho S6 only for clarity. Phospho S6 fluorescence intensity was reduced in TH-positive SNc neuron in Igf1 cKO mice, compared with CTRs (n = 4, P = 0.0047, t = 4.379 df = 6; two-tailed unpaired Student’s t test). Bar graphs show mean ± SEM. *P < 0.05.

Fig. 5.

Igf1 cKO mice show reduced levels of phospho Ser40 TH and are hypodopaminergic. (A) TH levels are not affected in Igf1 cKO mice. Pictures show representative image of TH immunofluorescence in SNc and VTA of CTR and Igf1 cKO mice. Graphs show TH fluorescence density (n = 6; SNc: P = 0.1984, t = 1.377 df = 10; VTA: P = 0.4535, t = 0.78 df = 10; two-tailed unpaired Student’s t test). (B) Igf1 cKO mice show reduced levels of phospho Ser40 TH in the SNc, but not in the VTA. Pictures show representative image of phospho Ser40 TH immunofluorescence in SNc and VTA of CTR and Igf1 cKO mice. Graphs show phospho Ser40 TH fluorescence density (n = 6; SNc: P = 0.0130, t = 3.014 df = 10; VTA: P = 0.6447, t = 0.4755 df = 10; two-tailed unpaired Student’s t test). (Scale bars: 20 μm.) (C) Total DA content from CTR and Igf1 cKO striata was measured and normalized vs. mg of tissue. Igf1 cKO mice are hypodopaminergic (n = 4; P = 0.0291, t = 2.852, df = 6; two-tailed unpaired Student’s t test). Graph bars show mean ± SEM, together with individual value of replicates.

Fig. 6.

IGF-1 modulates SNc DA neuron firing properties and Igf1 cKO SNc DA neurons display impaired autonomous and burst-like firing. SNc mDA neurons were identified by their sag in response to hyperpolarizing current injection, broad action potentials (>2 ms) (A), slow spontaneous firing (<4 Hz) (D), and expression of reporter fluorescence protein tdTomato (B). Representative image in B shows immunofluorescence for TH on a PFA-fixed brain slice, after patching. TH immunofluorescence colocalizes with Slc6a3^CreERT2/+- driven tdTomato expression. (Scale bar: 20 μm.) (C) IGF-1 treatment (50 ng/mL) decreased membrane resistance (MOhm) in CTR neurons. Igf1 cKO mDA neurons showed increased membrane resistance, compared with CTR, and 50 ng/mL IGF-1 treatment rescued membrane resistance values of Igf1 cKO mDA neurons [P CTR vs. Igf1 cKO = 0.0007, P CTR vs. CTR + IGF-1 = 0.0453, P Igf1 cKO vs. Igf1 cKO + IGF-1 = 0.0015; F_{(3, 36)} = 16.87]. (D) Shows representative spontaneous firing recorded from SNc DA neurons. Igf1 cKO neurons showed slower firing frequency (hertz) compared with CTR neurons. IGF-1 treatment (50 ng/mL) caused an increase in the frequency of spontaneous firing in CTR neurons and rescued the Igf1 cKO neurons’ phenotype [P CTR vs. Igf1 cKO = 0.0003, P CTR vs. CTR + IGF-1 < 0.0001, P Igf1 cKO vs. Igf1 cKO + IGF-1 = 0.0010; F_{(3, 36)} = 65.51]. (E) Igf1 cKO mDA neurons showed longer latency to trigger action potentials (milliseconds), compared with CTR. IGF-1 application (50 ng/mL) partially rescued Igf1 cKO neuron’s phenotype [P CTR vs. Igf1 cKO < 0.0001, P Igf1 cKO vs. Igf1 cKO + IGF-1 < 0.0001, P CTR vs. Igf1 cKO + IGF-1 = 0.0195; F_{(3, 30)} = 59.26]. (F) Traces show representative bursts of action potentials induced by local stimulation (gray bars). Igf1 cKO mice showed fewer action potentials (APs) per burst and a trend toward higher interspike interval within the burst, compared with CTR neurons [APs per burst: P CTR vs. Igf1 cKO < 0.0001; F_{(3, 42)} = 17.75]. Graphs show means ± SEM, together with individual values pooled from n = 3 mice for CTR and Igf1 cKO, n = 5 mice for CTR + IGF-1, and n = 4 mice for Igf1 cKO + IGF-1. Neurons responsive to 50 ng/mL IGF-1 treatments are plotted in the CTR + IGF-1 and Igf1 cKO + IGF-1 groups. Groups were compared with one-way ANOVA with Tukey’s correction for multiple comparisons. *P < 0.05.

Fig. 7.

Igf1 cKO mice show less spontaneous locomotion, impaired motor skill learning, and reduced novelty-induced exploration. (A) Representative cumulative tracks of mice in their home cage over 30 min. Igf1 cKO mice show less spontaneous movement (distance traveled, centimeters), compared with the CTRs (n = 10, nine males and one female per genotype. P = 0.0123, t = 2.784 df = 18; two-tailed unpaired Student’s t test). (B) Mice were tested on accelerating rotarod, over 3 days, and latency to fall (time, seconds) was scored. Igf1 cKO mice show impaired learning [n = 10, nine males and one female per genotype. P CTR vs. Igf1 cKO day 3 = 0.0310, day 2 = 0.1888, day 1 = >0.9999; F_{(2, 36)} = 3.477; two-way ANOVA with Bonferroni’s correction for multiple comparisons]. (C) Panel on the left shows examples of raw data from individual mice during the open-field test, as calculated between 0 min and 15 min during the task. Blue to red indicates low to high occupancy. Distance traveled (centimeters) has been calculated for 5-min time bins over a total of 30 min; Igf1 cKO mice showed reduced novelty-induced exploration compared with CTRs [n = 6; five males and one female per genotype. P CTR vs. Igf1 cKO time bin 1 = 0.0012, time bin 2 = 0.0115, time bins 3, 4, 5, and 6, P = >0.05; F_{(5, 50)} = 1.928; CTR, P time bin 1 vs. time bin 6 ≤0.0001; Igf1 cKO, P time bin 1 vs. time bin 6 = 0.7171; F_{(5, 50)} = 1.928; two-way ANOVA with Sidak’s correction for multiple comparisons]. Graphs show mean ± SEM. *P < 0.05.