The brain is composed of intricate circuits of connected neurons that communicate via a combination of electrical and chemical signals. Some signals (referred to as excitatory signals) increase the probability that the neuron receiving the chemical message will produce an electrical impulse. On the other hand, inhibitory messages decrease the likelihood of this activity. Both of these kinds of signals are fast, and act over milliseconds. There is also a diverse set of slower signals, referred to as neuromodulation, which regulates the faster signals. A signaling chemical called dopamine is involved in neuromodulation and is essential for rewarding behavior and complex motor actions. The importance of dopamine is clear from the profound lack of movement seen in individuals with Parkinson’s disease, which is caused by the death of dopamine producing brain cells.

Many nerve endings from dopamine-releasing neurons connect to a part of the brain’s reward system called the striatum. The neurons in this region are organized into two pathways that have opposing impacts on behavior. Dopamine activates different kinds of receptors called “G protein-coupled dopamine receptors” on neurons from each pathway. This allows dopamine to alter the activity of a protein called Protein Kinase A (or PKA) and alter the signaling state of these neurons.

The impact of dopamine on neural circuits in adults has been extensively studied. However it was unknown whether dopamine might influence how neural circuits are wired during brain development. Because the nerve endings from dopamine-releasing neurons reach the striatum before most excitatory connections between the neurons are formed, dopamine stands to influence the development of connections in the striatum.

Kozorovitskiy et al. have now investigated the role of neuromodulation in brain development in young mice. This involved measuring the formation of excitatory connections or synapses and the electrical activity of different striatal neurons during the maturation of brain circuits that occurs after birth. This analysis revealed that turning on dopamine receptors that increase PKA activity rapidly enhances the number of excitatory synapses on the neurons that express this receptor.

Kozorovitskiy et al. then used a variety of approaches to investigate whether there is cooperation between G protein-coupled receptors, PKA activity and a signaling molecule called glutamate in striatal development. This revealed a more general mechanism by which the activation of G-protein-coupled receptors interacts with glutamate (the primary excitatory signal sent between neurons) in order to produce new synapses. These results reveal a previously unknown role for neuromodulation in “wiring up” the brain and open the possibility of new therapies to treat neurodevelopmental and neurodegenerative disorders.

In the vertebrate basal ganglia, dopamine performs critical functions in motivated, goal-directed learning and behavior, transmitting signals related to rewarding stimuli and other salient experiences (Bromberg-Martin et al., 2010). Classical models of dopaminergic signaling in adult animals propose that the activity of dopamine-producing neurons encodes reward prediction errors and that released dopamine regulates the strength of corticostriatal excitatory glutamatergic inputs. This process is thought to favor the execution of a specific action that previously led to reward, at the expense of competing motor programs. An important element of this model is the distinct action of dopamine on the direct and indirect pathways, comprised of two classes of striatal spiny projection output neurons that express different types of dopamine receptors, project to separate downstream targets, and regulate complementary aspects of motor behavior (Bateup et al., 2010; Kravitz et al., 2010; Cui et al., 2013). Dopamine differentially regulates biochemical signaling (Nishi et al., 1999; 1997), neuronal activity and synaptic plasticity in these two classes of striatal neurons, promoting synaptic potentiation and depression through different receptor subtypes (Shen et al., 2008).

Less is known about the function of dopamine in the postnatal development of basal ganglia, when goal-oriented motor programs are first expressed and activity-dependent synapse formation occurs in the striatum. Therefore, we investigated whether dopamine modulates activity-dependent striatal plasticity and synapse formation during this time. In mice, the second postnatal week of life is a period of rapid growth of excitatory synapses on SPNs, a process driven by the release of glutamate from corticostriatal axons (Kozorovitskiy et al., 2012), among other striatal glutamatergic projections. Furthermore, the balance of activity in direct and indirect pathway SPNs (dSPNs and iSPNs, respectively) influences activity in the neocortex, potentially altering the firing patterns of corticostriatal neurons via recurrent circuit interactions (Kozorovitskiy et al., 2012; Oldenburg and Sabatini, 2015). Thus, positive feedback in recurrent circuits can amplify early glutamatergic signaling, rapidly changing network connectivity. If dopamine modulates glutamatergic synapse function in developing SPNs as it does in mature circuits, it likely interacts with glutamate-dependent neural circuit development during the early postnatal period. In dSPNs, dopamine acts on type 1 dopamine receptor (Drd1), a Gα_s-coupled GPCR to activate PKA, a critical enzyme involved in cell growth and plasticity (Nishi et al., 2000). In contrast, in iSPNs, dopamine activates type 2 dopamine receptors (Drd2), Gα_i-coupled GPCRs, decreasing PKA activity. Since PKA activity levels influence spinogenesis and synaptogenesis in cortical pyramidal neurons (Kwon and Sabatini, 2011), dopamine may directly set the threshold for activity-dependent synaptogenesis in striatal SPNs.

Dopamine is poised to regulate striatal signaling early in postnatal life, since embryonic striatum receives substantial dopaminergic innervation from substantia nigra pars compacta (SNc) and the ventral tegmental area (VTA), as is evidenced by neuroanatomical studies (Hu et al., 2004) and direct functional assays showing embryonic dopamine release (Ferrari et al., 2012). Two observations indirectly support the involvement of dopamine in the developmental wiring of the striatum. First, dysfunctions of the dopaminergic system are linked to numerous diseases that are postulated to have a developmental origin, including schizophrenia (Lau et al., 2013; Laruelle, 2014), obsessive compulsive disorder (Nikolaus et al., 2010) and Tourette Syndrome (Buse et al., 2013). Second, embryonic exposure to drugs of abuse that act on the dopaminergic system, such as methamphetamine (Heller et al., 2001), induces profound brain and behavioral abnormalities in the offspring.

Here we tested the hypothesis that neuromodulators, and specifically dopamine, regulate glutamate-dependent synaptogenesis during striatal development. We found a potentially general mechanism of developmental synaptogenesis that relies on the coordinated actions of glutamate and the activation of PKA via Gα_s-coupled receptors.

Consistent with early development of the dopaminergic projection from SNc/VTA to striatum, the dopamine transporter promoter is active and dopaminergic axons are found in the neonatal striatum on postnatal day 1 (P1) (Figure 1A). By P10 the axons of midbrain dopaminergic neurons densely innervate the striatum and express tyrosine hydroxylase, a critical enzyme for dopamine synthesis (Figure 1B), indicating that dopaminergic neuromodulation is poised to regulate synapse and cell function during this early developmental stage.

Figure 1 with 3 supplements see all

Download asset Open asset

To address whether neuromodulation regulates synapse formation in developing striatum, we examined the number and strength of glutamatergic synapses on SPNs in acute brain slices prepared from P8-13 mice, age-balanced across groups. Comparisons of dendritic spine density and spontaneous miniature excitatory post-synaptic currents (mEPSCs) were made between slices prepared from mice injected 1 hr previously with saline or ligands for GPCRs differentially expressed by dSPNs and iSPNs.

We observed that in acute mouse brain slices prepared 1 hr after systemic injection of the Drd1 agonist SKF 81297 (5 mg/kg), both dendritic spine density and mEPSC frequency in dSPNs were increased compared to saline-injected littermates (Figure 1C–G). The frequency of mEPSCs in Drd1 agonist exposed SPNs was 593 ± 168% of that in saline-treated controls (N = 7 and 8 neurons from 4 mice/group, Kruskal-Wallis test p = 0.0063, with p<0.05 for Dunn’s multiple comparison post hoc test). mEPSC amplitude was 10.8 ± 1.6 and 12.7 ± 1.1 pA for control and Drd1 agonist groups and was not significantly different (one way ANOVA, p = 0.313). Dendritic spine density was 0.29 ± 0.02 spines/µm and 0.60 ± 0.04 spines/µm for saline control and Drd1 agonist groups respectively (p<0.001 on Dunnett’s multiple comparison test after one-way ANOVA; N = 36 dendrites, 12 neurons from 3 mice and 21 dendrites in 7 neurons from 3 mice). Pretreating mice with a PKA antagonist H-89 (5 mg/kg) 30 min prior to agonist injection prevented the effects on mEPSC frequency and spine density (mEPSC frequency, 122 ± 47% of control with Drd1 agonist, N = 6 and 8 neurons from 2 mice/group, p>0.05 for Dunn’s multiple comparison post hoc test; spine density, 0.27 ± 0.02 spines/µm vs. 0.29 ± 0.02 spines/µm in control and Drd1 agonist treatment conditions, N = 18 dendrites in 6 neurons from 2 mice, and 24 dendrites in 8 neurons from 2 mice). These results suggest a PKA-dependent increase in the number of AMPA receptor-containing glutamatergic synapses induced by in vivo Drd1 activation. However, Drd1 receptor activation alone is insufficient to increase synapse number, as changes in mEPSC frequency or amplitude were not observed with incubation of acute brain slices in SKF 81297 in vitro (Figure 1—figure supplement 1. For mEPSC frequency, 107 ± 23% of ACSF control values with Drd1 stimulation; N = 6 neurons from 2 mice/group, t-test p = 0.83. The amplitude of mEPSCs was 12.3 ± 1.9 and 13.8 ± 2.2 pA for control and Drd1 stimulation groups respectively, t-test p = 0.51). Thus, dSPN excitatory synapse enhancement by a single, acute dose of a Drd1 receptor agonist alone requires intact circuitry, as it cannot be induced in the slice.

In contrast to the effects seen in dSPNs, in vivo administration of SKF 81297 had no effects on excitatory synapses in iSPNs (Figure 1H), consistent with the lack of expression of Drd1 receptor on these cells (Gerfen et al., 1990) (mEPSC frequency, 73.3 ± 30.1% of control values in presence of Drd1 agonist; N = 6 neurons from 4 mice and 8 neurons from 2 mice). iSPNs express Gα_s-coupled Adenosine 2a (A2a) receptor, which also enhances PKA activity (Higley and Sabatini 2010) – nevertheless, administration of an agonist of A2aRs (CGS 21680, 0.1 mg/kg) had no significant effect on iSPN dendritic spine density or mEPSC rate, within the rapid time-frame of 1 hr (140 ± 63% of control, p>0.05 for control vs. A2aR agonist post-hoc comparison; N = 11 neurons from 3 mice). Similarly, administering an agonist of Gα_i-coupled Drd2 receptors (quinpirole 0.5 mg/kg) also resulted in no significant differences from controls (132 ± 17.0% of control, p>0.05 for control vs. Drd2 agonist post hoc comparison; N = 8 neurons from 3 mice). However, co-administration of Drd1 and A2a receptor agonists increased mEPSC frequency and dendritic spine density in iSPNs, with no change in mEPSC amplitude, suggesting an increase in the number of functional excitatory synapses (Figure 1H) (mEPSC frequency, 470 ± 124% of control values with CGS 21680/SKF 81297, one-way ANOVA p<0.0008, Bonferroni post hoc p<0.05; N = 6 neurons from 4 mice and 5 neurons from 2 mice for this comparison. Dendritic spine density, 0.29 ± 0.02 spines/µm and 0.63 ± 0.05 spines/µm, for control vs. CGS 21680/SKF 81297, one-way ANOVA p<0.0001, Bonferroni post hoc p<0.0001; 31 dendrites, 9 neurons, 2 mice, and 7 dendrites, 4 neurons in 2 mice, for this comparison).

These results indicate that neither the activation of dopamine receptors, nor of Gα_s-coupled GPCRs, is sufficient to cell-autonomously enhance glutamatergic synapses on SPNs. Instead, neuromodulatory GPCR signaling may be a permissive, or facilitating, signal that acts in concert with other stimuli to induce synaptogenesis in vivo. We hypothesized that GPCR activation modulates glutamate-induced, activity-dependent synaptogenesis. Since Drd1 activation also increases dSPN excitability, which in turn may increase recurrent corticostriatal glutamatergic transmission onto SPNs by altering cortical activity (Oldenburg and Sabatini, 2015), then activation of this receptor alone in vivo will (1) activate PKA and (2) enhance glutamate release. This may be sufficient to drive synapse formation. Such stimulatory effects of dopamine on basal ganglia circuitry are consistent with the observation of Drd1 receptor activation-induced locomotion observed in pups, which normally show limited mobility at P8-13 (Dehorter et al., 2011; Wills et al., 2014) (Figure 1—figure supplement 2). This acute locomotion induction has lasting consequences for behavior and brain development, since priming with Drd1 receptor agonist administration at P9 potentiated agonist-evoked locomotion at P11 (Figure 1—figure supplement 3). In contrast to dSPNs, iSPNs would be expected to have more complex requirements for synaptogenesis, since iSPN activity is negatively coupled to intrastriatal glutamate release in the adult, via regulation of cortical activity (Oldenburg and Sabatini, 2015). We hypothesize that the cooperation of presynaptic glutamate release and postsynaptic PKA signaling explains why coincident activation of Drd1 receptors (to recurrently drive glutamate release) and A2aRs (to enhance PKA signaling) is required in vivo to increase synaptogenesis in iSPNs.

We directly tested this hypothesis ex vivo by examining the ability of exogenous glutamate exposure, through focal photolysis of MNI-L-glutamate, to trigger synaptogenesis in SPNs under different states of GPCR activation. Spatio-temporally controlled glutamate release can focally and rapidly induce de novo formation of a synapse-bearing dendritic spine on young cortical pyramidal neurons (Kwon and Sabatini, 2011). Cre-dependent AAV was used to induce GFP expression in Cre-expressing dSPNs or iSPNs (Figure 2A–C), allowing visualization of the cellular and dendritic morphology of sparsely labeled neurons of either pathway. Repeated glutamate uncaging (40 pulses at 2 Hz) with a variety of pulse durations (0.5–4 ms) was applied near a GFP-positive dendrite under conditions that maximally activate NMDA-type glutamate receptors (0 mM added extracellular Mg²⁺). In a subset of cases, this protocol induced de novo growth of a dendritic spine-like protrusion (Figure 2C). As a control for non-specific, photodamage-induced changes in tissue fluorescence, in a subset of trials, an interleaved stimulus of identical strength was delivered to a control site >5 µm away from any labeled dendrite (Figure 2B, Figure 2—figure supplement 1) (0.5 vs. 0 success probability at proximal and distal uncaging locations; 82 trials from 2 mice).

Figure 2 with 1 supplement see all

Download asset Open asset

As was previously described for layer 2/3 pyramidal neurons, P8-13 SPNs of both pathways were capable of de novo spinogenesis in response to glutamate uncaging (Kozorovitskiy et al., 2012; Kwon and Sabatini, 2011). The basal rate of spinogenesis (the probability of induction of a structure shaped like a dendritic spine), as well as its dependence on uncaging pulse width and proximity to the dendrite were similar for striatal SPNs and previously examined cortical neurons (Kwon and Sabatini, 2011). For both dSPNs and iSPNs, the success rate for inductions with stimulation 1–2 µm away from a labeled dendrite was ~20–50%, with longer uncaging pulses associated with a greater probability of spinogenesis (dSPNs, N = 146 and 174 trials in 10 mice for control group and in presence of GPCR agonist, respectively, balanced across conditions; iSPNs, N = 132 and 164 trials in 11 mice). The probability of spinogenesis increased in the presence of SKF 81297 for dSPNs, and CGS 21680 for iSPNs, respectively (the proportion of success trials was 0.23, 0.40, and 0.57 in the control condition, for varied length glutamate uncaging pulses, and 0.44, 0.62, and 0.76 in presence of Drd1 receptor agonist for dSPNs; for iSPNs, the respective proportions were 0.19, 0.46, and 0.49 in the control condition vs. 0.73, 0.74, and 0.70 for iSPNs with A2aR stimulation; two-tailed Z tests for proportions, p = 0.040, 0.015, and 0.018 for dSPNs, and p<0.0001, p = 0.018, and 0.008 for iSPNs). This enhancement was observed for all 3 uncaging pulse-widths tested (Figure 2D), and is also evident in across-cell averages within each mouse, for the 4 ms-long uncaging pulse condition (Figure 2E; proportion of successful trials, 0.48 vs. 0.80 for dSPNs in control and with Drd1 agonist treatment, and 0.48 vs. 0.72 for iSPNs; dSPNs, N = 3 and 4 mice/group, 68 and 75 trials; iSPNs, N = 3 and 4 mice/group, 61 and 103 trials).

Analysis of the time course of spine formation for successful inductions reveals that Gα_s-coupled receptor stimulation decreased the number of glutamate pulses necessary to induce spinogenesis (Figure 2F). Whereas spine growth occurs rapidly in all conditions, as observed in pyramidal neurons (Kwon and Sabatini, 2011), SKF 81297 and CGS 21680 accelerated the onset of spinogenesis for dSPNs and iSPNs, respectively. We compared the number of photoactivation pulses to threshold, defined as 50% fluorescence increase over baseline in the uncaging location, leading to an appearance of a spine-like extension from the dendrite. At 2 Hz stimulation, for successful trials, threshold pulse number was 21.3 without drug and 6.0 in the presence of Gα_s agonist for dSPNs, and 17.4 vs. 8.7 pulses for iSPNs (Mann-Whitney test, p = 0.0025 and p = 0.0048 for d- and iSPN comparisons, respectively; N = 12 and 13 induction trials for dSPN control vs. in the presence of agonist, with 16 and 17 induction trials for iSPNs, respectively). Given that glutamate was exogenously controlled and Gα_s receptors were stimulated by agonist bath application, these findings suggest the presence of cell-autonomous mechanisms for neuromodulatory regulation of glutamate-dependent spinogenesis.

If a similar interaction of neuromodulation and glutamate release controls synapse formation in vivo, it should be possible to bias rapidly developing networks towards increased synaptogenesis by briefly enhancing either PKA signaling or glutamate release. To test this prediction for PKA signaling, we used a pharmacogenetic approach based on Rs1, a modified Gα_s-coupled GPCR sensitive to a small, exogenous, blood-brain-barrier permeable ligand (Srinivasan et al., 2003). The efficacy of this GPCR coupling to Gα_s was examined in engineered HEK293 cells, in which GPCR coupling to Gα_s elevates cyclic AMP and activates a transcriptional cAMP response element, leading to the production of secreted alkaline phosphatase (SEAP) (Liberles and Buck, 2006). Transfection of HEK293 cells with Rs1, followed by incubation in the agonist RS 67333 and chemi-luminescent analysis of SEAP levels revealed a sub-micromolar EC₅₀ of Rs1 activation (Figure 3A).

Figure 3 with 1 supplement see all

Download asset Open asset

We tested the sufficiency of Gα_s-coupled GPCR activation in SPNs for enhancing corticostriatal transmission, using a combination of pharmacogenetic activation in vivo and optogenetic analysis of evoked corticostriatal transmission ex vivo (Figure 3B–G). We compared the magnitude of EPSCs evoked by channelrhodopsin (ChR2)-mediated activation of Rbp4-Cre expressing neurons, which include a dense corticostriatal projection, in SPNs in acute brain slices from animals with and without Rs1 expression. Rbp4-Cre; Ai32 mice were unilaterally injected in the striatum with a Cre-expressing AAV and Cre-dependent AAV encoding Rs1 (Hsiao et al., 2008) (Figure 3C–D). This co-injection resulted in broad expression in dorsal striatum SPNs of the injected hemisphere in both dSPNs and iSPNs (Figure 3E) (0% vs. 98% co-labeling, N = 750 and 728 neurons in control and injected hemispheres, from 3 mice). P8-13 infected mice were injected with the Rs1 ligand RS 67333 (3 mg/kg) and acute brain slices were prepared 1.5 hr later for analysis. Whole-cell voltage clamp recordings were used to monitor pharmacologically isolated, AMPAR-mediated blue light-evoked corticostriatal EPSCs, which were recorded from SPNs in dorsal striatum, located either in the control or in the infected hemisphere. Although this strategy results in ChR2 expression in SPNs, the brief time of expression from a single floxed ChR2 allele yielded negligible direct ChR2-evoked currents in SPNs under whole-cell voltage clamp configuration (e.g., Figure 3F). With this strategy, both the control and infected striatum were exposed to RS 67333 and comparisons were made across SPNs from control and infected hemispheres in the same animal. In the control condition, there was a strong age dependence of peak amplitude of EPSCs, ranging from an average of 10 pA to an average of 196 pA per age group, in a cohort of P8-P13/14 mice (Figure 3—figure supplement 1; light evoked EPSC amplitude across age, N = 14 neurons from 3 mice, 30 neurons from 6 mice and 14 neurons from 3 mice/group. Kruskal-Wallis test p<0.0001, for main effect and Dunn’s post hoc tests). Consistent with our hypothesis, light-evoked EPSC amplitude was increased in virally transduced SPNs, compared to control hemisphere SPNs (Figure 3G; light-evoked EPSC amplitude, 68.6 ± 29.9 pA control vs. 150.3 ± 55.8 pA injected hemisphere; N = 21 and 23 neurons from 4 mice/group, t-test p = 0.039). Thus, in vivo transient activation of a Gα_s-coupled GPCR in SPNs is sufficient to rapidly enhance corticostriatal connectivity, in the presence of otherwise normal corticostriatal transmission. The observed sufficiency of pharmacogenetic PKA activation to enhance corticostriatal connectivity for iSPNs (in contrast to stimulating iSPNs with an A2aR agonist, as in Figure 1H) stems from the likelihood that the concurrent pharmacogenetic activation of Gα_s cascades in dSPNs enhances corticostriatal glutamatergic drive via recurrent connections and regulating cortical activity (Oldenburg and Sabatini, 2015).

Conversely, to examine whether increased glutamate release in vivo is able to enhance corticostriatal transmission during this developmental period, a small blue light LED was mounted over somatosensory cortex on one hemisphere and used to activate ChR2-expressing Rbp4-Cre positive neurons in cortex (Figure 4). Mice were awake and freely moving within an enclosure, when a mild stimulus was delivered, consisting of 300 pulses of light over the course of 1 hr, given in bursts of 5 pulses at 20 Hz every minute (Figure 4A). This treatment induced moderate c-fos expression in the stimulated cortical region (Figure 4B). Acute slices were prepared immediately after the stimulation protocol and AMPAR-mediated light-evoked EPSC were recorded from dorsal striatum SPNs, located in the control or stimulated hemisphere. Both EPSC peak amplitude and charge transfer were increased on the stimulated side (Figure 4D, Figure 4—figure supplement 1) (Light-evoked EPSC amplitude, 30.5 ± 4.8 pA vs. 50.4 ± 6.6 pA for control and stimulated hemisphere comparison; charge transfer, 0.32 ± 0.14 pC and 0.89 ± 0.21 pC; N = 29–33 neurons/group, 4 mice/group). Paired analysis of EPSC amplitude within subjects across the two hemispheres is consistent with individual cell data (Figure 4D), confirming the successful induction of corticostriatal rewiring following cortical optogenetic stimulation (EPSC amplitude, control 16.2 ± 3.2 vs. 27.3 ± 3.2 pA in the stimulated hemisphere, paired t-test p = 0.026, 4 mice/group).

Figure 4 with 1 supplement see all

Download asset Open asset

Because stimulating layer 5 pyramidal neurons may alter intra-cortical connectivity, which could indirectly account for the observed differences in SPN responses, we tested whether this form of plasticity is striatally expressed by trimming off cortex from acute coronal brain slices with a fine scalpel prior to recording (Figure 4E, Figure 4—figure supplement 1). Both peak EPSC amplitude and charge transfer were increased on the stimulated side (light-evoked EPSC amplitude, 9.6 ± 2.4 pA vs. 23.4 ± 4.3 pA for control vs. stimulated hemisphere; charge transfer, 0.64 ± 0.14 and 1.2 ± 0.26 pC for the same comparison; N = 19 and 26 neurons from 4 mice/group). These data demonstrate that a major component of the functional reorganization induced through brief in vivo stimulation of corticostriatal afferents in young mice is locally expressed in the striatum. We further probed the pharmacological dependence of this form of plasticity and found that it was blocked by either pre-administration of PKA inhibitor H-89 (5 mg/kg), or by a combination of Drd1 and A2aR blockers SKF 83566 (0.5 mg/kg) and istradefylline/KW-6002 (2.5 mg/kg), administered 30 min before optogenetic stimulation (Figure 4F–G). Light-evoked EPSC amplitude, in presence of H-89, was 18.8 ± 4.4 pA vs. 13.8 ± 3.3 pA in neurons from control and stimulated hemispheres (t-test p = 0.376, N = 22 and 19 neurons from 3 mice/group). In presence of Drd1 and A2aR blockers, it was 34.2 ± 9.3 pA and 26.5 ± 5.3 pA, respectively (t-test p = 0.458, N = 12 and 15 neurons from 3 mice/group). Thus, SPN changes induced by glutamatergic stimulation require PKA signaling, which in turn is activated via Drd1 and A2aR for dSPNs and iSPNs, respectively. Increasing either pre-synaptic glutamatergic drive or enhancing striatal PKA signaling is sufficient to rapidly alter SPN synaptogenesis, supporting the idea of cooperative, neuromodulation-dependent circuit remodeling during striatal development.

In order to directly evaluate whether PKA activation and glutamate release within striatum are sufficient for strengthening corticostriatal synapses, we performed ex vivo plasticity induction using optogenetic stimulation of corticostriatal afferents in the presence of drugs activating PKA in SPNs (Figure 5). We found that the same optogenetic stimulation parameters that evoked PKA-dependent enhancements in evoked responses in vivo were sufficient to produce a moderate enhancement in light-evoked corticostriatal responses in vitro. In experiments using within-mouse controls, the amplitude of evoked responses was 528.8 ± 69 pA and 694.4 ± 79 pA, for neurons in control and stimulated slices respectively (Figure 5B–C) (Mann-Whitney test, p = 0.037, N = 64 and 61 neurons from 4 mice). In contrast, incubation in PKA agonists alone was insufficient to alter evoked response amplitude (Mann-Whitney test, p = 0.6, N = 19 and 14 neurons). These data directly confirm the existence of a mechanism for cooperative, neuromodulation-dependent circuit remodeling during striatal development.

Figure 5

Download asset Open asset

We examined the hypothesis that neuromodulation interacts with glutamate-dependent circuit wiring in the developing striatum. Instead of a privileged action of dopamine in striatal synapse development, we discovered a mechanism of regulated synaptogenesis requiring cooperative action of glutamatergic drive and neuromodulation. Our data reveal that neuromodulation, through the activity of PKA, enhances the probability of dendritic spine formation and lowers the amount of glutamate necessary to trigger spinogenesis in the developing striatum.

These findings must be integrated into the context of basal ganglia circuitry, where inhibitory outputs of the direct and indirect pathways form functionally complementary loops that provide polysynaptic recurrent positive and negative feedback for excitatory striatal synaptogenesis (Kozorovitskiy et al., 2012). Because dopamine tends to increase dSPN PKA activity and decrease iSPN PKA activity (models, Figure 6A–B), iSPN spinogenesis may be effectively yoked to dSPN development, as described below. Dopamine is abundant in the developing striatum, and it enhances PKA activity in dSPNs. This allows dSPNs to strongly respond to glutamate release from the first invading axons, becoming integrated into early basal ganglia circuits. On the other hand, iSPN activity and PKA levels are decreased by dopamine (Hernandez-Lopez et al., 2000) and their excitatory wiring may rely either on the positive feedback through direct pathway (Kozorovitskiy et al., 2012) to increase glutamate release in the striatum, or on striatal adenosine levels that may be driven by local activity. This scenario predicts a transient developmental bias in pathway innervation towards dSPNs, which we have previously observed in the frequency of mEPSCs, as well as in dendritic spine density at P14-15 (Figure 6C) (mEPSC frequency, 0.61 ± 0.1 and 0.35 ± 0.1 Hz for d- and iSPNs, t-test p = 0.019; mEPSC amplitude, 14.3 ± 0.5 and 14.0 ± 0.5 pA, p = 0.679; dendritic spine density, 0.74 ± 0.025 and 0.52 ± 0.013 spines/µm, p<0.0001; N = 55 neurons from 12 mice and 56 neurons from 11 mice for electrophysiology; a distributed subset of recorded neurons was imaged for dendritic spine density analyses, 67 dSPN dendrites from 22 neurons and 93 iSPN dendrites from 31 neurons). Thus, the palette of neuromodulators and their postsynaptic coupling in a given brain region or developmental time window orchestrate the integration of each neuronal population into functional circuitry.

Figure 6

Download asset Open asset

The observation that two types of Gα_s-dependent modulation appear to regulate synaptogenesis in d- and iSPNs raises the possibility that PKA activity, via Gα_s GPCRs, generally serves to potentiate activity-dependent synaptogenesis during the time of rapid synapse production. The array of Gα_s GPCRs expressed in the central nervous system is substantial and includes abundant receptors such as 5-HT₄ and 5-HT₇ serotonin receptors (Giulietti et al., 2014), as well beta-adrenergic (Nomura et al., 2014; Arriza et al., 1992) and corticotropin-releasing factor receptors (Fu and Neugebauer, 2008; Riegel and Williams, 2008). A general role for neuromodulators acting on Gα_s GPCRs during neural circuit development would transform our understanding of core principles in developmental neurobiology, greatly expanding the list therapeutic targets for neurodevelopmental and neurodegenerative disorders.

1. 1. Arriza JL
   2. Dawson TM
   3. Simerly RB
   4. Martin LJ
   5. Caron MG
   6. Snyder SH
   7. Lefkowitz RJ

   (1992)

   The g-protein-coupled receptor kinases beta ARK1 and beta ARK2 are widely distributed at synapses in rat brain

   The Journal of Neuroscience 12:4045–4055.

   • Google Scholar
2. 1. Bateup HS
   2. Santini E
   3. Shen W
   4. Birnbaum S
   5. Valjent E
   6. Surmeier DJ
   7. Fisone G
   8. Nestler EJ
   9. Greengard P

   (2010) Distinct subclasses of medium spiny neurons differentially regulate striatal motor behaviors

   Proceedings of the National Academy of Sciences of the United States of America 107:14845–14850.

   https://doi.org/10.1073/pnas.1009874107

   • Google Scholar
3. 1. Bromberg-Martin ES
   2. Matsumoto M
   3. Hikosaka O

   (2010) Dopamine in motivational control: rewarding, aversive, and alerting

   Neuron 68:815–834.

   https://doi.org/10.1016/j.neuron.2010.11.022

   • Google Scholar
4. 1. Buse J
   2. Schoenefeld K
   3. Münchau A
   4. Roessner V

   (2013) Neuromodulation in tourette syndrome: dopamine and beyond

   Neuroscience & Biobehavioral Reviews 37:1069–1084.

   https://doi.org/10.1016/j.neubiorev.2012.10.004

   • Google Scholar
5. 1. Bäckman CM
   2. Malik N
   3. Zhang Y
   4. Shan L
   5. Grinberg A
   6. Hoffer BJ
   7. Westphal H
   8. Tomac AC

   (2006) Characterization of a mouse strain expressing cre recombinase from the 3′ untranslated region of the dopamine transporter locus

   Genesis 44:383–390.

   https://doi.org/10.1002/dvg.20228

   • Google Scholar
6. 1. Cui G
   2. Jun SB
   3. Jin X
   4. Pham MD
   5. Vogel SS
   6. Lovinger DM
   7. Costa RM

   (2013) Concurrent activation of striatal direct and indirect pathways during action initiation

   Nature 494:238–242.

   https://doi.org/10.1038/nature11846

   • Google Scholar
7. 1. Dehorter N

   (2011) Onset of pup locomotion coincides with loss of NR2C/D-mediated cortico-striatal EPSCs and dampening of striatal network immature activity

   Frontiers in Cellular Neuroscience 5:24.

   https://doi.org/10.3389/fncel.2011.00024

   • Google Scholar
8. 1. Ferrari DC
   2. Mdzomba BJ
   3. Dehorter N
   4. Lopez C
   5. Michel FJ
   6. Libersat F
   7. Hammond C

   (2012) Midbrain dopaminergic neurons generate calcium and sodium currents and release dopamine in the striatum of pups

   Frontiers in Cellular Neuroscience 6:.

   https://doi.org/10.3389/fncel.2012.00007

   • Google Scholar
9. 1. Fu Y
   2. Neugebauer V

   (2008) Differential mechanisms of CRF1 and CRF2 receptor functions in the amygdala in pain-related synaptic facilitation and behavior

   Journal of Neuroscience 28:3861–3876.

   https://doi.org/10.1523/JNEUROSCI.0227-08.2008

   • Google Scholar
10. 1. Gerfen C
    2. Engber T
    3. Mahan L
    4. Susel Z
    5. Chase T
    6. Monsma F
    7. Sibley D

    (1990) D1 and D2 dopamine receptor-regulated gene expression of striatonigral and striatopallidal neurons

    Science 250:1429–1432.

    https://doi.org/10.1126/science.2147780

    • Google Scholar
11. 1. Giulietti M
    2. Vivenzio V
    3. Piva F
    4. Principato G
    5. Bellantuono C
    6. Nardi B

    (2014) How much do we know about the coupling of g-proteins to serotonin receptors?

    Molecular Brain 7:49.

    https://doi.org/10.1186/s13041-014-0049-y

    • Google Scholar
12. 1. Gong S
    2. Zheng C
    3. Doughty ML
    4. Losos K
    5. Didkovsky N
    6. Schambra UB
    7. Nowak NJ
    8. Joyner A
    9. Leblanc G
    10. Hatten ME
    11. Heintz N

    (2003) A gene expression atlas of the central nervous system based on bacterial artificial chromosomes

    Nature 425:917–925.

    https://doi.org/10.1038/nature02033

    • Google Scholar
13. 1. Gong S
    2. Doughty M
    3. Harbaugh CR
    4. Cummins A
    5. Hatten ME
    6. Heintz N
    7. Gerfen CR

    (2007) Targeting cre recombinase to specific neuron populations with bacterial artificial chromosome constructs

    Journal of Neuroscience 27:9817–9823.

    https://doi.org/10.1523/JNEUROSCI.2707-07.2007

    • Google Scholar
14. 1. Heller A
    2. Bubula N
    3. Freeney A
    4. Won L

    (2001) Elevation of fetal dopamine following exposure to methamphetamine in utero

    Developmental Brain Research 130:139–142.

    https://doi.org/10.1016/S0165-3806(01)00222-X

    • Google Scholar
15. 1. Hernandez-Lopez S
    2. Tkatch T
    3. Perez-Garci E
    4. Galarraga E
    5. Bargas J
    6. Hamm H
    7. Surmeier DJ

    (2000)

    D2 dopamine receptors in striatal medium spiny neurons reduce l-type Ca2+ currents and excitability via a novel PLC[beta]1-IP3-calcineurin-signaling cascade

    The Journal of Neuroscience 20:8987–8995.

    • Google Scholar
16. 1. Higley MJ
    2. Sabatini BL

    (2010) Competitive regulation of synaptic Ca2+ influx by D2 dopamine and A2A adenosine receptors

    Nature Neuroscience 13:958–966.

    https://doi.org/10.1038/nn.2592

    • Google Scholar
17. 1. Hsiao EC
    2. Boudignon BM
    3. Chang WC
    4. Bencsik M
    5. Peng J
    6. Nguyen TD
    7. Manalac C
    8. Halloran BP
    9. Conklin BR
    10. Nissenson RA

    (2008) Osteoblast expression of an engineered gs-coupled receptor dramatically increases bone mass

    Proceedings of the National Academy of Sciences of the United States of America 105:1209–1214.

    https://doi.org/10.1073/pnas.0707457105

    • Google Scholar
18. 1. Hu Z
    2. Cooper M
    3. Crockett DP
    4. Zhou R

    (2004) Differentiation of the midbrain dopaminergic pathways during mouse development

    The Journal of Comparative Neurology 476:301–311.

    https://doi.org/10.1002/cne.20230

    • Google Scholar
19. 1. Kozorovitskiy Y
    2. Saunders A
    3. Johnson CA
    4. Lowell BB
    5. Sabatini BL

    (2012) Recurrent network activity drives striatal synaptogenesis

    Nature7400 485:646–650.

    https://doi.org/10.1038/nature11052

    • Google Scholar
20. 1. Kravitz AV
    2. Freeze BS
    3. Parker PRL
    4. Kay K
    5. Thwin MT
    6. Deisseroth K
    7. Kreitzer AC

    (2010) Regulation of parkinsonian motor behaviours by optogenetic control of basal ganglia circuitry

    Nature 466:622–626.

    https://doi.org/10.1038/nature09159

    • Google Scholar
21. 1. Kwon H-B
    2. Sabatini BL

    (2011) Glutamate induces de novo growth of functional spines in developing cortex

    Nature 474:100–104.

    https://doi.org/10.1038/nature09986

    • Google Scholar
22. 1. Laruelle M

    (2014) Schizophrenia: from dopaminergic to glutamatergic interventions

    Current Opinion in Pharmacology 14:97–102.

    https://doi.org/10.1016/j.coph.2014.01.001

    • Google Scholar
23. 1. Lau C-I
    2. Wang H-C
    3. Hsu J-L
    4. Liu M-E

    (2013) Does the dopamine hypothesis explain schizophrenia?

    Reviews in the Neurosciences 24:389–400.

    https://doi.org/10.1515/revneuro-2013-0011

    • Google Scholar
24. 1. Liberles SD
    2. Buck LB

    (2006) A second class of chemosensory receptors in the olfactory epithelium

    Nature 442:645–650.

    https://doi.org/10.1038/nature05066

    • Google Scholar
25. 1. Madisen L
    2. Zwingman TA
    3. Sunkin SM
    4. Oh SW
    5. Zariwala HA
    6. Gu H
    7. Ng LL
    8. Palmiter RD
    9. Hawrylycz MJ
    10. Jones AR
    11. Lein ES
    12. Zeng H

    (2010) A robust and high-throughput cre reporting and characterization system for the whole mouse brain

    Nature Neuroscience 13:133–140.

    https://doi.org/10.1038/nn.2467

    • Google Scholar
26. 1. Nikolaus S
    2. Antke C
    3. Beu M
    4. Müller H-W

    (2010) Cortical GABA, striatal dopamine and midbrain serotonin as the key players in compulsive and anxiety disorders - results from in vivo imaging studies

    Reviews in the Neurosciences 21:119–139.

    https://doi.org/10.1515/REVNEURO.2010.21.2.119

    • Google Scholar
27. 1. Nishi A
    2. Snyder GL
    3. Greengard P

    (1997)

    Bidirectional regulation of DARPP-32 phosphorylation by dopamine

    The Journal of Neuroscience 17:8147–8155.

    • Google Scholar
28. 1. Nishi A
    2. Snyder GL
    3. Fienberg AA
    4. Fisone G
    5. Aperia A
    6. Nairn AC
    7. Greengard P

    (1999) Requirement for DARPP-32 in mediating effect of dopamine D2 receptor activation

    European Journal of Neuroscience 11:2589–2592.

    https://doi.org/10.1046/j.1460-9568.1999.00724.x

    • Google Scholar
29. 1. Nishi A
    2. Bibb JA
    3. Snyder GL
    4. Higashi H
    5. Nairn AC
    6. Greengard P

    (2000) Amplification of dopaminergic signaling by a positive feedback loop

    Proceedings of the National Academy of Sciences of the United States of America 97:12840–12845.

    https://doi.org/10.1073/pnas.220410397

    • Google Scholar
30. 1. Nomura S
    2. Bouhadana M
    3. Morel C
    4. Faure P
    5. Cauli B
    6. Lambolez B
    7. Hepp Régine

    (2014) Noradrenalin and dopamine receptors both control cAMP-PKA signaling throughout the cerebral cortex

    Frontiers in Cellular Neuroscience 8:247.

    https://doi.org/10.3389/fncel.2014.00247

    • Google Scholar
31. 1. Oldenburg IA
    2. Sabatini BL

    (2015) Antagonistic but not symmetric regulation of primary motor cortex by basal ganglia direct and indirect pathways

    Neuron 86:1174–1181.

    https://doi.org/10.1016/j.neuron.2015.05.008

    • Google Scholar
32. 1. Pologruto TA
    2. Sabatini BL
    3. Svoboda K

    (2003) ScanImage: flexible software for operating laser scanning microscopes

    BioMedical Engineering OnLine 2:13.

    https://doi.org/10.1186/1475-925X-2-13

    • Google Scholar
33. 1. Riegel AC
    2. Williams JT

    (2008) CRF facilitates calcium release from intracellular stores in midbrain dopamine neurons

    Neuron 57:559–570.

    https://doi.org/10.1016/j.neuron.2007.12.029

    • Google Scholar
34. 1. Srinivasan S
    2. Vaisse C
    3. Conklin BR

    (2003) Engineering the melanocortin-4 receptor to control g sub>_ssub> _signaling in vivo

    Annals of the New York Academy of Sciences 994:225–232.

    https://doi.org/10.1111/j.1749-6632.2003.tb03184.x

    • Google Scholar
35. 1. Schindelin J
    2. Arganda-Carreras I
    3. Frise E
    4. Kaynig V
    5. Longair M
    6. Pietzsch T
    7. Preibisch S
    8. Rueden C
    9. Saalfeld S
    10. Schmid B
    11. Tinevez J-Y
    12. White DJ
    13. Hartenstein V
    14. Eliceiri K
    15. Tomancak P
    16. Cardona A

    (2012) Fiji: an open-source platform for biological-image analysis

    Nature Methods 9:676–682.

    https://doi.org/10.1038/nmeth.2019

    • Google Scholar
36. 1. Shen W
    2. Flajolet M
    3. Greengard P
    4. Surmeier DJ

    (2008) Dichotomous dopaminergic control of striatal synaptic plasticity

    Science 321:848–851.

    https://doi.org/10.1126/science.1160575

    • Google Scholar
37. 1. Wills TJ
    2. Muessig L
    3. Cacucci F

    (2014) The development of spatial behaviour and the hippocampal neural representation of space

    Philosophical Transactions of the Royal Society B: Biological Sciences 369:20130409.

    https://doi.org/10.1098/rstb.2013.0409

    • Google Scholar

Author details

1. Yevgenia Kozorovitskiy

   1. Department of Neurobiology, Howard Hughes Medical Institute, Harvard Medical School, Boston, United States
   2. Department of Neurobiology, Northwestern University, Evanston, United States

   Contribution

   YK, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

2. Rui Peixoto

   Department of Neurobiology, Howard Hughes Medical Institute, Harvard Medical School, Boston, United States

   Contribution

   RP, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

3. Wengang Wang

   Department of Neurobiology, Howard Hughes Medical Institute, Harvard Medical School, Boston, United States

   Contribution

   WW, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

4. Arpiar Saunders

   Department of Neurobiology, Howard Hughes Medical Institute, Harvard Medical School, Boston, United States

   Present address

   Department of Genetics, Harvard Medical School, Boston, United States

   Contribution

   AS, Carried out a subset of mEPSC recording and dendritic spine imaging experiments in older mice, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

5. Bernardo L Sabatini

   Department of Neurobiology, Howard Hughes Medical Institute, Harvard Medical School, Boston, United States

   Contribution

   BLS, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   bernardo_sabatini@hms.harvard.edu

   Competing interests

   The authors declare that no competing interests exist.

• 3,932

  views

• 894

  downloads

• 61

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.