Dopamine’s secret agent: serotonin

Abstract

Dopamine suppresses GABA release from striatal terminals in the substantia nigra pars reticulata. Molinari and colleagues recently demonstrated that this suppression is frequency-dependent – instituting a high-pass filter on striatal “direct pathway” transmission – and does not require dopamine receptors. Rather, dopamine upregulates serotonin, activating presynaptic 5HT_1B receptors to exert its effects.

Dopamine is the belle of the ball when it comes to the regulation of basal ganglia function, but it doesn’t always act alone. Recent work by Molinari and colleagues [1] highlights how dopamine release in the substantia nigra pars reticulata (SNr) elevates serotonin tone via a non-canonical signaling mechanism, employing a “secret agent” to do its work. Their study emphasizes the importance of understanding complex mechanisms of neurotransmitter crosstalk.

In neuroscience research, it is often assumed that released neurotransmitters act by binding to their highaffinity receptors. This is the classic mode of chemical neurotransmission, as taught in ‘Neuroscience 101’. Molinari et al. therefore began their study by testing the reasonable hypothesis that dopamine regulates GABA release from the presynaptic terminals of striatal projection neurons expressing the D1 dopamine receptor (D1-SPNs) by activating presynaptic D1 receptors (Fig. 1A). Surprisingly, although the authors confirmed that dopamine inhibits GABA release from D1-SPN terminals, the D1 receptors expressed on these terminals did not appear to play a role: blocking them had no effect. The inhibition arose by a different mechanism.

Figure 1. Dopamine suppresses D1-SPN terminals in the SNr via presynaptic 5HT_1B receptors.

(A) Schematic representation of the “direct pathway” of the basal ganglia. Dopamine D1 receptorexpressing striatal projection neurons (D1-SPNs) send GABAergic projections to the substantia nigra pars reticulata (SNr), the output nucleus of the basal ganglia. Molinari et al. [1] examined whether dopamine (DA) release in the SNr from nearby substantia nigra pars compacta (SNc) neurons would suppress GABA release from D1-SPN terminals via presynaptic D1 receptors, as might be expected; however, this was not the case.

(B) Schematic representation of the model emerging from the new findings by Molinari et al. [1]. Somatodendritic DA release from SNc DA neurons boosts serotonin (5-HT) tone in the SNr, at least in part by acting as a competitive substrate for the serotonin transporter, SERT. 5-HT acts on presynaptic 5HT_1B receptors present on D1-SPN terminals, suppressing release. This suppression is strongest at low frequencies and thus imposes a high-pass filter on direct pathway transmission.

The D1-SPN synapses studied by Molinari et al. are important for regulating basal ganglia output. They form the so-called “direct pathway” of the basal ganglia, in which D1-SPNs in the striatum, the input nucleus of the basal ganglia, project directly to the SNr, the output nucleus of the basal ganglia [2]. According to the firing rate model of basal ganglia function, D1-SPNs should inhibit the tonic firing of SNr GABAergic neurons, thereby disinhibiting the postsynaptic targets of the SNr, including thalamus and brainstem motor centers, facilitating movement [2]. Within this framework, dopaminergic inhibition of direct pathway transmission at the level of the SNr could serve as a mechanism for filtering which direct pathway signals ultimately influence motor output. Given other studies suggesting that the loss of dopamine in the SNr is a crucial contributor to Parkinsonian motor deficits (e.g., [3]), this mechanism of dopamine action in the SNr may be highly relevant to motor disorders.

The findings by Molinari et al. add further key details to our understanding of the likely computational function of dopaminergic inhibition of direct pathway transmission. Molinari et al. demonstrate that SNr dopamine does not uniformly suppress direct pathway transmission but rather activates a frequency-dependent filter: dopamine strongly inhibited D1-SPN GABA release when these terminals were stimulated at low frequency (2Hz) but did not inhibit release at higher frequencies (10Hz). This high-pass filter on direct pathway transmission, present only when SNr dopamine is high, should selectively limit the epochs when direct pathway inhibition of SNr cells is effective.

The effects of dopamine on D1-SPN GABA release are consistent with previous reports [4,5] and supported by a variety of evidence. First, Molinari et al. observed effects on presynaptic release using FM1–43, a fluorescent dye that can be loaded into synaptic vesicles and reports exocytosis as a drop in fluorescence. Second, the authors observed presynaptic calcium dynamics in D1-SPN terminals using 2-photon imaging of GCaMP7s, a genetically encoded fluorescent calcium indicator. Third, they used electrophysiology to record inhibitory postsynaptic currents (IPSCs) from SNr GABA cells while optogenetically stimulating D1-SPN terminals. Across all these experiments, the authors found that dopamine application reduced D1-SPN GABA release. However, none of these effects were disrupted by dopamine antagonists or by genetic D1 receptor knockout.

If this dopamine-dependent high-pass filter does not depend on dopamine receptors, how is it effectuated? The answer seems to lie with serotonin. Serotonergic terminals from the dorsal raphe strongly innervate the SNr [6]. Molinari et al. found that dopamine increases serotonin tone in the SNr, likely by inhibiting the serotonin reuptake transporter, SERT. Dopamine can act as a competitive substrate for SERT, such that when serotonin and dopamine are present concurrently, less serotonin is removed from the synapse. Serotonin, in turn, activates presynaptic 5HT_1B receptors on D1-SPN terminals and lowers release probability (Fig. 1B) [1,4].

Returning to the logic of the firing-rate model described above, high-pass filtering of direct pathway signals when SNr dopamine is elevated may represent a mechanism by which SNr dopamine occludes marginal or ambiguous pro-movement signals from the striatum. Using optogenetic stimulation, Molinari et al. showed that a burst of dopamine neuron firing (20Hz, 1s stimulation) is sufficient to inhibit D1-SPN presynaptic terminals via 5HT_1B receptors. The implication is that dopamine neuron burst firing elicits somatodendritic release in the SNr [7] to activate the high-pass filter. If such bursting also incurs axonal dopamine release in the striatum, as expected, then dopamine would simultaneously facilitate the excitation of D1-SPN cell bodies [8], while providing a high-pass filter at their outputs. Therefore, high-frequency direct pathway activity would be both promoted and selected for amplification of signal-to-noise.

The study by Molinari et al. reveals a surprising mechanism that may have broad implications for understanding basal ganglia circuit function. Additional important questions about the regulation of SNr output by dopamine and serotonin remain to be addressed in future work. Serotonin exhibits other complex actions in the SNr, such as the depolarization of SNr cells through 5HT_2C receptors [5]. This 5HT_2C receptor activation likely enhances lateral inhibition amongst SNr cells to regulate ensembles of activity, but it is unclear when and how such effects interface with the serotonin-dependent suppression of low-frequency direct pathway inputs observed by Molinari et al. Under what conditions do dopamine and serotonin cooperate through the various available mechanisms in vivo? How might dopamine-serotonin interactions gate synaptic plasticity during motor learning paradigms? It also remains to be clarified when and how presynaptic D1 receptors on direct pathway terminals regulate transmission in the SNr, if at all.

The findings of Molinari et al. have relevance not only for healthy motor control but also for conditions such as Parkinson’s disease, dyskinesias, and even depression. In an acute dopamine depletion model (6-OHDA), direct pathway modulation of SNr is weakened [9]. However, progressive and chronic dopamine impairments may elicit distinct synaptic adaptations, and Molinari et al.’s study emphasizes the importance of scanning across physiological frequency ranges. Additionally, understanding how impairments in somatodendritic release in Parkinson’s disease, and levodopa treatment for Parkinson’s disease, may affect the function of the newly identified serotonergic mechanism will be important. Given that chronic fluoxetine, a selective serotonin reuptake inhibitor of the class that is commonly used to treat depression, sensitizes 5HT_1B receptors on D1-SPN terminals [4], the findings of Molinari et al. are also likely to have implications for depression treatment.