In DA Club of Reinforcement: Glutamate, It’s Your Birthday
2020; Cell Press; Volume: 107; Issue: 5 Linguagem: Inglês
10.1016/j.neuron.2020.07.036
ISSN1097-4199
AutoresAlvaro Nuno‐Perez, Manuel Mameli,
Tópico(s)Neural dynamics and brain function
ResumoDopamine release guides reward encoding, but the contribution of glutamate remains unclear. In this issue of Neuron, Zell et al. leverage the genetic ablation of dopamine synthesis from midbrain VGluT2 neurons to assess how glutamate shapes positive reinforcement. Dopamine release guides reward encoding, but the contribution of glutamate remains unclear. In this issue of Neuron, Zell et al. leverage the genetic ablation of dopamine synthesis from midbrain VGluT2 neurons to assess how glutamate shapes positive reinforcement. The year 1997 stands out for the seminal experiments performed by Schultz et al. (Schultz et al., 1997Schultz W. Dayan P. Montague P.R. A neural substrate of prediction and reward.Science. 1997; 275: 1593-1599Crossref PubMed Scopus (5400) Google Scholar), which became a milestone for modern neurobiology. A thirsty monkey faced two levers while learning to perform a simple-choice task. After illumination of a start cue, the monkey received a juice reward if he would exclusively press the left lever. While the animal engaged consecutive trials throughout the task, Schultz et al. (Schultz et al., 1997Schultz W. Dayan P. Montague P.R. A neural substrate of prediction and reward.Science. 1997; 275: 1593-1599Crossref PubMed Scopus (5400) Google Scholar) recorded the activity of midbrain dopamine neurons. Initially, these cells remained silent when the start cue lighted on, but they strongly increased their activity whenever the monkey would receive the reward. As the monkey continued to perform the task, however, the phasic increase in firing rate of dopamine neurons progressively shifted toward the start cue. These data represent the initial demonstration for the activity of midbrain dopamine neurons as the mechanistic underpinning for reinforcement learning. Many reports followed to further prove that dopamine release from ventral tegmental area (VTA) neurons is indeed sufficient to support positive reinforcement and reward learning, as well as to invigorate reward seeking (Figure 1). Thus, dopamine release is recognized as the primary reward signal in the brain (Keiflin and Janak, 2015Keiflin R. Janak P.H. Dopamine Prediction Errors in Reward Learning and Addiction: From Theory to Neural Circuitry.Neuron. 2015; 88: 247-263Abstract Full Text Full Text PDF PubMed Scopus (145) Google Scholar). If scientists had been satisfied without the need for further questioning after such pioneering work, this chapter of neurobiology would have ended and been set in stone. However, the inherent complexity of biological systems advocates for mechanisms’ going beyond the release of a sole neurotransmitter to underlie elaborate behaviors such as reinforcement. The VTA contains cells other than dopamine-producing ones, including some that synthesize and release glutamate (VTA VGluT2; Yamaguchi et al., 2015Yamaguchi T. Qi J. Wang H.L. Zhang S. Morales M. Glutamatergic and dopaminergic neurons in the mouse ventral tegmental area.Eur. J. Neurosci. 2015; 41: 760-772Crossref PubMed Scopus (76) Google Scholar). VTA VGluT2 represent ±30% of the entire midbrain neuronal population (Yamaguchi et al., 2015Yamaguchi T. Qi J. Wang H.L. Zhang S. Morales M. Glutamatergic and dopaminergic neurons in the mouse ventral tegmental area.Eur. J. Neurosci. 2015; 41: 760-772Crossref PubMed Scopus (76) Google Scholar). In situ hybridization approaches revealed co-expression of VGluT2 and Tyrosine-hydroxylase (Th), indicative that these neurons can produce both glutamate and dopamine (Yamaguchi et al., 2015Yamaguchi T. Qi J. Wang H.L. Zhang S. Morales M. Glutamatergic and dopaminergic neurons in the mouse ventral tegmental area.Eur. J. Neurosci. 2015; 41: 760-772Crossref PubMed Scopus (76) Google Scholar). Electrophysiological recordings in cultured neurons and acute brain slices support that dopamine-releasing neurons can drive excitatory postsynaptic currents via AMPA receptor activation (Stuber et al., 2010Stuber G.D. Hnasko T.S. Britt J.P. Edwards R.H. Bonci A. Dopaminergic terminals in the nucleus accumbens but not the dorsal striatum corelease glutamate.J. Neurosci. 2010; 30: 8229-8233Crossref PubMed Scopus (351) Google Scholar). Notably, co-immunoprecipitation of VGluT2 and the vesicular monoamine transporter in the Nucleus Accumbens (NAc) Figure 1 shell, together with the observation that glutamate entry promotes monoamine storage, suggests that glutamate-dopamine co-release might occur from the same vesicles (Hnasko et al., 2010Hnasko T.S. Chuhma N. Zhang H. Goh G.Y. Sulzer D. Palmiter R.D. Rayport S. Edwards R.H. Vesicular glutamate transport promotes dopamine storage and glutamate corelease in vivo.Neuron. 2010; 65: 643-656Abstract Full Text Full Text PDF PubMed Scopus (259) Google Scholar). The axons of VTA VGluT2 cells innervate and make functional synapses with virtually all medium-sized spiny neurons (MSNs) and interneurons of the NAc (Cai and Ford, 2018Cai Y. Ford C.P. Dopamine Cells Differentially Regulate Striatal Cholinergic Transmission across Regions through Corelease of Dopamine and Glutamate.Cell Rep. 2018; 25: 3148-3157.e3Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar; Hnasko et al., 2012Hnasko T.S. Hjelmstad G.O. Fields H.L. Edwards R.H. Ventral tegmental area glutamate neurons: electrophysiological properties and projections.J. Neurosci. 2012; 32: 15076-15085Crossref PubMed Scopus (172) Google Scholar; Qi et al., 2016Qi J. Zhang S. Wang H.L. Barker D.J. Miranda-Barrientos J. Morales M. VTA glutamatergic inputs to nucleus accumbens drive aversion by acting on GABAergic interneurons.Nat. Neurosci. 2016; 19: 725-733Crossref PubMed Scopus (105) Google Scholar). Moreover, optogenetic activation of either VTA VGluT2 cell bodies or their terminals in the NAc supports self-stimulation. Altogether, these previous observations open the possibility that this cell population might contribute to reinforcement learning and reward processing (Yoo et al., 2016Yoo J.H. Zell V. Gutierrez-Reed N. Wu J. Ressler R. Shenasa M.A. Johnson A.B. Fife K.H. Faget L. Hnasko T.S. Ventral tegmental area glutamate neurons co-release GABA and promote positive reinforcement.Nat. Commun. 2016; 7: 13697Crossref PubMed Scopus (87) Google Scholar). This scenario raises a brainteaser—does dopamine release from VTA VGluT2 neurons promote reinforcement? Or can glutamate itself be independently rewarding as well? Up until now, the literature lacked direct demonstrations that (1) midbrain glutamate neurons functionally release dopamine and that (2) glutamate alone can drive positive reinforcement without the concomitant neuromodulatory component. In this issue of Neuron, Zell et al. (Zell et al., 2020Zell V. Steinkellner T. Hollon N.G. Warlow S.M. Souter E. Faget L. Hunker A.C. Jin X. Zweifel L.S. Hnasko T.S. VTA Glutamate Neuron Activity Drives Positive Reinforcement Absent Dopamine Co-release.Neuron. 2020; (Published online June 30, 2020)https://doi.org/10.1016/j.neuron.2020.06.011Abstract Full Text Full Text PDF Scopus (16) Google Scholar) employ genetic manipulations to shed light on this matter, focusing on the functional relevance of glutamate release in the complete absence of dopamine co-transmission from VTA VGluT2 neurons. To test this, the authors began by generating mice in which the TH gene was conditionally deleted solely in VGluT2-expressing neurons by crossing VGluT2-Cre with Th-floxed mice. Using immunohistochemistry and in situ hybridization, the authors showed a complete lack of Th expression in glutamatergic neurons of these mice when compared to control littermates. Along with this observation, Th immunoreactivity also decreased within terminals projecting to the NAc. This genetic manipulation thereby sets the perfect experimental ground for understanding whether a subpopulation of glutamate neurons within the VTA co-releases dopamine. Cre-dependent expression of the excitatory opsin Channelrhodopsin-2 (ChR2) allowed the control of VTA VGluT2 neuronal activity. Optical stimulation of ChR2-expressing terminals within the NAc, coupled with fast-scan voltammetry, elicited robust dopamine transients in control mice but not in mice lacking Th from VTA VGluT2 neurons. Thus, glutamate neurons within the midbrain not only possess the machinery to produce dopamine, but they are also capable of releasing it onto downstream targets. VGluT2 modulates dopamine storage within vesicles, thereby controlling glutamate-dopamine co-release (Hnasko et al., 2010Hnasko T.S. Chuhma N. Zhang H. Goh G.Y. Sulzer D. Palmiter R.D. Rayport S. Edwards R.H. Vesicular glutamate transport promotes dopamine storage and glutamate corelease in vivo.Neuron. 2010; 65: 643-656Abstract Full Text Full Text PDF PubMed Scopus (259) Google Scholar). It is therefore plausible that deleting Th from VGluT2 neurons might in turn affect glutamate release. To address this issue, the authors examined optically evoked excitatory postsynaptic currents onto NAc MSNs. Glutamate release from VGluT2 neurons was qualitatively similar across D1- and D2-expressing MSNs. Furthermore, such glutamate release remained intact in mice lacking Th in VGluT2 neurons. Therefore, this genetic line represents an ideal tool for defining the duality between glutamate and dopamine from VGluT2 neurons for reward encoding and reinforcement. Two behavioral paradigms were employed by the authors: the intracranial self-stimulation (ICSS) and the real-time place preference (RTPP). In the ICSS configuration, mice could either nose-poke into hole A to trigger an optogenetic stimulation of VGluT2 terminals within the NAc or into hole B as the control without stimulation. Over sessions, mice developed a preference for the active hole. Notably, ablating Th from VTA VGluT2 neurons did not affect such behavior. In the RTPP paradigm, mice could freely access two distinct compartments, only one of which was paired with the optical stimulation of ChR2. Here, resembling the behavior of their control counterparts, mice devoid of dopamine release from VTA VGluT2 neurons spent less time in the active chamber, although they increased their approach rate into the active side. A caveat of this approach is the deletion of TH throughout developmental stages within the VGluT2 neuronal population. Indeed, VTA VGluT2 expression is larger early after birth than in adulthood. To overcome this issue, the authors employed an elegant approach through viral-based CRISPR/Cas9, enabling the disruption of dopamine synthesis and release from VGluT2 cells solely during adulthood. This led to a robust reduction in Th expression and in the dopamine transients recorded by fast-scan voltammetry. Similar to the results obtained with the Cre-Lox mice, CRISPR/Cas9-mediated ablation of Th exclusively in VTA VGluT2 neurons did not affect the ICSS of glutamatergic terminals within the NAc. Altogether, these results demonstrate that the activation of VTA VGluT2 neurons is sufficient to drive self-stimulation and approach behavior in the absence of concomitant dopamine release. The work from Zell et al. (Zell et al., 2020Zell V. Steinkellner T. Hollon N.G. Warlow S.M. Souter E. Faget L. Hunker A.C. Jin X. Zweifel L.S. Hnasko T.S. VTA Glutamate Neuron Activity Drives Positive Reinforcement Absent Dopamine Co-release.Neuron. 2020; (Published online June 30, 2020)https://doi.org/10.1016/j.neuron.2020.06.011Abstract Full Text Full Text PDF Scopus (16) Google Scholar) helps explain many previous studies and also generates many new questions spanning from the molecular to the circuit level. A point openly discussed by the authors is the display of a counter-intuitive avoidance behavior during the RTPP assay despite the increase in approach behavior. This observation is consistent with previous experimental results (Qi et al., 2016Qi J. Zhang S. Wang H.L. Barker D.J. Miranda-Barrientos J. Morales M. VTA glutamatergic inputs to nucleus accumbens drive aversion by acting on GABAergic interneurons.Nat. Neurosci. 2016; 19: 725-733Crossref PubMed Scopus (105) Google Scholar; Yoo et al., 2016Yoo J.H. Zell V. Gutierrez-Reed N. Wu J. Ressler R. Shenasa M.A. Johnson A.B. Fife K.H. Faget L. Hnasko T.S. Ventral tegmental area glutamate neurons co-release GABA and promote positive reinforcement.Nat. Commun. 2016; 7: 13697Crossref PubMed Scopus (87) Google Scholar). The authors speculate that the firing patterns preferred by animals differ between glutamate and dopamine release. Although mice favor brief stimulation of VTA VGluT2 neurons, longer stimulation of dopamine neurons is optimal for driving positive reinforcement. Yet, this raises the outstanding question of which activity patterns are specifically entrained in VTA VGluT2 neurons when confronting natural reinforcements. This is not trivial because how the midbrain achieves the discrimination between dopamine release from different neuronal sources remains a challenge. Understanding the neuronal firing during naturalistic reward paradigms might allow for the design and experimental replay of similar activity patterns. Would this artificial emulation recapitulate similar reinforcing processes? From a cellular and circuit standpoint, glutamate release from VTA VGluT2 neurons equally targets D1 and D2 MSNs. How is this synaptic control organized? One scenario is that a single VTA VGluT2 neuron releases glutamate onto both D1- and D2-expressing neurons. Alternatively, a more refined organization would consist of specific subsets of VGluT2 neurons devoted to the independent control of either D1 or D2 MSNs. The latter scenario would increase the entropy of the midbrain for the control of its downstream targets. Interestingly, VTA VGluT2 neurons also synapse onto NAc GABAergic interneurons (Qi et al., 2016Qi J. Zhang S. Wang H.L. Barker D.J. Miranda-Barrientos J. Morales M. VTA glutamatergic inputs to nucleus accumbens drive aversion by acting on GABAergic interneurons.Nat. Neurosci. 2016; 19: 725-733Crossref PubMed Scopus (105) Google Scholar). More refined experiments, as well as mathematical modeling, might help further clarify the computations performed by NAc neuronal populations upon VTA glutamate release, as opposed to the glutamate stemming from alternative brain regions (i.e., the cortex, amygdala, and hippocampus). Overall, Zell et al. (Zell et al., 2020Zell V. Steinkellner T. Hollon N.G. Warlow S.M. Souter E. Faget L. Hunker A.C. Jin X. Zweifel L.S. Hnasko T.S. VTA Glutamate Neuron Activity Drives Positive Reinforcement Absent Dopamine Co-release.Neuron. 2020; (Published online June 30, 2020)https://doi.org/10.1016/j.neuron.2020.06.011Abstract Full Text Full Text PDF Scopus (16) Google Scholar) provide convincing evidence for the functional relevance of VTA glutamate release in reinforcement and reward-guided behaviors (Figure 1). Moving forward, the conclusions as well as the approaches of this study can be adapted to investigate alternative behavioral conditions, such as avoidance, innate behaviors, or pathological phenotypes. Determining how the glutamate-dopamine duality shapes behavioral processes will certainly be enlightening. This work is supported by funding from the Swiss National Science Foundation ( 31003A ) to M.M. The authors declare no competing interests. VTA Glutamate Neuron Activity Drives Positive Reinforcement Absent Dopamine Co-releaseZell et al.NeuronJune 30, 2020In BriefActivation of VTA glutamate neurons leads to dopamine co-release in nucleus accumbens. Zell et al. genetically block this dopamine signal to show that VTA glutamate projections to nucleus accumbens can reinforce behaviors independently. These findings establish a parallel dopamine-independent mesolimbic pathway capable of supporting positive reinforcement. Full-Text PDF Open Archive
Referência(s)