Corticostriatal mechanisms of action learning and habit formation

已鉴定神经元功能的活体研究技术的不断发展 我们的主要目标之一是了解基底节内的回路如何有助于控制和学习动作。这一系列研究的一个重要组成部分是了解该回路中特定神经元亚型在动作和动作序列的学习、规划、启动和执行过程中的活动。我们与Champalimaud神经科学研究所的Rui Costa博士和NIAAA分子生理学实验室的Steven Vogel博士合作，开发了一种体内纤维光度测量技术，用于在体内进行纹状体的时间相关单光子计数(TCSPC)荧光测量。将这项技术与在直接或间接途径的纹状体介质棘突投射神经元中选择性表达荧光钙指示蛋白(GCaMP)相结合，使我们能够测量反映神经元激活的钙瞬变。我们发现，在运动序列的启动过程中，这两条路径上的神经元是共同活跃的，特别是当动物离开植入纤维探针的半球时。 我们现在正在将这项技术扩展到测量不同脑深部区域不同神经元亚型的细胞内钙信号。我们还在探索使用我们最近开发的一种新的单纤维慢性植入光度系统来测量清醒、自由活动的小鼠突触前元件中钙瞬变的可能性。我们感兴趣的是测量细胞内钙以外的信号分子，以评估这些信号通路在行为执行过程中的动态。基于Forster共振能量转移(FRET)的遗传编码检测器已被开发用于测量各种细胞内信号分子。我们的光纤光度系统的TCSPC能力应该允许我们利用这些分子的供体荧光团的荧光寿命变化来测量FRET的变化。在细胞系中的实验表明，我们的系统具有检测这种变化的灵敏度，我们目前正在体内测试一些检测器。最终，这些实验应该允许我们测量皮质纹状体回路中已识别的神经元中的各种细胞内信号。 眶前额-纹状体突触上的大麻素1受体控制目标定向行为和习惯性行为之间的平衡 我们以前的研究表明，内源性大麻素和CB1大麻素受体都参与了皮质纹状体突触传递的短期和长期抑制。此外，实验室中的研究表明，CB1受体在工具性习惯学习中发挥着关键作用，这是在食物奖励的杠杆按压任务中测得的。然而，这些受体表达在突触前轴突终末在皮质-基底神经节回路中的许多位置，在那里它们可以参与这种行为。眶前叶皮质(OFC)投射到背内侧纹状体(DMS)，在维持目标导向的工具杠杆按压方面具有重要作用，可能是通过提供关于结果价值的信息。因此，我们推测CB1介导的OFC到DMS突触的抑制可能通过抑制这种与结果相关的信号来帮助习惯性反应的发展。为了测试这种低血压，我们试图有选择地敲除这一通路中的CB1受体，并检查对工具性学习和表现的影响。我们使用携带CB1受体的等位基因的小鼠，以允许在Cre重组酶活跃的细胞中有条件地敲除该蛋白。我们最初使用了三种方法在OFC中选择性地敲除CB1。第一种策略是用他莫昔芬诱导表达Cre的小鼠，用FLOLED-CB1小鼠饲养，并将他莫昔芬注射到OFC中。我们还向OFC注射了一种病毒，其中Cre的表达是由一个启动子驱动的，该启动子在大多数(如果不是所有的)神经元中都工作得很好，以及一种含有CamKIIpha启动子的病毒来驱动OFC投射神经元的表达。在所有这三种条件下，小鼠在工具性任务中表现出目标定向反应(例如，它们缺乏习惯学习)。然而，OFC投射神经元支配着大脑的几个区域。为了确定CB1在投射到DMS的神经元中的作用，我们使用了双病毒注射策略，在OFC中感染Cre重组酶结构，使其依赖于Flipase的表达，并在CB1-Flobled小鼠的DMS中表达含有Flipase的病毒。通过这种交叉策略，我们再次在小鼠中观察到习惯学习受损，这表明CB1在OFC-DMS投射神经元中处于从目标导向向习惯学习的转变过程中。在这些实验中使用的所有小鼠在其他方面都是健康的，在仪器任务中表现出正常的杠杆按压率、奖赏提取和消耗。总体而言，我们的发现支持这样的观点，即内源性大麻素/CB1介导的抑制OFC-DMS谷氨酸能传递在从目标导向到习惯性动作控制的转变中具有重要作用，可能是通过抑制关于结果价值的信息。

Continued Development of Techniques for In Vivo Investigation of the Function of Identified Neurons One of our major aims is to understand how circuitry within the basal ganglia contributes to the control and learning of actions. An important component of this line of research is gaining an understanding of the activity of specific neuronal subtypes within this circuitry during the learning, planning, initiation and performance of actions and action sequences. We have worked with Drs. Rui Costa of the Champalimaud Neuroscience Institute and Steven Vogel of the Laboratory of Molecular Physiology at NIAAA to develop a technique for in vivo fiber photometry to perform Time Correlated Single Photon Counting (TCSPC) fluorimetry in the striatum in vivo. Coupling this technique with selective expression of a fluorescent calcium indicator protein (GCaMP) in either direct or indirect pathway striatal medium spiny projection neurons has allowed us to measure calcium transients reflective of neuronal activation. We found that neurons in the two pathways were co-active during initiation of movement sequences, particularly when the animal was moving away from the hemisphere in which the fiber probe was implanted. We are now extending this technique to measure intracellular calcium signals in different neuronal subtypes in different deep brain regions. We are also exploring the possibility of measuring calcium transients in presynaptic elements in awake, freely-moving mice using a new single-fiber chronically-implantable photometry system that we recently developed. We are interested in measuring intracellular signaling molecules other than calcium, in order to assess the dynamics of these signaling pathways during behavioral execution. Genetically-encoded detectors based on Forster Resonance Energy Transfer (FRET) have been developed to measure a variety of intracellular signaling molecules. The TCSPC capability of our fiber photometry system should allow us to measure changes in FRET using fluorescence lifetime changes in the donor flourophore of these molecules. Experiments in cell lines indicate that our system has the sensitivity to detect such changes, and we are currently testing some detectors in vivo. Ultimately, these experiments should allow us to measure a variety of intracellular signals in identified neurons within the corticostriatal circuitry. Cannabinoid 1 Receptors on Orbitofrontal-Striatal Synapses Control the Balance between Goal-Directed and Habitual Behavior Our previous studies indicated that endocannabinoids and CB1 cannabinoid receptors are involved in both short- and long-lasting depression of synaptic transmission at corticostriatal synapses. Furthermore, studies in the laboratory have shown that CB1 receptors play a crucial role in instrumental habit learning, as measured in a food-rewarded lever-pressing task. However, these receptors are expressed on presynaptic axon terminals in many locations throughout the cortico-basal ganglia circuitry where they could participate in such behaviors. The orbitofrontal cortex (OFC) projects to the dorsomedial striatum (DMS), and has important roles in maintaining goal-directed instrumental lever-pressing, presumably by providing information about outcome value. Thus, we reasoned that CB1-mediated depression at OFC to DMS synapses may aid in the development of habitual responding by dampening this outcome-related signal. To test this hypoethsis, we sought to selectively knock out CB1 receptors in this pathway and examine effects on instrumental learning and performance. We used mice carrying a floxed allele of the CB1 receptor to allow for conditional knockout of the protein in cells where the Cre recombinase is active. We initially used three approaches to selectively knock out CB1 in the OFC. The first strategy used a tamoxifen-inducible Cre-expressing mouse bred with the floxed-CB1 mouse, and tamoxifen injection into OFC. We also injected into the OFC a virus in which Cre expression is driven by a promoter that works well in most, if not all, neurons, as well as a virus containing a CamKIIalpha promoter to drive expression in OFC projection neurons. Under all three of these conditions, mice showed goal-directed responding in the instrumental task (e.g. they were deficient in habit learning). However, OFC projection neurons innervate several brain regions. To determine the role of CB1 in neurons projecting to the DMS, we used a dual-virus injection strategy in which a Cre recombinase construct that confers flipase-dependent expression was infected in OFC and a virus containing flipase was expressed in DMS of CB1-floxed mice. With this intersectional strategy we again observed impaired habit learning in mice, implicating CB1 in OFC-DMS projection neurons in transition from goal-directed to habit learning. All of the mice used in these experiments were otherwise healthy and showed normal lever-pressing rates, reward retrieval and consumption in the instrumental task. Overall, our findings support the idea that endocannabinoid/CB1-mediated suppression of OFC-DMS glutamatergic transmission has important roles in transition from goal-directed to habitual action control, presumably through suppression of information about outcome value.