Bidirectional modulation of pain-related behaviors in the zona incerta

Central amygdala neurons expressing protein kinase C-delta (CeA-PKCδ) are sensitized following nerve injury and promote pain-related responses in mice. The neural circuits underlying modulation of pain-related behaviors by CeA-PKCδ neurons, however, remain unknown. In this study, we identified a functional monosynaptic inhibitory neural circuit that originates in CeA-PKCδ neurons and terminates in the ventral region of the zona incerta (ZI), a subthalamic structure previously linked to pain processing. Behavioral experiments further show that chemogenetic inhibition of GABAergic ZI neurons is sufficient to induce bilateral hypersensitivity in uninjured mice as well as contralateral hypersensitivity after nerve injury. In contrast, chemogenetic activation of GABAergic ZI neurons reverses nerve injury-induced hypersensitivity, demonstrating that silencing of the ZI is required for injury-induced behavioral hypersensitivity. Our results identify a previously unrecognized inhibitory efferent pathway from CeA-PKCδ neurons to the ZI and demonstrate that ZI-GABAergic neurons can bidirectionally modulate pain-related behaviors in mice.

. 110 As summarized in Table 1, dense to moderate labeling was consistently seen in  Representative high magnification images of the CeA in a coronal brain slice of a ChrimsonR-tdTomato injected mouse. ChrimsonR-transduced cells are shown in red and neurons immunostained for PKCδ in cyan. The merged image is shown on the right panel. Lower insets depict higher magnification images of the areas delineated by the white box in the upper images. White arrowheads highlight representative transduced cells that are also positive for PKCδ. Scale bars represent 100 µM for low magnification and 10 µM for high magnification images. (C) Summary diagram illustrating CeA-PKCδ neuron efferent projections within the brain. Forebrain regions are shown in yellow, hypothalamic structures in red, thalamus in green, midbrain in blue and pons in fuchsia. The thickness of the arrows depict the density of labeling (sparse, moderate or dense). (D) Low magnification representative images of brain regions with axonal terminals from CeA-PKCδ cells. Insets in each image are high magnification images depicting axonal terminals within the regions delineated by the boxes in the respective low magnification images. Scales are 1000 µM for low magnification images and 20 µM for high magnification images. bed nucleus of stria terminalis medial (BNST-MA); bed nucleus of stria terminalis fusiform nucleus (BNST-Fu); extended amygdala (EA); substantia innominate (SI); lateral preoptic area (LPO); globus pallidus (GP); lateral hypothalamus (LH); Subthalamic nucleus (STh); Zona incerta (ZI); Parasubthalamic nucleus (PSTh); periaqueductal grey (PAG); substantia nigra (SNR); ventral tegmental area (VTA); pedunculopontine tegmental nucleus (PTg); laterodorsal tegmental nucleus (LDTg); reticular formation (mRT); lateral parabrachial (LPB); locus coeruleus (LC). See Figure1 -figure supplement 1.  Table 1. Semi-quantitative analysis of the density of axonal terminals in brain regions from 5 PKCδ-Cre mice stereotaxically injected with a cre-dependent adeno-associated virus anterograde tracer (ChrimsonR, mCherry, EGFP) into the CeA. Rightmost column is from experiment 265945645 of the Mouse Brain Connectivity Atlas of the Allen Brain Institute (http://connectivity.brain-map.org/). -no expression; + sparce; ++ moderate; +++ dense The ZI was among the regions identified as an efferent target of CeA-PKCδ 133 neurons in our anatomical experiments ( Figure 1C). This subthalamic brain region was   VGAT-Cre mice were stereotaxically injected with AAV-DIO-hM4Di-mCherry into the ZI. Current-clamp recordings were obtained from mCherry-positive cells in acute ZI slices 2 weeks after the injection. Representative traces of whole-cell current-clamp recording obtained from ZI neurons transduced with hM4Di-mCherry before (left) and after (right) bath application of 10 µM CNO (lower panel) or vehicle (top panel). Action potential were elicited using 500-ms depolarizing current injection that evoked 2 to 5 action potential before bath the application. The same amplitude of depolarizing current injection was used before and after bath application. Summary graphs depicting the mean ± SEM number of spikes before and after bath treatment are shown on the right panel (n = 6 neurons per treatment; *p<0.05 for ACSF vs CNO).   activation of VGAT-positive ZI neurons led to significant (p < 0.0001) reversal of cuff-245 induced hypersensitivity in the hindpaw ipsilateral to cuff implantation, while no 246 measurable effects were seen in saline-injected or CNO-injected mCherry-control mice.

247
The effects of activation of VGAT-positive ZI neurons were specific to nerve injury as 248 withdrawal thresholds in sham treated mice or the hindpaw contralateral to sciatic nerve 249 cuff were comparable between groups. As illustrated in Figure 7, reversal of cuff-induced 250 hypersensitivity was also modality-specific as behavioral responses to cold and heat  injected in the ZI (Figure 2). These results are consistent with previous studies that have 290 shown that somatostatin-expressing neurons in the CeA, which have virtually no overlap

299
In the present study, we show that chemogenetic manipulation of the activity of 300 VGAT-positive ZI neurons bidirectionally modulate pain-related behaviors in a modality-301 specific manner (Figures 4-7). Thus, we observe bilateral hypersensitivity to tactile (but 302 not thermal) stimulation following inhibition of VGAT-positive ZI neurons in the absence 303 of nerve or tissue injury (Figures 4-5). In contrast, chemogenetic activation of VGAT-

324
Baseline responses in sham animals or the hindpaw contralateral to sciatic nerve  the same process of sciatic nerve exposure and stretching but no tubing was implanted.

418
After the procedure was complete, the skin above the thigh was closed with wound clips.

419
The mice were subjected to at least one-week recovery period before performing the 420 behavior tests.    depolarizing current (10 to 120 pA) was injected to elicit between 2 and 5 action potentials.

512
The current injection repeated every 15 seconds until the cell fired stably and consistently.

513
Following this stabilization, 10 additional recordings were acquired before bath application