Evidence in primates supporting the use of chemogenetics for the treatment of human refractory neuropsychiatric disorders

Abstract

Non-human primate (NHP) models are essential for developing and translating new treatments that target neural circuit dysfunction underlying human psychopathology. As a proof-of-concept for treating neuropsychiatric disorders, we used a NHP model of pathological anxiety to investigate the feasibility of decreasing anxiety by chemogenetically (DREADDs [designer receptors exclusively activated by designer drugs]) reducing amygdala neuronal activity. Intraoperative MRI surgery was used to infect dorsal amygdala neurons with AAV5-hSyn-HA-hM4Di in young rhesus monkeys. In vivo microPET studies with [¹¹C]-deschloroclozapine and postmortem autoradiography with [³H]-clozapine demonstrated selective hM4Di binding in the amygdala, and neuronal expression of hM4Di was confirmed with immunohistochemistry. Additionally, because of its high affinity for DREADDs, and its approved use in humans, we developed an individualized, low-dose clozapine administration strategy to induce DREADD-mediated amygdala inhibition. Compared to controls, clozapine selectively decreased anxiety-related freezing behavior in the human intruder paradigm in hM4Di-expressing monkeys, while coo vocalizations and locomotion were unaffected. These results are an important step in establishing chemogenetic strategies for patients with refractory neuropsychiatric disorders in which amygdala alterations are central to disease pathophysiology.

Keywords: DREADDs; amygdala; anxiety; behavioral inhibition; clozapine; depression; non-human primate; rhesus; stress.

Graphical abstract

Figure 1

Representative images of hM4Di-HA expression in the rhesus amygdala (A) The rhesus amygdala is composed of several nuclei, of which the central nucleus (medial and lateral divisions in purple) and basal nucleus (magnocellular and intermediate divisions in blue) were stereologically analyzed in subject P1. (B) Co-labeling of NeuN (left panels; greyscale) and HA (middle panels; red) in subject P1 revealed that neuronal expression of hM4Di-HA varied between subregions of the amygdala, with little coexpression observed in the central nucleus (top panels) compared to the magnocellular and intermediate portions of the basal nucleus (bottom panels). Images are maximum intensity projections integrated across the z stack. Scale bar (white) is 50 μm. (A) adapted with permission from Paxinos et al.

Figure 2

Autoradiographic demonstration of hM4Di-HA expression in the rhesus amygdala (A) Autoradiograms from subject P2 unilaterally injected with AAV5-hSyn-HA-hM4Di (hM4Di-HA) into the amygdala, showing nonspecific and total [³H]clozapine (3.5 nM) binding. The white dotted line depicts the amygdala boundaries. (B) Densitometric quantification of specific binding ([³H] clozapine total binding – nonspecific: total binding was assessed in three coronal slices) in the viral-infected amygdala compared to the uninfected amygdala. Error bars represent +/- SEM.

Figure 3

Effects of clozapine administration on freezing behavior during NEC Log transformed freezing duration following administration of vehicle or two different doses of clozapine (n = 5/group). Compared to vehicle, there was a significant reduction in freezing duration following 0.5 mg/kg clozapine but not 0.1 mg/kg clozapine (∗∗p < 0.01).

Figure 4

Effects of 0.03 mg/kg and 0.1 mg/kg clozapine on plasma ACTH and cortisol levels in animals selected for the DREADD behavioral experiment As part of characterizing individual doses of clozapine, effects on ACTH (upper) and cortisol (lower) were assessed following administration of vehicle, 0.03, and 0.1 mg/kg clozapine to the same animals (n = 8). ACTH was significantly reduced by both doses, whereas significant reductions in cortisol occurred at the higher dose (∗p < 0.05, ∗∗p < 0.01).

Figure 5

DREADDs experiment methods (A) Timeline for behavioral testing of monkeys receiving AAV5-hSyn-HA-hM4Di (n = 5) and cage-mate controls (n = 5). (B) Overlap of the infused area across all subjects as assessed by the gadolinium signal detected with MRI. The colors represent the number of animals with gadolinium signal at each voxel. The overlap of the gadolinium injection clouds across all five experimental animals is depicted in yellow demonstrating consistent coverage of the dorsal amygdala region using the MRI-guided targeted injection procedure. (C) Neurons expressing hM4Di-HA in a section containing the accessory basal nucleus in a maximum intensity projection image at 40× integrated across the z stack; scale bar (yellow) represents 100 μm. Neurons are identified with NeuN in gray, and hM4Di-HA expression with HA immunoreactivity in red. Note, the neuron seen in center of the field of view is prominently expressing hM4Di-HA in its soma, as well as in its neuronal extensions.

Figure 6

[¹¹C]DCZ binding in the amygdala demonstrates hM4Di-HA expression in vivo The average difference (hM4Di-HA minus control) in [¹¹C]DCZ binding signal from two of the hM4Di-HA animals in the DREADDs behavioral experiment is overlayed on a rhesus monkey brain MRI template. The average difference image is thresholded at 40%. Greater binding is observed bilaterally in the amygdala, where the AAV5-hSyn-HA-hM4Di virus was infused and where gadolinium signal was observed across all five hM4Di-HA subjects. The peak difference in the right amygdala reached 53% greater binding signal in the hM4Di-HA animals. No other brain regions exceeded this threshold. See also Figure S2.

Figure 7

Quantification of hM4Di-HA expression in the amygdala (A) Intraoperative MRI image of the gadolinium signal after the infusion into the dorsal amygdala of AAV5-hSyn-HA-hM4Di mixed with gadolinium. (B) Acetylcholinesterase staining of a tissue section containing the amygdala from subject E1, demonstrating considerable staining in the basal nucleus and a lack of staining in the central nucleus. (C) A depiction of the amygdala nuclei as derived from acetylcholinesterase staining. The box outlined in red represents the region of one of the sections that was used in the stereological analysis. (D and E) A 200 × 200 × 10 μm subregion within the accessory basal nucleus that was part of the stereological analysis (D), which is a magnification of the green square in (E); scale bar (yellow) represents 50 μm. Red fluorescence indicates HA immunoreactivity, and gray fluorescence indicates NeuN immunoreactivity. (E) A tissue section used for stereological assessment of the co-expression of HA (red) and NeuN (greyscale). White squares indicate subregions where co-expression was quantified. (F) A heatmap of the percent of neurons, identified with NeuN, that are co-expressing HA, reflecting the stereological analysis performed on the section depicted in (E). See also Figure S3.

Figure 8

DREADD-induced change in anxiety-related behavioral response The clozapine-induced change in freezing duration (clozapine-vehicle) occurring during the NEC component of the HIP for the hM4Di-HA and control animals. The five different symbols each identify an hM4Di-HA subject and its cage-mate control. Symbols in black represent the difference between clozapine and vehicle at baseline prior to surgery (pre). Symbols in red represent the difference between clozapine and vehicle after hM4Di-HA expression that is averaged across the two rounds of testing (post). A significant group by session (pre/post) interaction was observed (F_1,8 = 14.89, p < 0.01). Post hoc testing revealed a significantly greater decrease in freezing for the hM4Di-HA animals post compared to pre (∗∗p < 0.01), whereas the control animals did not significantly differ post compared to pre.