Chemogenetic Activation of Melanopsin Retinal Ganglion Cells Induces Signatures of Arousal and/or Anxiety in Mice

Abstract

Functional imaging and psychometric assessments indicate that bright light can enhance mood, attention, and cognitive performance in humans. Indirect evidence links these events to light detection by intrinsically photosensitive melanopsin-expressing retinal ganglion cells (mRGCs) [1-9]. However, there is currently no direct demonstration that mRGCs can have such an immediate effect on mood or behavioral state in any species. We addressed this deficit by using chemogenetics to selectively activate mRGCs, simulating the excitatory effects of bright light on this cell type in dark-housed mice. This specific manipulation evoked circadian phase resetting and pupil constriction (known consequences of mRGC activation). It also induced c-Fos (a marker of neuronal activation) in multiple nuclei in the hypothalamus (paraventricular, dorsomedial, and lateral hypothalamus), thalamus (paraventricular and centromedian thalamus), and limbic system (amygdala and nucleus accumbens). These regions influence numerous aspects of autonomic and neuroendocrine activity and are typically active during periods of wakefulness or arousal. By contrast, c-Fos was absent from the ventrolateral preoptic area (active during sleep). In standard behavioral tests (open field and elevated plus maze), mRGC activation induced behaviors commonly interpreted as anxiety like or as signs of increased alertness. Similar changes in behavior could be induced by bright light in wild-type and rodless and coneless mice, but not melanopsin knockout mice. These data demonstrate that mRGCs drive a light-dependent switch in behavioral motivation toward a more alert, risk-averse state. They also highlight the ability of this small fraction of retinal ganglion cells to realign activity in brain regions defining widespread aspects of physiology and behavior.

Figure 1

Chemogenetic Activation of mRGCs (A and B) After intravitreal injection of a viral vector (AAV2-hSyn-DIO-hM3Dq-mCherry) to Opn4^Cre:Z/EGFP mice, immunohistochemical staining revealed transgene (mCherry; red) expression in GFP-positive neurons (green) in an en face view of retinal whole mounts (A) and retinal section (B). The retinal section shows the expression of transgene in cells of the retinal ganglion and inner nuclear cell layers (GCL and INL) of hM3Dq Opn4^Cre/+ mice. Notice the different soma sizes of transduced cells (arrows), with DAPI stain in blue. A monochrome version of mCherry staining and more details on hM3Dq expression are provided in Figure S1. (C) Representative images of eyes under infrared illumination from hM3Dq-expressing mice held in darkness, prior to (top) and at 20, 120, and 180 min after intraperitoneal (i.p.) injection of CNO (5 mg/kg). For more details, see Figure S2. (D) Representative double-plotted actogram of wheel running activity of a hM3Dq mouse housed under a light:dark cycle (the light phase is indicated in yellow) until day 12 followed by constant darkness. The star shows the start of a 2 hr presentation of 0.25 mg/mL CNO in drinking water; the red line to left represents the eye fit through activity onsets. A representative double-plotted actogram of a control mouse is shown in Figure S3. (E) Change in circadian phase CNO application to hM3Dq-expressing (open bars; n = 4) and control (filled bars; n = 5) mice (means ± SEM; Mann-Whitney U test, p = 0.015). (F) Representative micrographs of coronal section through the SCN labeled for c-Fos (dark) from hM3Dq and control mice after CNO administration (5 mg/kg, i.p., at CT14). (G) Mean (±SEM) number of c-Fos positive cells mm⁻² in SCN sections from hM3Dq (n = 6) and control (n = 6) mice (two-tailed unpaired t test,^∗∗p < 0.01).

Figure 2

c-Fos Activity Mapping after Chemogenetic Activation of mRGCs (A and B) Representative micrographs of coronal sections showing c-Fos labeling (dark) in (A) the paraventrical hypothalamic nucleus (PVN), the dorsomedial hypothalamus/dorsal hypothalamic area (DMH/DHA), lateral (perifornical) hypothalamic area (LH), and ventrolateral preoptic nucleus (VLPO); and (B) amygdala (Amg), intralaminar thalamic nuclei (ITL), paraventricular thalamus (PVT), lateral habenula (LHb), and nucleus accumbens (NAc). Images to right of each panel are from control mice, and those to left are from unilateral hM3Dq-expressing animals (transduced eye to right of presented image, except in LH, VLPO, Amg, and NAc, where only the contralateral region is shown). (C) Mean (±SEM) number of c-Fos-positive cells mm⁻² in all brain regions (bilaterally) after CNO administration (5 mg/kg, i.p., at CT14) in hM3Dq-expressing (open bars; n = 8) and control (filled bars; n = 8) mice (two-tailed unpaired t test, ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001). A complete summary of c-Fos data is provided in Table S1. BLA, basolateral amygdala; CeA, central nucleus of the amygdala. (D) Brain diagram illustrating target areas analyzed.

Figure 3

Chemogenetic and Light Activation of mRGCs Alters Performance in Behavioral Tests (A) Time spent in center over 10 min under dim far-red illumination in an open arena in hM3Dq-expressing and control Opn4^Cre/+ mice treated with CNO or saline. Open bars depict data from hM3Dq-expressing mice, and filled bars depict data from control Opn4^Cre/+ mice. (B–D) Time spent in open arms (B) and the number of entries to open (C) and closed (D) arms of an elevated plus maze under dim far-red illumination in hM3Dq and control Opn4^Cre/+ mice treated with CNO or saline. (E) Time spent by visually intact (Opn4^Cre/+) and rodless and coneless (rd1;Cnga3^−/−) mice in center of an open arena under bright or dim far-red light. (F) Time spent by visually intact (Opn4^Cre/+) and melanopsin knockout (Opn4^−/−) mice in center of an open arena under bright or dim far-red light. All behavioral tests undertaken between CT14 and CT17; n = 9–13 per group, two-way ANOVA with post hoc Bonferroni correction, ^∗p < 0.05, ^∗∗p < 0.01. All graphs depict mean ± SEM, with open bars depicting hM3Dq-expressing mice and closed bars depicting control Opn4^Cre/+ mice, except in (E) and (F), where closed bars depict dim light and hatched bars depict bright light. In all cases, dim far-red light was 2.91 μW cm⁻², λ > 680 nm, and bright light was white, 217 μW cm⁻².