Figure 1.
Conditional knockout of VGLUT2 from mRGCs.
(A) Breeding scheme for obtaining conditional VGLUT2 knockouts (cKO) and controls. (B) Mice were genotyped by PCR with allele-specific primers. (C–H) VGLUT2 immunoreactivity is detected in melanopsin-immunoreactive retinal ganglion cells in wild-type (WT) (C, F) and control mice (D, G) but not cKO mice (E, H). (C, F) flatmount retina; (D, E, G, H) retinal slices. Arrows point to retinal ganglion cells that are immunopositive for melanopsin. Scale bar: 50 um. (I) Mean optokinetic response, a measure of image-forming vision, was not different between control (n = 8 eyes) and cKO (n = 7 eyes) mice (P = 0.4, Student’s t-test). (J, K) Typical action potential responses of neonatal (P8) mRGCs to light in control (J) and cKO (J) retinas. These electrophysiological findings demonstrate that mRGCs in VGLUT2 cKO pups retained their light responses.
Figure 2.
Glutamatergic signaling from mRGCs is required for photoaversion and light-evoked ultrasonic vocalizations in neonatal mice.
(A) Schematic of photoreception and visual pathways in the retina of neonatal mice. Until postnatal day 10, mice rely on intrinsic photoreception in mRGCs for light-mediated functions as visual signaling from cones, rods and bipolar cells has not yet matured. After P10, visual signals from rods and cones excite RGCs including mRGCs. The schematic shows that neurotransmission from mRGCs relies on glutamate and PACAP. (B) Experimental setup to test neonatal photoaversion (see methods). (C) Examples of movement and vocal responses to light in control and in cKO mice. Exposure to light evoked increased movement and distress USVs in the control but not in the cKO mouse. The timing of individual USVs is marked by ticks at the top of the graph. Locomotive movement is presented as the amount of net movement in kiloPixels per frame (see Methods). (D, E) Summary of light-induced responses. The number of USVs (D) and the amount of net movement (E) significantly increased in the control mice (n = 7 for USVs and n = 9 for movement) but not in the cKO mice (n = 7 for USVs and n = 10 for movement). Red lines are median values. Asterisk (*) indicates P<0.05; (**) indicates P<0.01; (n.s.) indicates P>0.05 (Wilcoxon signed rank test).
Figure 3.
Removal of glutamatergic signaling from mRGCs unmasks secondary signaling pathway for pupillary light responses.
(A) Schematic of photoreception in the retina of adult mice. Visual signals from photoreceptors are relayed to mRGCs and other retinal ganglion cells via VGLUT1-dependent signaling. MRGCs signal to brain via VGLUT2-dependent signaling and PACAP. (B) Experimental setup to test the pupillary light reflex (PLR). Adult mice were dark adapted and then placed into a restraining chamber. The consensual PLR was monitored with an infrared camera. (C) Typical PLR responses in control and cKO mice. (D) Mean normalized pupil diameter in control and cKO mice during 1 min before light onset, during 1 min light and during 1 min after the light was turned off. Pupil diameter was measured every 2 sec and then normalized to 10-sec baseline before the light onset. Shaded areas mark one standard deviation from the mean. (E) Percent maximum constriction in response to 1 min light. Pupils constricted on average 80% in control mice (n = 8) but only 50% in cKO mice (n = 15). Error bars mark one standard deviation from the mean. Asterisks (***) indicate P<0.001 (Student’s t-test). The incomplete constriction in cKO mice was not due to defects in the iris sphincter muscle or its cholinergic innervation as topical application of pilocarpine constricted the cKO pupils to the control levels (n = 4). (F) Mean time to maximum constriction during 1 min after light onset was 11 sec in control mice (n = 8) and 36 sec in cKO mice (n = 15). Error bars mark one standard deviation from the mean. Asterisks (***) indicate P<0.001 (Student’s t-test).