JGP study (Horie et al. https://doi.org/10.1085/jgp.202413749) explains why mice lacking TRPM1 exhibit oscillatory firing of their retinal ganglion cells, and suggests that the same mechanism causes similar oscillations in other outer retinal diseases.
Retinal ganglion cells (RGCs) are the output neurons of the retina that relay visual signals to the brain. Under certain pathological conditions, RGCs can display spontaneous oscillatory activity. The resulting “noise” disrupts visual information processing and can cause hallucinations, but the mechanisms underlying these oscillations are unclear. In this issue of JGP, Horie et al. reveal why RGC oscillations occur in Trpm1 knockout mice, a model for congenital stationary night blindness, and suggest that the same mechanism drives oscillations in rd1 mice, a model for the degenerative disease retinitis pigmentosa (1).
Clockwise from top left: Sho Horie, Katsunori Kitano, Masao Tachibana, and Chieko Koike.
TRPM1 forms cation channels at the dendritic tips of ON bipolar cells (BCs) (2). TRPM1 channels are inhibited by the glutamate receptor mGluR6, but when photons of light interrupt the release of glutamate from rod and cone photoreceptors, the channels open and depolarize ON BCs. The ON signal is then transmitted either directly to RGCs or indirectly via amacrine cells. This ON response is lost in the absence of TRPM1 or mGluR6, and mice lacking either protein serve as models for congenital stationary night blindness. But Trpm1 knockout mice also exhibit oscillatory firing of their RGCs, a phenotype that is not observed in mGluR6-deficient rodents (3). “We wondered what the difference is that causes pathological oscillations in Trpm1 KO but not mGluR6 KO retinas,” says Chieko Koike, a professor at Ritsumeikan University in Japan.
Horie et al. show that the axon terminals of RBCs (green) are smaller and closer to the photoreceptor layer in Trpm1 KO retinas (second from left) than they are in wild-type (left) or mGluR6 KO (second from right) retinas. This reduces glutamatergic inputs to AII ACs, resulting in oscillatory firing of RGCs in Trpm1 KO, but not mGluR6 KO, mice. Notably, similar changes in the morphology of RBCs can be observed in rd1 retinas (right), suggesting that a similar mechanism causes pathological oscillations in retinitis pigmentosa.
Koike and colleagues, including first author Sho Horie and co-corresponding authors Katsunori Kitano and Masao Tachibana, found that both ON and OFF RGCs oscillate at ∼8–9 Hz in Trpm1 KO retinas, and that these oscillations are in-phase between RGCs of the same type, but anti-phase between pairs of ON and OFF cells. This suggested that the oscillations are driven by synaptic inputs from upstream retinal neurons that connect to both ON and OFF RGCs, rather than by any intrinsic properties of the RGCs themselves.
“Using whole-cell clamp techniques, we confirmed that the synaptic inputs to RGCs are periodic, and determined that ON and OFF RGCs receive periodic excitatory and inhibitory inputs, respectively,” Koike explains.
This led Koike and colleagues to focus on AII amacrine cells (AII ACs). These retinal interneurons make inhibitory, glycinergic synapses onto OFF RGCs, while also connecting via gap junctions to ON cone BCs that, in turn, make excitatory, glutamatergic synapses onto ON RGCs. Horie et al. found that glycine receptor antagonists reduced oscillations in OFF RGCs, while both ionotropic glutamate receptor antagonists and gap junction blockers suppressed oscillations in ON RGCs, supporting the idea that AII ACs drive the oscillations in Trpm1 KO retinas.
AII ACs are themselves excited by glutamatergic inputs from rod bipolar cells (RBCs), which are ON BCs and therefore regulated by TRPM1. Horie et al. found that compared with wild-type mice, the axon terminals of RBCs were smaller and shifted toward the photoreceptor layer in Trpm1 KO animals. RBCs were also significantly hyperpolarized in the absence of Trpm1. Together, these morphological and functional changes should reduce glutamatergic signaling from RBCs to AII ACs. Computer simulations revealed that this, along with the hyperpolarization of ON cone BCs in Trpm1 KO retinas, drives oscillations in AII ACs that, in turn, cause oscillations in RGCs.
In mGluR6 KO mice, in contrast, the size and position of RBC axon terminals were normal, suggesting that in these animals, glutamatergic signaling from RBCs to AII ACs is maintained and can prevent the development of oscillations. Koike and colleagues think that because TRPM1 levels are only slightly reduced in the absence of mGluR6, channel activity is sufficient to support normal RBC maturation.
Intriguingly, rd1 mice that are used as a model for retinitis pigmentosa show, in addition to a progressive loss of photoreceptors, pathological RGC oscillations that are very similar to those seen in Trpm1 KO mice (4). The drastic structural changes in rd1 retinas, however, have made it difficult to determine the cause of oscillations in these animals. Horie et al. found that TRPM1 levels are significantly reduced at the dendritic tips of RBCs in rd1 mice and that, similar to Trpm1 KO, the axon terminals of RBCs are small and stop short of AII ACs.
“This strongly suggests that functional and morphological alterations in RBCs are a common mechanism contributing to pathological oscillations in retinitis pigmentosa,” Koike says. “RBCs therefore emerge as a promising therapeutic target for suppressing these aberrant retinal signals.”
Koike and colleagues now plan to confirm that the morphological changes in RBCs do, indeed, reduce synaptic input to AII ACs in Trpm1 KO mice, and investigate whether the pathological oscillations can be suppressed by expressing TRPM1 or other cation channels such as channelrhodopsin.
