Science - USA (2020-05-01)

INSIGHTS | PERSPECTIVES

sciencemag.org SCIENCE

PHOTO: NCMIR/TOM DEERINCK/SCIENCE SOURCE

and project to the expected brain targets of
ipRGCs such as the suprachiasmatic nucleus
(SCN), the circadian pacemaker that regu-
lates the body clock, and the olivary pretec-
tal nucleus (OPN), a structure involved in
pupillary light reflex. Furthermore, by opto-
genetically activating these GAD-expressing
ipRGCs, the authors show that these cells
make functional GABAergic synapses with
neurons in the SCN. Among SCN neurons
showing postsynaptic responses to activated
GAD-expressing RGC axons, ~30% received
GABAergic inhibitory inputs.
These findings allude to a previously un-
known level of complexity in ipRGCs, which
are distinct among RGCs for their intrinsic
photosensitivity. Unlike other RGCs that
solely rely on rod and cone photoreceptors in
the outer retina for light detection, ipRGCs

express the photopigment melanopsin and
thereby can be activated by light even in the
absence of rod and cone photoreceptor in-
puts ( 10 , 11 ). There are at least six different
types of ipRGCs in the mouse retina ( 12 , 13 ).
Future characterization of the morphology
and physiology of these GABAergic ipRGCs
in the retina will clarify whether these in-
hibitory minorities belong to the existing
type(s) or form a new type in the ipRGC fam-
ily. Notably, Sonoda et al. found that most of
the SCN neurons that received optogeneti-
cally activated GABAergic retinal inputs also
received simultaneous glutamatergic retinal
inputs, raising the intriguing possibility that
GABAergic ipRGCs may corelease glutamate
and GABA from their axonal terminals; this
awaits further testing.
Given the prominent axonal projections
of GABAergic ipRGCs in the SCN and the

OPN, Sonoda et al. examined the effect of

GABAergic signaling from these ipRGCs

on two non–image-forming behaviors me-

diated by these nuclei: respectively, cir-

cadian photoentrainment and pupillary

light reflex. Circadian photoentrainment

is the synchronization of the body’s inter-

nal clock with the solar cycle. Pupillary

light reflex assists the adaptation of vision

to different light levels by controlling the

diameter of the pupil to light intensity.

Glutamatergic ipRGC inputs to the SCN

and OPN inform these brain nuclei about

ambient light level of the visual environ-

ment ( 14 ). The retinal inputs interact with

the rest of the brain circuitries to synchro-

nize internal circadian rhythm with the so-

lar cycle and adjust pupil diameter. When

Sonoda et al. disrupted the GABAergic out-

put channel from the retina by genetically

deleting Gad in ipRGCs, they found that

the mutant mice were able to perform cir-

cadian photoentrainment more effectively

at lower light levels and had a more sensi-

tive pupillary light reflex in dim light com-

pared with that of control animals. Thus,

the authors propose that the GABAergic

ipRGC inputs dampen the sensitivity of

SCN and OPN circuits that are driven by

the more common glutamatergic ipRGC

inputs so that the non–image-forming be-

haviors are more resilient to smaller fluc-

tuations in brightness. 

How do GABAergic ipRGC inputs feed

into the circuits of non–image-forming

brain structures? Sonoda et al. have be-

gun exploring the wiring specificity of

GABAergic ipRGCs to SCN neuronal sub-

populations. They observed a similar pro-

portion of GABAergic retinal inputs to va-

soactive intestinal peptide (VIP)–positive

and VIP-negative SCN neurons, whereas

glutamatergic retinal inputs are more fre-

quently detected in the VIP-positive group.

Interpreting these results would require

a more comprehensive understanding of

SCN circuitry than is currently available.

Notably, a recent study in mice that com-

bined genetic labeling and a state-of-the-art

imaging technique called correlated light

and electron microscopy reveals exceptional

diversity in ipRGC axonal morphology and

presynaptic bouton connectivity with target

neurons in the SCN and the OPN ( 15 ).

In the SCN, ipRGC axonal terminals

preferentially target distal dendrites and

are enriched in a population of putative

GABAergic neurons that make dendro-den-

dritic synapses with each other ( 15 ). Future

efforts in determining the precise axonal

targeting patterns of GABAergic ipRGC ax-

ons in the SCN and the OPN will provide

insights into how the opposing retinal sig-

nals from GABAergic and glutamatergic

ipRGCs are integrated to set the sensitivity

of non–image-forming behaviors at the cir-

cuit level. Furthermore, Sonoda et al. also

showed evidence for GABAergic ipRGC

projections to image-forming regions such

as the medial posterior superior colliculus

and the shell of the dorsal lateral geniculate

nucleus. Assessing how this inhibitory reti-

nal channel can dually influence conscious

visual behavior would also be of future in-

terest. Overall, the work of Sonoda et al. has

opened the door for the study of functional

GABAergic ipRGC circuits in higher-order

visual centers and offers new and exciting

hypotheses for future exploration. j

REFERENCES AND NOTES

1. J. R. Sanes, R. H. Masland, Annu. Rev. Neurosci. 38 , 221
   (2015).


2. S. C. Massey, in Progress in Retinal Research, N. N.
   Osborne, G. Chader, Eds. (Pergamon, 1990), vol. 9, pp.
   399–425.


3. T. Sonoda et al., Science 368 , 527 (2020).


4. J. R. Wilson, M. J. Friedlander, S. M. Sherman, Proc. R.
   Soc. 8221 , 411 (1984).


5. T. Sakurai, T. Miyamoto, Y. Okada, Neurosci. Lett. 109 ,
   299 (1990).


6. R. R. Mize, G. D. Butler, J. Comp. Neurol. 371 , 633 (1996).


7. B.-C. Yu, C. B. Watt, D. M. K. Lam, K. R. Fry, Brain Res.
   439 , 376 (1988).


8. N. Lugo-García, R. E. Blanco, Brain Res. 564 , 19 (1991).


9. P. C. Rogers, D. V. Pow, Vis. Neurosci. 12 , 1143 (1995).


10. D. M. Berson, F. A. Dunn, M. Takao, Science 295 , 1070
    (2002).


11. S. Hattar et al., Science 295 , 1065 (2002).


12. L. Lazzerini Ospri et al., Annu. Rev. Neurosci. 40 , 539
    (2017).


13. L. E. Quattrochi et al., J. Comp. Neurol. 527 , 297 (2019).


14. C. S. Colwell, Nat. Rev. Neurosci. 12 , 553 (2011).


15. K.-Y. Kim et al., Cell Rep. 29 , 628 (2019).

ACKNOWLEDGMENTS

This work was supported by NIH 1R01NS109990 and NSF

GRFP DGE-1746045 (J.D.).

10.1126/science.abb7529

The retina is connected to the brain by retinal ganglion cells (orange) and their optic nerve fibers (red).

472 1 MAY 2020 • VOL 368 ISSUE 6490