Far-red LED light alters circadian rhythms and elicits dark-adapted ERG responses in rodents

Abstract

Rodents are assumed to be blind to red light, thus red light is often used in the dark phase of a light/dark cycle to facilitate study procedures using nocturnal rodents. However, effects of red light in dark phase on behaviors and circadian rhythms in rodents are not yet clear. Thus, we evaluated effects of various long wavelength red light-emitting diode (LED) light on circadian rhythm and electroretinogram (ERG) in C57BL/6J mice and Wistar Han rats. Animals were implanted with telemetry devices to measure body temperature, heart rate, blood pressure, and locomotor activity for circadian rhythm assessment. In contrary to infra-red light, all visible long wavelength red lights, including the far-red LED light with a peak at 741 nm, induced significant alterations in circadian rhythms and dark-adapted rod photoreceptor-mediated ERG responses in mice and/or rats. However, far-red light did not elicit light-adapted cone photoreceptor-mediated ERG responses in both mice and rats. These findings demonstrate that rodents can perceive all spectrum of long wavelength red lights that are visible to humans, and exposures of red lights in dark phase interfere with their circadian rhythms. A dim far-red LED with peak wavelength in the range of 740-760 nm is recommended to use in the dark phase of a rodent room, and potential impacts are considered when using red light >2 photopic lux.

Fig 1. Light spectrum.

White LED light, 642 ± 10 nm (peak wavelength± peak half width) red, 666 ± 11 nm photo-red, 741 ± 23 nm far-red, and 940 ± 26 nm infra-red LED lights were used in circadian experiments at approximately 50 µW/cm² inside of cages (i.e., 172, 60, 25, 1, and 0 lux, respectively). The 850 ± 9 nm infra-red LED was used in ERG experiments.

Fig 2. Circadian entrainment or phase shift.

A) Light protocol for mouse entrainment experiment. Delayed 12:12h light:dark cycle (dusk signal) was used to entrain mice or to observe phase shift. B) Light protocol for rat entrainment experiment. Advanced 12:12h light/dark cycle (dawn signal) was used to entrain rats or to observe phase shift. C) Double plotted mouse body temperature (TEMP) actogram (i.e., 2 consecutive days plotted next to each other). D) Double plotted rat heart rate (HR) actogram. E) Circadian period obtained from onset of TEMP circadian rhythm in mice. F) Circadian period obtained from onset of HR circadian rhythm in rats. G) 12h light-induced TEMP circadian phase responses (partial phase response curve) in mice (details see S5 Fig). Daily phase shift was calculated from TEMP onset and subtracted by daily phase shift in DD. Data were then binned into 2-hour circadian time (CT) intervals. H) 12h light-induced HR circadian phase responses (partial phase response curve) in rats (details see S6 Fig). White bar: 12h light-on phase; black bar: 12h light-off phase. Open rectangle = light-on period. CT 0-12 = subjective light phase; CT 12-24 = subjective dark phase. LD = 12:12h white light:dark; FD = 12:12h far-red light:dark; PD = 12:12h photo-red light:dark; DD = 24h constant darkness. Data: mean ± SD. N = 4 mice/8 rats. *Significantly different between groups (E&F), or different from FD and zero (G), or different from PD and zero (H) (P < 0.05, ANOVA). #Significantly different from PD (G) (P < 0.05, ANOVA).

Fig 3. Free-running circadian rhythm under constant light condition.

A) Double plotted mouse body temperature (TEMP) actogram. B) Double plotted rat heart rate (HR) actogram. C) Circadian period obtained from onset of mouse TEMP circadian rhythm. D) Circadian period obtained from onset of rat HR circadian rhythm. LD = 12:12h light:dark; DD = constant darkness; FF = constant far-red; PP = constant photo-red; RR = constant red; LL = constant white light. Data: mean ± SD. N = 4 mice/ 8 rats. *Significantly different between light conditions (P < 0.05, ANOVA).

Fig 4. Two-hour test light in the dark phase.

A & C & E) Mouse body temperature (TEMP) profiles. B&D) Rat heart rate (HR) profiles. F) Double plotted mouse TEMP actogram. G) Double plotted rat HR actogram. White bar: light phase; black bar: dark phase. Black line: standard light/dark cycle; red line: 2h-test light in the dark phase of a standard light/dark cycle. Test lights were placed at ZT 16-18 daily for approximately one week as indicated. ZT 0-12 = light phase; ZT 12-24 = dark phase. LD = 12:12h white light:dark. Profile data: means. N = 4 mice/ 7 rats.

Fig 5. Twelve-hour test light in the whole dark phase.

A & C & E) Mouse body temperature (TEMP) profile. B & D & F & G) Rat heart rate (HR) profiles. H) Double plotted mouse TEMP actograms. I) Double plotted rat HR actograms. J) Rat HR circadian amplitude (mean ± SD). White bar: light phase; black bar: dark phase. Black line: standard light/dark cycle; red line: 12h-test light in the whole dark phase of a standard light/dark cycle. Test lights were placed at ZT 12-24 daily for approximately one week as indicated. ZT 0-12 = light phase; ZT 12-24 = dark phase. LD = 12:12h white light:dark; LI = 12:12h white light:infra-red; LF = 12:12h white light:far-red; LP = 12:12h white light:photo-red; LR = 12:12h white light:red light. Profile data: means. N = 4 mice/ 7 rats. *P < 0.05, ANOVA.

Fig 6. Dark-adapted and light-adapted electroretinogram (ERG) responses in C57BL/6J mice.

A) Representative dark-adapted ERG traces recorded after LED light flashes at 4200 µW/cm² for 5 ms. Animals were adapted to the darkness for at least 2 hours prior to the dark-adapted ERG recording. B) Representative light-adapted ERG traces recorded after LED light flashes at 4200 µW/cm² for 5 ms. Animals were exposed to white LED light at 42 µW/cm² for 10 min for light adaptation prior to the testing. C) Dark-adapted b-wave amplitude. D) Light-adapted b-wave amplitude. E) Dark-adapted b-wave latency. F) Light-adapted b-wave latency. G) Mean dark-adapted ERG b-wave amplitude from all levels of stimuli. H) Mean light-adapted ERG b-wave amplitude from all levels of stimuli. Data: mean ± SD. N = 7. *Significantly different between groups (P < 0.05, ANOVA).

Fig 7. Dark-adapted and light-adapted electroretinogram (ERG) responses in Wistar Han rats.

A) Representative dark-adapted ERG traces recorded after LED light flashes at 4200 µW/cm² for 5 ms. Animals were adapted to the darkness for at least 2 hours prior to the dark-adapted ERG recording. B) Representative light-adapted ERG traces recorded after LED light flashes at 4200 µW/cm² for 5 ms. Animals were exposed to white LED light at 42 µW/cm² for 10 min for light adaptation prior to the testing. C) Dark-adapted b-wave amplitude. D) Light-adapted b-wave amplitude. E) Dark-adapted b-wave latency. F) Light-adapted b-wave latency. G) Mean dark-adapted ERG b-wave amplitude from all levels of stimuli. H) Mean light-adapted ERG b-wave amplitude from all levels of stimuli. Data: mean ± SD. N = 6. *Significant different between groups (P < 0.05, ANOVA).

Fig 8. Free-run circadian period and ERG responses.

A) Mouse body temperature free-run circadian period under different wavelength constant LEDs at 50 µW/cm² (Aschoff effect). Assuming free-run circadian period under constant white light is equivalent to that under constant monochromatic 550 nm LED which is close to the maximum responses; and free-run circadian period under constant darkness is equivalent to that under constant 940 nm infra-red which is close to the minimum responses. LED IC50 wavelength was obtained through sigmoid fitting [Y = Bottom + (X^Hillslope)*(Top-Bottom)/(X^HillSlope + IC50^HillSlope)} the mean data. That is to say, LED light with a peak wavelength equal to IC50 will have 50% of maximum responses which correspond to white LED exposure. Greater IC50 wavelength means more sensitive to long wavelength LED exposures. B) Rat heart rate free-run circadian period under different wavelength constant LEDs at 50 µW/cm² (Aschoff effect). C) LED IC50 wavelength obtained from free-run circadian period (under constant light at 50 µW/cm²) and dark-adapted b-wave amplitude (at 4200 µW/cm² for 5 ms) – comparisons between mice and rats. D) LED wavelength dependent dark-adapted b-wave amplitude in C57BL/6J mice. Sigmoid fit was performed for each flash light intensity level. E) LED wavelength dependent dark-adapted b-wave amplitude in Wistar Han rats. F) LED IC50 wavelength from dark-adapted and light-adapted b-wave amplitudes (at 4200 µW/cm² for 5 ms) – comparisons between dark-adapted and light-adapted responses. G) LED wavelength dependent light-adapted b-wave amplitude in C57BL/6J mice. H) LED wavelength dependent light-adapted b-wave amplitude in Wistar Han rats. I) Correlation between mean free-run circadian period (n = 8 at 50 µW/cm² constant lights) and mean dark-adapted b-wave amplitude (n = 6 at 4200 µW/cm² for 5 ms) obtained from Wistar Han rats with different wavelength LEDs. Each dot corresponds to each light source used. Data: mean ± SD. Circadian data N = 4 mice/8 rats; ERG data N = 7 mice/6 rats. *Significantly different between groups (P < 0.05, ANOVA).

Fig 9. Quantification of species-specific α-opic light exposures associated with circadian and ERG responses in rodents.

A) Mouse body temperature circadian responses plotted with mouse melanopic EDI (equivalent daylight illuminance). Circadian response (delta tau): lengthening of free-running circadian period under a constant test light (e.g., far-red, photo-red) vs constant dark (DD). Hill function y = b2+(b1-b2)/(1+(x/b3)^b4) was used to fit the data obtained from the circadian experiment in individual animals to generate species-specific α-opic light exposures associated with 5% (EC5) and 50% (EC50) of maximum responses and circadian responses associated with species-specific 0.1 lux melanopic and 0.1 lux rod-opic EDI light exposures. B) Mouse delta tau plotted with mouse melanopic irradiance. C) Rat heart rate circadian responses (delta tau) plotted and fitted with rat melanopic EDI. D) Rat delta tau plotted with rat melanopic irradiance. E) Mouse delta tau plotted with mouse rod-opic EDI. F) Mouse delta tau plotted with human photopic illuminance (brightness to human eyes) of the light sources. G) Rat delta tau plotted with rat rod-opic EDI. H) Rat delta tau plotted with human photopic illuminance of the light sources. I). Mouse dark-adapted ERG responses (b-wave amplitude) plotted with mouse melanopic EDI. J). Mouse ERG b-wave amplitude plotted with mouse rod-opic EDI. K) Rat dark-adapted ERG b-wave amplitude plotted with rat melanopic EDI. L) Rat ERG b-wave amplitude plotted with rat rod-opic EDI. M) Mouse ERG b-wave amplitude plotted with human photopic illuminance of the light sources. N) Rat ERG b-wave amplitude plotted with human photopic illuminance of the light sources. O) Mouse α-opic EDIs at 0.1 lux melanopic EDI for D65 (standardized daylight), white, red, photo-red, far-red, and infra-red (850nm) LEDs. P) Mouse circadian (delta tau) and dark-adapted ERG (b-wave amplitude) responses at 0.1 lux melanopic EDI for different light sources. Q) Rat α-opic EDIs at 0.1 lux melanopic EDI for D65, white, red, photo-red, far-red, and infra-red (850nm) LEDs. R) Rat circadian (delta tau) and dark-adapted ERG (b-wave amplitude) responses at 0.1 lux melanopic EDI for different light sources. S) Mouse α-opic EDIs for white, red, photo-red, and far-red LEDs that associate with 5% of maximum dark-adapted ERG responses (ERG EC5). T). Illustrate 5% of max ERG responses (ERG EC5) that correspond with panel S). U) Rat α-opic EDIs for white, red, photo-red, and far-red LEDs that associate with 5% of maximum dark-adapted ERG responses (ERG EC5). V). Illustrate 5% of max ERG responses (ERG EC5) that correspond with panel U). Dotted lines indicate recommended 0.1 lux melanopic EDI night light limit for lab animals. For D65, it would be 0.1 lux α-opic EDI for other opsins as shown in panel O) and Q). However, this is not the case for different long wavelength LEDs, which exhibit much greater rod and M-cone opic light exposures at 0.1 lux melanopic EDI. In addition, an α-opic measurement represents the level of light exposure to a particular opsin. However, this species-specific α-opic light exposure may not necessarily contribute to circadian or dark-adapted ERG responses. Data: mean ± SD. Circadian data: N = 4 mice/ 7 rats; ERG data: N = 6 mice/ 5 rats.