Intrinsically Photosensitive Retinal Ganglion Cells (ipRGCs) Are Necessary for Light Entrainment of Peripheral Clocks

The experimental procedures used in this study were approved by the Institutional Animal Care and Use Committee at the University of Minnesota and are in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Prior to tissue collection, mice were humanely euthanized by isoflurane or carbon dioxide exposure followed by cervical dislocation and all efforts were made to minimize suffering. The animal room was maintained at a 12-h light/dark (LD) cycle (i.e., lights on from 06:00 to 18:00 h). Zeitgeber times (ZTs) of ZT0 and ZT12 were used as the lights-on and lights-off times, respectively. The light intensity at the surfaces of the cages was approximately 500 lux. Mice were fed normal, commercial rodent chow and provided with water ad libitum.

We generated Opn4::TeNT mice by inter-crossing Opn4^Cre/+ mice (Cre recombinase inserted in the melanopsin genetic locus)[15] with the R26^{floxstop-TeNT} mice in which the tetanus light chain subunit (TeNT) is inserted downstream of a loxP-flanked STOP sequence into the R26 locus [16]. Opn4::TeNT mice (Opn4^Cre/+, R26^TeNT/TeNT) were compared to control littermates (Opn4^+/+, R26^TeNT/TeNT) with the exception of multi-electrode array (MEA) and pupillary light reflex measurements where we also included Opn4^+/+, R26^TeNT/TeNT and Opn4^Cre/+, R26^TeNT/+ mice. We used female and male mice in all experiments unless indicated.

Immunohistochemistry of brain slices or retinas was performed essentially as described [12]. Briefly, mice were anesthetized with isoflurane and perfused transcardially with 1× phosphate-buffered saline (PBS), pH 7.4, followed by 20 ml of 4% paraformaldehyde in PBS. The brains were extracted and fixed overnight in the same fixative at 4°C. Each brain was sectioned at 100 μm on a Vibratome 1000 (Vibratome, St. Louis, MO). For flat-mount analysis of the retina, the eyes were enucleated and the anterior portion and lens discarded, and the remaining “eyecup” was further fixed in 4% paraformaldehyde in PBS for 1 hr at 4°C. Retinas were isolated then rinsed in PBS (pH 7.4) for 1 hour. Free-floating sections of brain or whole retinas were blocked and permeabilized overnight at 4°C in PBS supplemented with 10% goat serum and 0.5% Triton X-100. The sections were then incubated in primary antibody for two-three days at 4°C, rinsed with PBS, and incubated overnight in secondary antibody at 4°C. After rinsing in PBS for 1 hour, sections were mounted directly to glass slides and covered with the antifade agent Vectashield (Vector Lab, Burlingame, CA). Some retinas were processed as vertical (transretinal) sections using a cryostat (Leica, Buffalo Grove, IL). In this case, the posterior eyecup was immersion fixed in 4% paraformaldehyde in PBS for 15 or 30 min at room temperature. After washing in PBS, the eyecup was cryoprotected in graded sucrose solutions (10, 20, and 30% in PBS) and frozen in embedding medium. Vertical sections (10–20 μm) were cut and collected on superfrost slides. Antibody staining and processing of the retinal sections followed the procedures employed for flat-mount retinas with the exception that the incubation in primary antibodies was abbreviated to an overnight period.

The primary antibodies used in this study were: rabbit anti-melanopsin (1:500)[17], rabbit anti-C-FOS (1:10000)(Calbiochem), mouse anti-glutamine synthetase (1:200) (BD Transduction Laboratories, Franklin Lakes, NJ), goat anti-calretinin (1:200)(Swant, Switzerland), mouse anti-PKC alpha (1:200)(Novus Biologicals, Littleton, CO). We used donkey anti-rabbit Alexa-488 and donkey anti-mouse Alexa-594, donkey anti-goat Alexa-594, as secondary antibodies (Invitrogen, Grand Island, NY, 1:750). Confocal z stacks were acquired using an Olympus FluoView FV1000 confocal microscope with 20X, 40X, or 60X oil-immersion objectives. Editing of images was limited to adjusting the brightness and contrast levels using ImageJ software (NIH, Bethesda, MD). Experiments were replicated with a minimum of 3 mice.

Mice under ketamine-xylazine (100–10 mg/kg im) anesthesia were implanted with CMA 20 (10 mm) probes (Harvard Apparatus, Holliston, MA) subcutaneously in the dorsal neck region. Following tethering to a liquid swivel (Instech Labs, Plymouth Meeting, PA), mice were sampled beginning 2 days post-surgery at 60 min intervals for 3 days under a 12:12 h light-dark (LD) cycle. Free corticosterone in dialysate was assayed by ¹²⁵I-corticosterone RIA kit (MP Biomedicals, Santa Ana, CA). A dialysate pool was found to dilute in parallel with the corticosterone standard curve (data not shown); dialysate samples were diluted 1:3 to insure that values fell on the linear part of the standard curve. The in vivo microdialysis data were analyzed using the CircWave 1.4 software (Dr. R. Hut, http://www.euclock.org↗) to test the daily rhythmicity of corticosterone levels and the 24-h rhythm was confirmed if p < 0.05.

Control and Opn4::TeNT mice were interbred with mPER2^Luc mice [18]. Mutant mice (Opn4^Cre/+, R26^TeNT/TeNT, PER2^Luc/+) were compared to control mice (Opn4^+/+, R26^TeNT/TeNT, PER2^Luc/+). Mice were sacrificed with CO₂ asphyxiation/cervical dislocation 2–4 h before lights-off (ZT 8–10). Peripheral tissues (adrenal gland, cornea, lung, liver, spleen and anterior pituitary) were rapidly dissected and placed in ice-cold Hank's solution. Each tissue explant was placed on a membrane (0.4μm, 30 mm in diameter, Millicell cell culture inserts; Millipore, Billerica, MA, USA) in a 35 mm Petri dish, sealed with vacuum grease, and cultured in 1.1 ml DMEM containing 10 mM HEPES, 0.1 mM Luciferin (Promega, Madison, WI), 25 units/ml penicillin, 25 μg/ml streptomycin (Gibco, Grand island, NY), and 2% B27 (Gibco) as described previously [19]. Cultures were incubated at 37°C, and bioluminescence was monitored for 1 min at 7.5-min intervals using a LumiCycle luminometer (Actimetrics, Evanston, IL). Bioluminescence recordings were detrended using a 24 h moving average-subtraction method and smoothed by a 2 h moving average. The first peak in the smoothed data after 24 h in vitro was used as a phase marker. Phase angles and variances of circular plots were calculated using Oriana (Kovach Computing Services, Wales, UK).

Mice (6–24 weeks of age) were housed within a temperature-controlled facility in individual cages equipped with running wheels that were maintained in ventilated chambers that permitted control of the light-dark cycles. The animals were maintained on a 12-h light (~ 150 lux, white light): 12-h dark (LD 12:12 h) (T24) cycle for at least 3 weeks and then transferred to the ventilated chambers. The intensity of light at 360, 440, 480, 540, 570 and 610 nm was 1.29 · 10¹², 2.05 · 10¹², 2.21 · 10¹², 2.38 · 10¹², 2.19 · 10¹² and 2.16 · 10¹² photons · cm⁻² · s⁻¹, respectively. Irradiance measurements were made with a calibrated radiometer model S370 (UDT Instruments, San Diego, CA) fitted with 10 nm bandpass filters. Manipulation in constant dark (DD) conditions were performed under dim red light. Light negative masking was assessed by recording running wheel activity for mice submitted to a T7 (3.5-h light: 3.5-h dark) cycle [20] over several days. Activity data were recorded continuously by a PC system and displayed and analyzed by using CLOCKLAB software (Actimetrics). The free-running period was calculated (days 1–10 in DD) by using a χ2 periodogram. The amplitude of the circadian component was estimated from a normalized Fourier spectrum by using 10 days in DD. Original data were collected at one-min intervals.

Mice were anesthetized by intraperitoneal injection of a solution containing ketamine (100 mg/ml) and xylazine (10 mg/ml) at a dose of 0.1 ml /100 g body weight. The vitreous body of the left eye was injected with 2 μl of a 0.1% solution of cholera toxin B protein (CTB) conjugated to Alexa Fluor 488 (ThermoFisher Scientific, Waltham, MA) in sterile PBS. Following the injection, the needle was held in place for 30 sec to prevent leakage. The same procedure was repeated for the right eye with injection of CTB conjugated to Alexa Fluor 594. Mice were perfused with 4%oParaformaldehyde in PBS intracardially 2 days after the injection. Brain were processed and imaged as described above for immunohistochemistry.

After removal from the eye, a patch of retina about 4–10 mm² was mounted on a Multi-electrode array (Multichannel Systems, Reutlingen, Germany), ganglion cell side down, and perfused with oxygenated Ames’ medium at 35°C supplemented with 20 μM CNQX ((6-cyano-7-nitroquinoxaline-2,3-dione)) and 50 μM D-APV (D(−)-2-Amino-5-phosphonopentanoic acid) to block glutamatergic transmission. The activity of ganglion cells was recorded via 256 electrodes 30 μm in diameter spaced every 100 μm apart and arranged in a 16×16 square grid (Multi Channel Systems MCS GmbH). Full-field visual stimuli were presented during recordings using a high brightness LED (LuxeonStar 5, luxeonstar.com↗) with a peak wavelength of 480 nm. The light stimulation protocol consisted of 1min stimuli of increasing irradiance (5.10¹⁰ photons/cm²/s to 5.10¹³ photons/cm²/s). The current through the LED was controlled using custom electronics and software written in Matlab (Mathworks, Natick, MA) and aligned with the physiological recording with a resolution of ±100 μs. Signal is acquired from all 256 channels @ 10 kHz. Negative thresholds for spike detection were set at 5 times the standard deviation of the noise on each channel. Spike cutouts, consisting of 1 msec preceding and 2 msec after a suprathreshold event, along with a time stamp of the trigger were written to hard disk. For each electrode, these spike cutouts were sorted into trains of a single cell after recording using Offline Sorter (Plexon, Denton, TX). Data analysis and display were performed using Neuroexplorer (Plexon) and custom software written in Matlab.

Mice were implanted with an acrylic headpost. After at least 1 week of recovery from the headposting surgery, they were tested for PLR. On the day of the testing, 15 min before the light stimulation, dark-adapted mice received a drop of tropicamide in the left eye. Before the recordings, mice were briefly anesthetized with isofluorane and restrained in a custom-made animal holder. The animal holder was placed inside a light tight box with the left eye apposed against an opening of an integrating sphere. Light from a 300 Watt Xenon Arc lamp light source (Sutter Instrument, Novato, CA, USA) was filtered, collimated and delivered to the integrating sphere through a liquid light guide. An inline 480 nm filter, a filter wheel with a neutral density filter and a Lambda 10–3 optical filter changer with SmartShutter^TM were used to control the spectral quality, intensity and duration of light. Light intensity was measured with apower meter (Melles Griot, Rochester, NY). The mouse’s right eye was illuminated by an IR LED and recorded with a high precision LINX video camera (Imperx Inc., Boca Raton, FL) equipped with an IR filter at a sample rate of 7Hz. We recorded sequences consisting of 1 min of darkness, 1 min of monochromatic 480 nm light (1.10¹⁰ photons/cm².s and 1.10¹⁴ photons/cm².s) and finally 3 min of darkness. Digital movies of pupil constriction were analyzed with a custom Matlab (Mathworks®) program. We extracted the pupil diameter. The mean diameter measured during the first period of darkness (1 min) of each sequence served as the baseline for normalization of the recordings.

Statistical analyses were performed using OriginPro 8.1 (Origin Lab, Northampton, MA) or GraphPad InStat (GraphPad, LaJolla, CA). Statistical comparison of means was performed using a Student's t test or one-way analysis of variance (ANOVA) followed by Tukey–Kramer multiple comparison tests. In some cases, comparison of means were performed by Kruskal-Wallis non-parametric ANOVA, followed by Dunn’s multiple comparison test. Data are presented as mean ± SE.

To suppress ipRGC neurochemical signaling, we used targeted expression of tetanus toxin light chain fragment (TeNT) on ipRGCs. TeNT inhibits chemical neurotransmission by cleaving the synaptic vesicle membrane protein VAMP2/Synaptobrevin-2 with consequent decrement or suppression of synaptic signaling of neurons that express it [21]. Selective expression of TeNT in ipRGCs was achieved by breeding Opn4^Cre/+ mice in which Cre recombinase is inserted in the melanopsin (Opn4) genetic locus [15] with R26^{floxstop-TeNT} mice in which expression of the TeNT is dependent on Cre recombinase activity [16]. Mutant mice (Opn4^Cre/+, R26^TeNT/TeNT) r, showed no apparent abnormalities during development although an extensive assessment of their behavior or anatomy has not been carried out.Retinas of adult Opn4^Cre/+, R26^TeNT/TeNT mice had an apparent normal morphology with comparable thickness of the outer nuclear and inner nuclear layers (S1 Fig). Immunostaining for various immunocytochemical markers (calretinin, Protein Kinase C-α, glutamine synthetase) revealed similar expression in the mutant mice (S1 Fig). Melanopsin staining in whole mount retinas or retinal sections showed the presence of ipRGCs including the outer stratifying M1 cells, inner stratifying M2 and displaced M1 cells. Comparable numbers of melanopsin-positive retinal ganglion cells were found in the control (Opn4^+/+, R26^TeNT/TeNT) and mutant (Opn4^Cre/+, R26^TeNT/TeNT) retinas (S2 Fig). Although TeNT is fused to GFP, we failed to detect the presence of GFP in the mutant retinas possibly due to the low levels of expression of TeNT. Injection of fluorescently conjugated cholera toxin into the eye, which labels all ganglion cell fibers from the retina, showed that fibers that innervated NIF visual targets such as the SCN (S3 Fig), intergeniculate leaflet (IGL) and ventral lateral geniculate nucleus (vLGN) were preserved in the Opn4^Cre/+, R26^TeNT/TeNT mice (S3 Fig).

We examined the light-evoked responses of ipRGCs in isolated retinas of mutant and control mice using multi electrode array (MEA) recordings (Fig 1). We recorded from early postnatal mice given that ipRGCs are the only light responsive cells at this developmental stage [22, 23]. As expected, we observed light responses with increasing firing rates with higher levels of light stimulation. During the 60 second period of light simulation the firing rate at higher light intensities reached a maximum followed by substantial progressive decline. Upon termination of the light stimulus, ipRGCs continued to display substantial spike activity that persisted for several seconds. The relatively sluggish initiation and termination of spike responses are consistent with previous reports of melanopsin-evoked light responses [23]. Overall there were no significant differences across genotypes indicating that expression of TeNT in ipRGCs did not impact their overall responsiveness to light.

To examine whether expression of TeNT in ipRGCs suppressed NIF behaviors, we measured consensual PLR at two light intensities on awake control and mutant mice. Following a four-hour dark period adaption, PLR was evoked by a 60 second 480 nm light stimulus and pupil responses were recorded in the contralateral eye using a low-light video camera. In control littermate mice (Opn4^+/+, R26^TeNT/TeNT and Opn4^+/+, R26^+/+) the pupils constricted both at low (1X10¹⁰ photons/cm².s) and high (1X10¹⁴ photons/cm².s) light intensities (Fig 2A and 2B). In contrast, Opn4^Cre/+, R26^TeNT/TeNT mice had a significant deficit of PLRs even at high light intensities as reflected by the relatively small constriction of pupils (normalized pupil constriction: ctrl Opn4^+/+, R26^TeNT/TeNT: 0.66 ± 0.15; mut: -0.01 ± 0.01; p<0.001, unpaired t-test). Interestingly, an intermediate level of pupil constriction (0.17 ± 0.07) was observed in the Opn4^Ccre/+, R26^TeNT/+ mice indicating a gene dosage effect on the TeNT-mediated synaptic silencing of ipRGCs Lack of light-evoked responses in the Opn4^Cre/+, R26^TeNT/TeNT was not due to malformation of the pupillary constriction apparatus as topical application of 100 mM carbachol in the mutant mouse eye (n = 3) was able to elicit full pupillary constriction. Thus we conclude that PLRs are deficient in the Opn4^Cre/+, R26^TeNT/TeNT mice in accordance with previous observations on the role of ipRGCs in this particular NIF behavior [24, 25]. Based on these results we used the Opn4^+/+, R26^TeNT/TeNT as control mice and the Opn4^Cre/+, R26^TeNT/TeNT as mutant mice and referred as Opn4::TeNT hereafter.

The immediate early gene c‐Fos has been used as a marker of light‐induced neuronal activity in SCN neurons with marked increase in its expression with light pulses delivered at subjective night [26, 27]. Because the Opn4::TeNT mice did not show light entrainment in locomotor activity (see following section), we used the onset of locomotor activity at circadian time (CT)12 as a phase reference point for both control and mutant mice. Mice were kept in constant dark (DD) for two days and then a 30 min light pulse was administered at CT16; mice were kept in DD for 60 min until perfusion. Using immunohistochemistry, we confirmed that c-FOS as measured by the number of immunopositive cells is induced in the SCN of control mice (n = 151.7 ± 16.6)(Fig 3). By comparison, the Opn4::TeNT mice showed significantly lower levels of induction of c-FOS within the SCN (n = 16.8 ± 4.0) (p<0.001, One-way ANOVA, Bonferroni's post hoc test for pairwise comparison); which was not significantly different from sham control mice (n = 14.0 ± 3.2)(p>0.05, One-way ANOVA, Bonferroni's post hoc test for pairwise comparison). These findings indicate the lack of ipRGC neurochemical signaling to the SCN in the Opn4::TeNT mice.

Behavioral analysis of circadian locomotor activity was carried out to ascertain whether Opn4:TeNT mice align their activity to LD cycles. In these experiments, mice were individually housed in cages containing a running wheel, and their daily locomotor activity was recorded. The animals were initially exposed to regular LD cycles (12:12 h). In these conditions, the control mice showed normally entrained daily patterns of activity, initiating their nightly bouts of activity at the beginning of the dark period and exhibited nocturnal rhythms in activity as reflected by the period length (tau) of 23.98 ± 0.01 (n = 12)(Fig 4A and 4B). All Opn4::TeNT mice, however, recorded in LD 12:12 conditions failed to entrain their circadian rhythms to external light cues and showed a “free running” rhythm with a tau of 23.46 ± 0.08 (n = 12). There was no difference in free-running periods of Opn4::TeNT mice kept on LD 12:12 or DD conditions (tau = 23.56 ± 0.11, n = 12)) (p>0.05, One-way ANOVA, Bonferroni's post hoc test for pairwise comparison) suggesting that the SCN circadian clock in these animals oscillates independently of external light conditions.

Suppression of locomotor activity in nocturnal rodents by light (light negative masking) is another NIF behavior associated with light signaling by ipRGCs [25]. This behavior was assessed by exposing the mice to an ultradian light regimen of 3.5:3.5 light/dark cycles (T7) for two to three weeks (Fig 4C). This light regimen measures the effects of light independent of circadian rhythms as the mice fail to entrain to light cycles that move across the circadian cycle [28]. Activity in the dark was normalized to the total activity to yield the relative activity in the dark. The control group exhibited a relative activity in the dark of 0.92 ± 0.02 indicating a strong bias for activity to the dark periods. The mutant group, however, showed a relative activity in the dark of 0.55 ± 0.02 consistent with near equal activity during the light and dark periods (Fig 4C and 4D).

Robust self-sustaining PER2::LUC circadian rhythms have been measured in vitro for the SCN and peripheral tissues using the mPer2^Luc knockin mice [18]. In these mice, PER2::LUC rhythms measured in vitro likely reflect the in vivo circadian rhythms of PER2 protein [29]. Thus to examine the effect of ipRGC silencing on the photic entrainment of peripheral circadian oscillators, we crossed the Opn4::TeNT with mPER2^Luc mice. We kept the mutant and control mice on regular LD cycles (12:12 h) with free access to running wheels for several days and sacrificed the mice when their circadian locomotor activity became approximately antiphasic (8–12 h phase difference) (Fig 5A). Under these conditions, explant cultures of peripheral tissues from mutant and control mice showed robust rhythms of PER2::LUC that were approximately antiphasic (Fig 5B and Fig 6A). Similar to what has been previously reported [18], peripheral tissues of control mice displayed prominent circadian oscillations in PER2::LUC bioluminescence with a peak in the dark phase of the cycle between CT 17–20 (Fig 6B). However, explant cultures from the Opn4::TeNT mice showed peak PER2::LUC expression in the light phase of the cycle between CT 7–10 suggesting the lack of photoentrainment of peripheral tissues in vivo (Fig 6B). The average period for all tissues were comparable in control and mutant mice (p>0.05, unpaired two-tailed Student’s t-test), with the exception of adrenals (p = 0.03, unpaired two-tailed Student’s t-test) (Fig 5C).

In order to further assess the effects of light on the circadian rhythmicity of peripheral tissues, we also housed the animals in LD cycles (12:12 h) in the absence of running wheels. We reasoned that in control mice the peripheral tissues should maintain their circadian PER2::LUC rhythms in phase to light-dark cycles and in phase to each other. The Opn4::TeNT mice on the other hand were expected to show circadian oscillations independent of LD cycles with no phase relationships to each other. Analysis of peak PER2::LUC expression using phase coherence maps of the peak time on the second day show clustered expression of PER2::LUC rhythms in all tissues examined in the control mice (Fig 7) (Rayleigh uniformity test, P < 0.001). However in the tissues derived from Opn4::TeNT mice, there was considerable dispersion of the peak times for each tissue examined with a low degree of phase coherence (Rayleigh uniformity test, P > 0.05) with exception of adrenals (Rayleigh uniformity test, P = 0.02). These results suggest that light failed to serve as an entrainer for the peripheral tissues examined in the absence of ipRGC neurochemical signaling.

In nocturnal rodents, baseline plasma corticosterone concentrations vary predictably peaking immediately before the onset of activity [30]. Thus we asked the question whether the circadian pattern of corticosterone in mice is entrained by light in an ipRGC-dependent manner. Mice were housed in a 12:12 h LD cycle in cages containing running wheels for three weeks before the implantation of the microdialysis probe. The LD schedule was then maintained during the microdialysis sampling in cages without running wheels. Subcutaneous corticosterone, which is highly synchronized to plasma corticosterone rhythms in rats [31], was measured at 1-hour intervals for two-three days using in vivo microdialysis. The periodicity and phase of corticosterone rhythms in dialysate was assessed using CircWave software which aims to produce a Fourier-curve that describes the data by adding as many harmonics to the wave as the data allow(Fig 8). In both experimental groups, the subcutaneous corticosterone showed a circadian rhythm (CircWave, p<0.0001)) with peak times aligned to the expected onset of running wheel activity. However, whereas the corticosterone peak aligned with the beginning of the dark phase in the control group, the mutant animals showed a peak corticosterone that was not synchronized to LD cycles (Fig 8). Taken together, our data show that the lack of ipRGC signaling prevents synchronization of corticosterone rhythms to environmental LD cycles.