Daytime Light Intensity Modulates Spatial Learning and Hippocampal Plasticity in Female Nile Grass Rats (Arvicanthis niloticus)

Abstract

Light has pervasive effects on the physiology and behavior of mammals. Several human studies have shown that light modulates cognitive functions, however, the mechanisms responsible for the effects of light remain unclear. Our previous work using diurnal male Nile grass rats (Arvicanthis niloticus) revealed that reduced illuminance during the day leads to impairments in hippocampal-dependent spatial learning/memory, reduced CA1 dendritic spine density, and attenuated hippocampal brain-derived neurotrophic factor (BDNF) expression in males. The present study examined the impact of ambient light intensity on hippocampal functions in female grass rats and explored sex differences in behavioral and hippocampal responses. Female grass rats were housed in either a 12:12hr bright light-dark (brLD, 1000 lux) or dim light-dark (dimLD, 50 lux) cycle for four weeks. dimLD group showed impaired spatial memory in the Morris Water Maze task and reduced CA1 apical dendritic spine density, similar to prior observations in males. However, the behavioral deficits seen in females were more severe than those seen in males, with dimLD females showing no evidence of long-term retention over the 24-hour periods between training sessions. In contrast to the attenuated hippocampal BDNF expression found in dimLD males, there was no significant difference in the expression of BDNF and of its receptor TrkB between females in brLD and dimLD. The results suggest that, as seen in male grass rats, reduced illuminance during the day impairs hippocampal-dependent spatial memory and hippocampal plasticity in female diurnal grass rats, but the underlying signaling pathways responsible for the effects of light restriction may differ between the sexes.

Introduction

The day/night light-dark cycle is the most reliable and predictable cue in the environment influencing our brain and behavior. For diurnal mammals including humans, bright light during the day entrains circadian rhythms, promotes wakefulness and arousal, and influences emotion and cognition (Challet E, 2007;Smale L et al., 2003;Vandewalle G et al., 2009). Insufficient daylight exposure has a negative impact on emotion and cognition, as seen in patients suffering from seasonal affective disorder (SAD), in which the typical symptoms include depression, anxiety, low motivation and cognitive impairment (Rosenthal NE et al., 1984). The onset and remission of the symptoms in the fall and spring, respectively, is associated with seasonal fluctuation in the amount of sunlight that individuals receive. Before its full remission in the spring, the SAD symptoms can be alleviated by bright light therapy, further supporting the notion that the cause of SAD is light deficiency. In addition to the changes in ambient light, reduced photoreception can result from eye diseases, such as glaucoma or age-related macular degeneration, which have been associated with cognitive impairments (Clemons TE et al., 2006;Harrabi H et al., 2015). The impact of light on cognitive function has also been documented in non-clinical populations independently of seasonal changes in photoperiod, as brighter light has been found to improve standard test scores of school children and productivity at work (Baron R et al., 1992;Heschong L et al., 2002;Mills P et al., 2007;Mott M et al., 2012). These findings collectively suggest that the amount of daylight is positively related with superior cognitive function, but the underlying mechanism for this effect of light is unclear.

Our previous work utilized a diurnal rodent model, the Nile grass rat (Arvicanthis niloticus) to study the effects of ambient light on hippocampal function (Soler JE et al., 2018). We found that male grass rats housed in dim light during the day had impairments in the Morris water maze (MWM), a spatial learning/memory task, compared to those housed in bright light, and that the behavioral deficits were accompanied by reduced brain-derived neurotrophic factor (BDNF) and dendritic spine density in the hippocampus. The results suggest that ambient light modulates spatial learning and memory through structural and functional changes within the hippocampus. Since these findings pertain to only males, whether or not light affects female grass rats in a similar manner remains unknown. Although it might be intuitive to expect female grass rats to show the same behavioral and hippocampal responses following the same light paradigm, that is not necessarily our hypothesis, as males and females differ in various brain functions including learning and memory (Andreano JM and Cahill L, 2009;Cahill L, 2006;Koss WA and Frick KM, 2017). Particularly, sex differences in spatial learning and memory have been documented in numerous research efforts using humans and rodents (Hyde JS, 2016;Jonasson Z, 2005). In humans, consistent sex differences favoring males have been found in virtual MWM tasks, route learning and spatial rotation (Koss WA and Frick KM, 2017). In rodents, the results are less consistent, with sex differences found in some studies but not in others (Koss WA and Frick KM, 2017). For the MWM task, a sex difference was detected with a male advantage for rats, and a small female advantage in mice. Many factors can influence the sex differences in spatial memory tasks. In rats, the male advantage in MWM was greater when the animals were raised in isolation or without receiving pre-training trials (Jonasson Z, 2005), suggesting a possible interaction between sex and social stimulation and/or sex and stress, such that environmental, social, and experiential factors affect learning and memory in a sex-specific way. Therefore, it is possible that the ambient lighting conditions may also have different impacts on the cognitive performance of each sex.

The objective of the present study is to expand our previous findings on male grass rats by investigating the effects of light on hippocampal function in female grass rats, and to explore potential sex differences in behavioral and hippocampal structural and molecular responses following chronic daytime light deficiency. The majority, if not all, of the rodent studies on sex differences in learning and memory use nocturnal species, i.e. laboratory rats or mice. Diurnal and nocturnal species have adapted to different temporal niches through the entrainment of their circadian system by light as well as by showing opposite responses to acute presentations of light, to achieve optimal behavioral competence during the day or night, respectively (Smale L,Lee T and Nunez AA, 2003; Yan L et al., 2018). Our studies on spatial learning and memory in a diurnal animal model under different light intensities during the daytime, the active phase of the species, will fill a gap in the literature and provide insight into how ambient light modulates hippocampal function for both sexes in diurnal mammals, like humans.

Experimental Procedures

Subjects

Female unstriped Nile grass rats (Arvicanthis niloticus) used in all experiments were obtained from the breeding colony at Michigan State University. The animals were initially housed in a 12:12 hr light-dark (LD, ~300 lux during the day) cycle in plexiglass cages (34×28×17 cm) with food (PMI Nutrition Prolab RMH 2000, Brentwood, MO, USA) and water available ad libitum. During the experiment, animals were housed under either a 12:12hr bright light-dark (brLD, ~1000 lux during the day) or dim light-dark (dimLD, ~ 50 Lux) cycle as in our prior studies (Deats SP et al., 2014;Leach G et al., 2013;Soler JE,Robison AJ,Nunez AA and Yan L, 2018). Fluorescent light fixtures (Jesco Lighting, SP4–26SW/30-W) were utilized for the behavioral experiment, four cabinet lights (two in the front and two in the back) were attached to the top level of every row within a cage rack. The color temperature for these fixtures was approximately 3,000K. For enrichment, a PVC tube was provided in the cages. This also served as a hut for the grass rats to hide, thus direct light exposure during the experiment was voluntary.

Morris Water Maze

Female grass rats (n=8/lighting condition) were used in this experiment. Animals were singly housed in either brLD or dimLD for 4 weeks prior to being trained on the Morris Water Maze (MWM). During the 4th week of housing in each lighting condition, the animals were handled daily for 10 minutes in their home cage in the behavioral testing room. In the following week, animals were trained and tested for MWM. The handling, training and testing was performed the same way as described in a previous study using male animals (Soler JE,Robison AJ,Nunez AA and Yan L, 2018). Animals were trained and tested during zeitgeber time (ZT, ZT0 is lights on) 5–7; the light intensity in the testing room was ~ 300 lux. Training on the MWM was performed using a circular pool (60 cm depth × 122 cm diameter) with a platform (15-cm diameter) located 2cm under the water level and approximately 30cm away from the perimeter of the pool. The water was made opaque with non-toxic white paint and kept at 26±2°C, with different geometrical cues posted up on each wall in the room for spatial orientation. In order to be certain of normal motor abilities, the animals performed one-day cued-platform training in which the platform was visible before exposure to the hidden-platform training (Vorhees CV and Williams MT, 2006). All animals tested in the task located the platform in less than two minutes when it was visible. In the following 5 days training session, two training trials were completed each day with each trial at a maximum of two minutes in length with a 30-second inter-trial interval. Animals that failed to locate the platform in the 2-minute period were guided in the direction of the platform and given a latency of 120 seconds. Twenty-four hours after the final training session, reference memory was tested with the platform removed from the MWM and each grass rat swam for one minute. Training and testing sessions were recorded and analyzed using Noldus Ethovision (XT 8.5, Noldus Information Technology, Netherlands) by an experimenter who was blind to the experimental conditions.

Morphometry

Behaviorally naïve female grass rats (no experience with training or testing) were housed under either brLD or dimLD for 4 weeks prior to bilateral injection of herpes simplex virus expressing green fluorescent protein (HSV-GFP, Harvard Massachusetts General Hospital Viral Core Facility) into the dorsal hippocampus. The surgical procedure, histology and confocal microscopic analyses were performed as in the previous study using male grass rats (Soler JE,Robison AJ,Nunez AA and Yan L, 2018). In brief, 0.5μl of HSV-GFP was infused at a rate of 0.1μl/minute at the following coordinates from bregma: −0.1mm A-P, ± 2.0 mm L-M, and −2.7mm D-V from the surface of the brain. Animals were perfused transcardially 48h post-surgery. Brains sections at 100 μm thickness were mounted onto subbed glass slides with ProLong® Diamond Antifade Mountant (ThermoFisher Scientific, Waltham, MA), and were examined using a Nikon A1Rsi laser scanning confocal microscope at 1000x magnification. A z-stack was obtained for each sample to observe detailed morphology of dendritic spines, and was reconstructed to three dimensions using the NeuronStudio freeware morphometric program utilizing the rayburst algorithm (Rodriguez A et al., 2008). Five neurons (two dendritic segments/neuron) were analyzed per animal. Dendritic segments were randomly chosen 50–150 μm away from the soma for analysis; segments included were ~1.5μm in diameter for both groups, with even viral expression and no overlap with neighboring dendrites. Dendritic spines were classified on three parameters: (1) presence or absence of a neck; (2) head diameter; and (3) head/neck aspect ratio (Bourne JN and Harris KM, 2008;Hering H and Sheng M, 2001). Both thin and mushroom subtypes have visible necks, but where the mushroom subtype has a head which has a markedly larger diameter compared to the neck, the thin subtype does not have a notable difference between the head and neck diameter. The stubby subtype has a large head, but lacks the presence of the neck. Each subtype of dendritic spines was analyzed from 20μm segments of two distinct dendritic branches per neuron.

Immunohistochemistry (IHC)

Animals tested in the MWM were left undisturbed for two days, and then they were transcardially perfused at ZT 5–7 with saline followed by 4% paraformaldehyde. Brains were processed for BDNF IHC followed by quantitative analysis following the same procedures as described in our previous study using males (Soler JE,Robison AJ,Nunez AA and Yan L, 2018). Briefly, 3 alternate sets of 40μm sections were collected; one set of 10 sequential sections containing the dorsal hippocampus was incubated with anti-BDNF primary antibody (1:5000, ab101747, Abcam, Cambridge, UK). The signals were visualized using 3,3’-Diaminobenzidine (DAB) and 4% Nickel Sulfate. BDNF-immunoreactive (ir) cells were counted on photomicrographs of the dorsal hippocampus by an experimenter who was blind to the experimental conditions. BDNF-ir cells were only counted if they exhibited immunoreactivity in both the nucleus and cytoplasm of the cell, those that only showed partial immunoreactivity were not included in the counts. Cells were counted within the CA1, CA3 and Dentate Gyrus (DG) subregions of the hippocampus with a 200μm × 400μm counting box (Soler JE,Robison AJ,Nunez AA and Yan L, 2018).

Western Blot

Behaviorally naïve female grass rats were used for Western blot analysis. The animals were housed in either brLD or dimLD for 4 weeks prior to brain tissue collection at ZT 5–7 following decapitation. Flash frozen brains were sectioned coronally at 200 μm thickness, thaw-mounted onto a slide. The CA1 subregions were punched out from the slice using a 1-mm (diameter) micropuncher (Harris Micropunch, Hatfield, PA) and stored at −80 °C. For the analysis of BDNF (n=8/condition), tissue punches were then homogenized in radioimmunoprecipitation assay (RIPA) lysis buffer (sc-24948; Santa Cruz Biotechnology, Santa Cruz, CA) for protein extraction according manufacturer’s instruction. Protein concentrations were measured with the Bradford assay method (Bio-Rad; Hercules, CA). From each animal, 20 μg total protein was run on precast gels (4–20% Tris-Glycine Mini; NuSep, Germantown, MD) and transferred to nitrocellulose membranes (iBlot Gel Transfer Stacks; Invitrogen by Thermo Fisher Scientific). Membranes were treated with REVERT Total Protein Stain Kit (P/N 926–11016; LI-COR, Lincoln, NE) to quantify total protein for western blot normalization using LI-COR Odyssey CLx Imaging System. After total protein imaging, membranes were washed with REVERT reversal solution to remove total protein stain and the membranes were proceeded for BDNF immunoblotting. Membranes were incubated in Odyssey Blocking Buffer TBS (OBB-TBS) on shaker for 1 hour at room temperature, followed by incubation in guinea pig anti-BNDF primary antibody (1:1000; AGP-021; Alomone Labs, Jerusalem, Israel) at 4 °C for 5 days. The BDNF antibody that was utilized detected both mBDNF and proBDNF, as well as any potential dimers or tetramers of BDNF. The membranes were then incubated with a IRDye 800CW secondary antibody (1:10,000, P/N 925–32411, LI-COR, Lincoln, NE). The fluorescence intensity of proBDNF (~35KD) and mature BDNF (mBDNF, ~16KD) were detected and quantified by LI-COR Odyssey CLx Imaging System and normalized to the total protein fluorescence intensity measured from the same animal. The ratio of fluorescence intensity for proBDNF or mBDNF over total protein of each animal were calculated and used for statistical analysis. A second cohort of female grass rats (n= 5–6/condition) were used for analyzing tropomyosin receptor kinase B (TrkB) and its phosphorylation at the tyrosine 816 (Tyr816) site (pTrkB). CA1 tissue punches were collected as for BDNF assay above. To preserve phosphorylated sites, tissue punches were homogenized in RIPA buffer with phosphatase inhibitors (PhosSTOP™, Millipore Sigma). 100 μg total protein from each animal was loaded on the gel and was analyzed following the same procedures as for BDNF above. A rabbit anti-phospho-TrkB (Tyr816) (1:500, Millipore, Cat#ABN1381) was used to first detect the pTrkB (~140KD). Preadsorption was performed with control peptides to verify the specificity of the antibodies. The membrane was then stripped with Restore™ PLUS Western Blot stripping buffer (Thermo Scientific, Ref# 46430) before being re-incubated with a rabbit anti-TrkB (~140KD; 1:1000, Alomone, Cat#ANT-019) to detect total TrkB (~140KD). For both pTrkB and total-TrkB detection, an IRDye® 680RD goat anti-rabbit IgG (1:10000, 0.1mg, LI-COR, cat#925–68071) was used. The phosphorylation ratio of TrkB was determined using the ratio of fluorescent intensity between pTrkB (Tyr816) and total TrkB.

Data analysis

Statistical analysis was performed using SPSS (version 24, IBM, Armonk, North Castle, NY). For the MWM, the latency to reach the platform was analyzed within each trail using a 2 × 5 Mixed ANOVAs with lighting condition as the between–subjects factor and training days as the repeated measures factor. In the case that there was a significant interaction, repeated measure one-way ANOVA was used to analyze the effects of training days within each condition; when there was no interaction, only main effects were interpreted. Two-tailed independent samples student’s t-tests were used to assess group differences on the amount of time spent in the goal quadrant, swim speed, and thigmotaxis (i.e. time spent swimming within 10cm of edge) during the probe tests. Dendritic spine density, the number of BDNF-ir cells and the level of mBDNF, proBDNF and TrkB phosphorylation ratio were compared between lighting conditions using two-tailed independent samples student’s t-tests. The threshold for statistical significance for all analyses was established at p < 0.05.

Results

Chronic dim light housing impairs MWM performance of female grass rats

During the first trial over the 5 training days (Fig. 1A), female rats in the brLD group located the platform more effectively when compared to those in the dimLD group (Fig. 1A; main effect of training days: F(4,56)= 8.332, p < 0.001; main effect of lighting condition: F(1,14)= 7.657, p < 0.05). There was also a significant interaction between lighting condition and training days (F(4,56)= 7.830, p < 0.001). For brLD, the repeated measures ANOVA revealed a significant effect of training days (F(4, 28)=14.361, p <0.001); for dimLD, the training days had no significant effect (F(4, 28)=1.29, p > 0.05). During the second trial of each day (Fig. 1B), which was conducted 30 seconds after the first one each day, both groups showed improved performance over training days (F(4,56)= 7.914, p < 0.001) without significant differences between the two groups (main effect of lighting condition: F(1,14)= 1.181, p > 0.05) or interactions between training days and light condition (F(4,56)= 0.547, p > 0.05). During the probe trial when the platform was removed from the pool, animals in the brLD group mainly focused their search within the target goal quadrant while those in the dimLD group displayed a random search pattern (Fig. 1C). Quantitative analysis revealed that animals in the dimLD group spent significantly less time searching within the goal quadrant when compared to the brLD group (Fig. 1D; t(14)= 3.134, p < 0.01), and the performance of dimLD animals was not different from the chance level of 15 seconds (one sample t-test, t(7)= 2.03, p > 0.05). Independent samples t-tests revealed that there were no significant differences between lighting conditions in either swim speed (t(14)= 0.936, p> 0.05) or thigmotaxis (t(14)= −0.161, p > 0.05).

Figure 1.

MWM performance of female grass rats housed in either brLD or dimLD conditions over 4 weeks. (A) Latency of the animals to find the platform in trial 1 (24-hour delay) over the course of 5 training days. Animals housed in brLD located the platform significantly faster than the animals housed in dimLD (main effect of training days: F(4,56)= 8.332, p < 0.001; main effect of lighting condition: F(1,14)= 7.657, p<0.05); interaction between lighting condition and training days (F(4,56)= 7.830, p<0.001). (B) Latency of the animals to find the platform in trial 2 (30-second delay) over the course of 5 training days. Both groups expressed improvement over the training days (F(4,28)=14.361, p<0.001) with no significant differences between the groups or interactions between training days and light condition. (C) Representative track plots of the search patterns used by brLD and dimLD animals during the probe trial (goal quadrant in red). (D) Grass rats in the dimLD condition spent significantly less time in the goal quadrant compared to grass rats in the brLD condition; horizontal line at 15 seconds indicates chance level performance. Date are shown as mean ± sem. *, p < 0.05

Chronic dim light housing leads to attenuated CA1 dendritic spine density in female grass rats

CA1 apical dendritic spine density was analyzed by each subtype (i.e. mushroom, thin and stubby) using 3D dendritic spine reconstruction from HSV-GFP transduced CA1 neurons (Fig. 2A.) Quantitative analysis (Fig. 2B) revealed a greater density of mushroom spines (t(9)= 5.357, p < 0.001) in the brLD compared to the dimLD group. There were no significant differences between the two groups in thin or stubby dendritic spines (p> 0.05). The total spin density of three subtypes combined is significantly higher in the brLD group (t(9)= 2.879, p = 0.018).

Figure 2.

Hippocampal CA1 apical dendrites visualized and quantified by expression of HSV-GFP. (A) Visualization of dendritic spines using HSV-GFP expression. (B) Quantification of the dendritic density for each spine subtype. Scale bar, 5μm. Date are shown as mean ± sem. *, p < 0.001

Photic modulation of hippocampal BDNF expression and TrkB phosphorylation in female grass rats

BDNF-ir in the hippocampus was comparable between female grass rats housed in brLD or dimLD condition (Fig. 3A). Quantitative analysis revealed no significant difference in the number of BDNF-ir cells in CA1, CA3 or DG (Fig. 3B, p> 0.05). The results were verified by Western blot in the CA1, there was no significant difference in the level of either mature BDNF or proBDNF between the two lighting conditions ((Fig. 4A, B, ps > 0.05). Additionally, the phosphorylation of TrkB (Tyr816) in the CA1 was also not significantly different between females in either lighting condition (Fig. 4C, D, p > 0.05).

Figure 3.

Hippocampal BDNF-ir cells in female grass rats housed in brLD or dimLD condition. (A) Representative photomicrographs of BDNF immunochemical staining in the CA1, CA3, and DG of the hippocampus in brLD and dimLD animals. (B) Quantitative analysis of the number of BDNF-ir cells in brLD and dimLD conditions in each hippocampal subregions. Date are shown as mean ± sem. Scale bar, 100 μm. The rectangle shows the size of counted area (200 μm × 400 μm). No significant differences were observed between the counts in brLD and dimLD condition in any of the subregion (p > 0.05).

Figure 4.

Hippocampal CA1 expression of BDNF and TrkB in female grass rats. A, representative Western blot bands of mature BDNF (mBDNF) and proBDNF. B, relative expression level of mBDNF and proBDNF. C, representative Western blot bands of phosphor-TrkB at Tyr816 (pTrkB) and total TrkB. D, phosphorylation rate of TrkB (Try816). Date are shown as mean ± sem. No significant differences were observed between brLD and dimLD in any of the measures (p > 0.05).

Discussion

Using female Nile grass rats, the present study revealed that chronic dim lighting conditions (dimLD) resulted in impaired performance in the MWM, a hippocampus-dependent spatial task, along with reduced dendritic spine density in the CA1 subregion. Another marked difference was that in contrast to the downregulation of BDNF in dimLD compared to brLD in males, the hippocampal BDNF expression remained unchanged in females exposed to the dimLD condition. Together, these results suggest that daytime light deficiency negatively impacts hippocampal function in both males and females, but through distinct neural mechanisms, with downregulation of BDNF being involved in males but not in females. Alternatively, a yet unknown mechanism common to both sexes and unrelated to changes in BDNF may be responsible for the effects of dim light on behavior and hippocampal morphology.

Female ovarian hormones have been shown to influence hippocampal functions, such that the strength of spatial learning/memory, CA1 dendritic spine density and BDNF expression all fluctuate with the estrous cycle (Chan CB and Ye K, 2017;Koss WA and Frick KM, 2017). Female grass rats do not have spontaneous estrous cycles, and remain in diestrus when they are not co-housed with males (McElhinny TL et al., 1997). Thus, there was no need to control for ovulatory cycle phases when they were trained and tested in the MWM task (Fig. 1). Following a 30-second delay during the training session (trial 2, Fig. 1B), the dimLD group performed similarly to the brLD group. This suggests that after completing the first training trial, animals in both groups encoded newly acquired information regarding the location of the platform to perform the second trial effectively. However, the impairments of the dimLD group became evident when assessed after 24h delays. Analysis of the performance of the dimLD females across the first trials of the training phase, showed no evidence of learning (trial 1, Fig. 1A). The absence of a learning curve in the dimLD group during trial 1seems to indicate that there was a lack of consolidation of newly acquired information because there is no progression in latency scores as training days go on. When reference memory was assessed, the dimLD females also performed at chance level during the probe trial (Fig. 1D). If appropriate acquisition and consolidation of the MWM task occurred, animals would spend most of their time in the probe trial searching for the platform in the goal quadrant. By impairing the consolidation process during training, animals in the dimLD group exhibit poor reference memory by searching indiscriminately across all quadrants in the MWM.

The behavioral data seem to suggest that short-term memory (STM) is conserved and long-term memory (LTM) is impaired. A possible explanation for this occurrence may be that early phase of long-term potentiation (E-LTP) remains intact, but chronic daytime light deficiency may disrupt the induction and/or maintenance of late-phase LTP (L-LTP). STM is governed by E-LTP, which only depends on the activation of existing proteins and intracellular release of Ca²⁺ to traffic receptors and kinases to the synapse (Sweatt JD, 1999). On the other hand, LTM relies upon altered gene expression and protein synthesis required for the persistence of L-LTP (Blum S et al., 1999;Goelet P et al., 1986;Wang SH et al., 2010). Therefore, it may be possible that chronic dim lighting conditions do not affect aspects of LTP that only consist of activating existing synaptic machinery, but impair those that are required for strengthening and stabilizing the synapse.

When the MWM performance of females in the present study was compared with male counterparts in our previous study (Soler JE,Robison AJ,Nunez AA and Yan L, 2018), although the results are generally consistent, females in dimLD seemed to exhibit greater behavioral deficits in the MWM task (Fig. 5). When the MWM performance during the first trial of each training day was compared, animals were housed, females showed comparable performance as males in brLD condition (Fig. 5A), but much worse performance in females compared to males in the dimLD condition (Fig. 5B). In dimLD conditions, in contrast to the modest but significant improvement over training days seen in males, there was no improvement over training days in females. The difference between females and males in dimLD points to a higher vulnerability in females to the detrimental effects of chronic daytime light deficiency on a hippocampal-dependent memory task. This higher level of susceptibility in females may also apply to other challenges that affect hippocampal functioning and may account for the female-bias seen in some neurological or psychiatric disorders e.g., Alzheimer’s disease and depression (Launer LJ et al., 1999;Nolen-Hoeksema S, 2001).

Figure 5.

Sex difference in MWM performance following reduced daytime illumination. MWM performance in trial 1 (24-hour delay) over the 5 training days was compared between female and male grass rats. The female graphs are replotted with data shown in Figure 1, male graphs are replotted using data published in a previous study (Soler JE,Robison AJ,Nunez AA and Yan L, 2018) with permission. (A) In brLD condition, females and males showed comparable performance in MWM task (main effect of sex: F(1,14)=1.01, p > 0.05; interaction between sex and training days: F(4,56)=1.54, p > 0.05). (B) In dimLD condition, males outperformed females in the MWM. There was a significant interaction between sex and training days (F(4,56)=3.739, p < 0.01). Post-hoc comparison revealed a significant difference on training day 4 and 5 (*, p < 0.05). A significant main effect of training days was present in males (F(4,28)=1.29, p < 0.01), but absent in females (F(4,28)=1.29, p > 0.05).

Previous work using male grass rats revealed attenuated CA1 dendritic spine density, especially the mushroom and stubby type of spines following dimLD housing (Soler JE,Robison AJ,Nunez AA and Yan L, 2018). Consistent with those observations, female grass rats exposed to dim illumination showed fewer mushroom spines compared to their brLD counterparts (Fig. 2). Spine morphology has been linked to the function and stability of a synapse. Mushroom and stubby spines have larger heads that are positively correlated with the size of the post-synaptic density (PSD) area and the number of NMDA and AMPA receptors and docked presynaptic vesicles (Dobrunz LE and Stevens CF, 1997;Harris KM and Stevens JK, 1989;Nusser Z et al., 1997;Schikorski T and Stevens CF, 2001). Consequently, Ca²⁺ influx from NMDA receptors leads to the activation of calcium/calmodulin-dependent protein kinase II (CaMKII). With repeated high-frequency Ca²⁺ influx, CaMKII undergoes autophosphorylation that confers constitutive Ca2⁺-independent kinase activity, and allows enhancement of AMPA receptors that are already at the synapse and the trafficking of additional AMPA receptors to the synapse, critical mechanisms for LTP (Shonesy BC et al., 2014). Additionally, enhanced CaMKIIα activity is restricted to dendritic spines that undergo enlargement (Lee SJ et al., 2009). The spine morphology of grass rats was analyzed in behaviorally naïve animals that were not trained in the MWM. Thus, the lower density of mushroom and stubby spines in the dimLD group indicates that there are less mature synapses available to engage memory formation even before the MWM training. Behavioral experience, i.e. training in the MWM, can induce neural plasticity, and leads to the development and/or maturation of spines corresponding to the strengthening of neural connections that are needed for memory formation (De Roo M et al., 2008). Although not directly tested, it is possible that compared to the difference seen in naïve animals (Fig. 2), the differences in spine morphology between brLD and dimLD groups would be more salient following MWM training and testing, and could potentially reveal sex differences in spine morphology not seen in naïve animals here.

In contrast to the impaired spatial memory and attenuated hippocampal spine density, daytime dim light housing had no significant effect on hippocampal BDNF expression in female grass rats. There was no significant difference in the number of hippocampal BDNF-ir cells (Fig. 3) or BDNF protein in CA1 (Fig. 4) between females in dimLD and brLD conditions. On the other hand, a significant reduction of BDNF-ir cells (Soler JE,Robison AJ,Nunez AA and Yan L, 2018) and BDNF in CA1 (data not shown) following dimLD housing was observed in male grass rats. To further analyze BDNF signaling, we also examined its high-affinity binding receptor, TrkB. Ligand-mediated phosphorylation of TrkB at Tyr816 leads to phospholipase-Cγ (PLCγ) mobilization of intracellular Ca²⁺ stores, a vital step for LTP maintenance (Emfors & Bramham, 2003). However, no significant differences were observed in Tyr816 phosphorylation between the two lighting conditions. The results in BDNF and TrkB expression suggest that the modulation of daytime light intensity on hippocampal function may involve different signaling pathways in males and females. It should be noted that the analyses of BDNF and TrkB were conducted following 4 weeks of animals housed in each lighting condition, and thus the results do not obviate the possibility of changes in BDNF-TrkB pathways at earlier time points. Such earlier changes could contribute to the reduction in mushroom spines and the behavioral deficits in MWM, and by the time the structural and behavioral changes became evident, the BDNF-TrkB pathways in dimLD group could have reached the “new normal” steady state of the system, no longer showing significant differences from the brLD group. On the other hand, sex-specific responses in hippocampal BDNF have been observed in other rodents following various paradigms/treatments. For example, isolation or maternal deprivation reduces BDNF expression solely in male but not female mice (Kikusui T et al., 2009;Viveros MP et al., 2010;Weintraub A et al., 2010); while enrichment induces more BDNF expression in females than in males (Bakos J et al., 2009;Zhu SW et al., 2006). The findings from diurnal grass rats contribute to the existing literature on sex-specific regulation of hippocampal BDNF, and suggest the possibility of distinct neural mechanisms underlying the modulatory effects of ambient light on hippocampal function in males and females.

The cellular mechanisms or molecules responsible for deficits in hippocampal synaptic plasticity and spatial memory in female grass rats housed in dimLD in the absence of changes in BDNF remain unknown. In addition to BDNF, transcription factors of the nuclear factor-κB (NF-κB) family have been implicated in synaptic plasticity (Ahn HJ et al., 2008;Freudenthal R et al., 2005;Freudenthal R et al., 2004). Many members of the NF-κB family activate the common intracellular kinase pathways that are also the target of BDNF, to execute synaptic strengthening in a similar fashion, and thus could be an alternative molecule regulated in female grass rats in dimLD. Furthermore, astroglial NF-κB has been shown to play a sex-specific role in learning and memory, such that overexpression of NF-κB inhibitor in transgenic mice leads to impairments in hippocampal-dependent tasks in females without affecting males (Bracchi-Ricard V et al., 2008). Future studies will test the possible involvement of the transcription factors of the NF-κB family in photic modulation of hippocampal functions in grass rats.

The present study expanded our previous findings on the effects of light on the hippocampal functions in male diurnal grass rats, to show that the daytime dim light condition leads to impaired hippocampal function in females as well, but with more salient behavioral deficits in females compared to males. Furthermore, the results suggest that ambient lighting conditions activate sex-specific neural responses within the hippocampus to modulate spatial learning and memory as well as hippocampal dendritic spine morphology. For humans, the amount of light we are exposed to varies over the seasons and over the life span, with older people being particularly at risk for insufficient light exposure. Older adults living in residential care homes experience a mean daytime light exposure that is less than 500 lux (Nioi A et al., 2017;Shochat T et al., 2000). In a group of individuals 60–100 years old, the median duration of light above 1,000 lux that they experienced was only 9 min a day, and within that group, for those with advanced cognitive decline, this duration dropped to 1 min (Shochat T,Martin J,Marler M and Ancoli-Israel S, 2000). Given our findings in the grass rat model, it is reasonable to expect that the low level of illuminance and the extremely short duration of bright light will impair the already fragile hippocampus of the aging brain and accelerate aging-related cognitive decline. Furthermore, the prevalence of dementia and Alzheimer’s Disease is higher in women than in men (Launer LJ,Andersen K,Dewey ME,Letenneur L,Ott A,Amaducci LA,Brayne C,Copeland JR,Dartigues JF,Kragh-Sorensen P,Lobo A,Martinez-Lage JM,Stijnen T and Hofman A, 1999;Munro CA, 2014), suggesting a possible sex differences for humans with respect to vulnerability to environmental challenges, which may include light deficiency. A better understanding of the neural mechanisms underlying impaired learning and memory in both males and females will contribute to gender-specific strategies for the prevention and treatment of cognitive impairments.

Highlights.

• Dim light housing during the day leads to impaired spatial learning and memory in female diurnal grass rats.

• Female grass rats housed in dim light had reduced CA1 dendritic spine density.

• Dim light housing had no significant effect on hippocampal BDNF expression and TrkB phosphorylation in female grass rats.

Abbreviations

ANOVA

Analysis of variance

BDNF

Brain-derived neurotrophic factor

brLD

bright light-dark

CA1

Cornu Ammonis 1

CA3

Cornu Ammonis 3

CaMKII

calcium/calmodulin-dependent protein kinase II

DAB

3,3’-Diaminobenzidine

DG

Dentate gyrus

dimLD

dim light-dark

E-LTP

Early long-term potentiation

GFP

Green fluorescent protein

HSV

Herpes-simplex virus

IHC

Immunohistochemistry

LD

Light-dark

LTP

Long-term potentiation

L-LTP

Late long-term potentiation

LTM

Long-term memory

MWM

Morris Water Maze

NF-κB

Nuclear factor-κB

PLCγ

Phospholipase-Cγ

SAD

Seasonal Affective Disorder

STM

Short-term memory

TrkB

Tropomyosin receptor kinase B

Tyr

Tyrosine

ZT

Zeitgeber time