SAD rats: Effects of short photoperiod and carbohydrate consumption on sleep, liver steatosis, and the gut microbiome in diurnal grass rats

Abstract

Seasonal affective disorder (SAD) is a recurrent depression triggered by exposure to short photoperiods, with a subset of patients reporting hypersomnia, increased appetite, and carbohydrate craving. Dysfunction of the microbiota–gut–brain axis is frequently associated with depressive disorders, but its role in SAD is unknown. Nile grass rats (Arvicanthis niloticus) are potentially useful for exploring the pathophysiology of SAD, as they are diurnal and have been found to exhibit anhedonia and affective-like behavior in response to short photoperiods. Further, given grass rats have been found to spontaneously develop metabolic syndrome, they may be particularly susceptible to environmental triggers of metabolic dysbiosis. We conducted a 2×2 factorial design experiment to test the effects of short photoperiod (4h:20h Light:Dark (LD) vs. neutral 12:12 LD), access to a high concentration (8%) sucrose solution, and the interaction between the two, on activity, sleep, liver steatosis, and the gut microbiome of grass rats. We found that animals on short photoperiods maintained robust diel rhythms and similar subjective day lengths as controls in neutral photoperiods but showed disrupted activity and sleep patterns (i.e., a return to sleep after an initial bout of activity that occurs ~13h before lights off). We found no evidence that photoperiod influenced sucrose consumption. By the end of the experiment, some grass rats were overweight and exhibited signs of non-alcoholic fatty liver disease (NAFLD) with micro- and macro-steatosis. However, neither photoperiod nor access to sucrose solution significantly affected the degree of liver steatosis. The gut microbiome of grass rats varied substantially among individuals, but most variation was attributable to parental effects and the microbiome was unaffected by photoperiod or access to sucrose. Our study indicates short photoperiod leads to disrupted activity and sleep in grass rats but does not impact sucrose consumption or exacerbate metabolic dysbiosis and NAFLD.

Introduction

Seasonal Affective Disorder, or SAD, is a syndrome with uncertain etiology characterized by recurrent depressive episodes during autumn and winter which subsequently abate each spring or summer (Sohn and Lam, 2005). In the United States, it is estimated that 2–5% of individuals may meet the diagnostic criteria for SAD, with prevalence rates being 3–4 times higher in women (Rosen et al., 1990; Magnusson and Partonen, 2005) and increasing with latitude (Rosen et al., 1990; Booker and Hellekson, 1992). Importantly, the incidence of SAD may be increasing owing to behavioral changes including increased exposure to indoor lighting, reduced exposure to natural outdoor lighting, decreases in the total amount of daily sleep, and increasing percentages of individuals that engage in shift work (Woo and Postolache, 2008; Fonken and Nelson, 2011). Although SAD is widely recognized as a type of seasonal depression, individuals with SAD frequently exhibit circadian dysfunction and sleep disruption, as well as carbohydrate craving and weight gain (Donofry et al., 2014; Sandman et al., 2016). However, not all individuals that suffer from SAD exhibit these symptoms and instead this appears to be a significant subtype that is particularly prevalent among women (Rohan et al., 2009).

Though the etiology of SAD remains uncertain, a variety of potential underlying biological mechanisms have been proposed (Sohn and Lam, 2005; Rohan et al., 2009), with disruption of brain monoamine transmission being the most frequently implicated (Neumeister et al., 2001; Praschak-Rieder and Willeit, 2012). Dysfunction in serotonin signaling circuits has been reported in some patients with SAD (Mc Mahon et al., 2016). This is consistent with other major depressive disorders, as selective serotonin reuptake inhibitors (SSRIs) are the most widely prescribed antidepressants. How SSRIs reduce depression is highly controversial, but recent evidence suggests serotonin receptor activation may be necessary for adult neurogenesis (Olivas-Cano et al., 2023), which is increasingly thought to influence mood (Becker and Wojtowicz, 2007). The relationship between carbohydrate craving and mood in SAD may also stem from dysfunction in serotonin signaling systems as the consumption of carbohydrates leads to increased brain tryptophan levels, and, consequently, increased serotonin production (Mantantzis et al., 2019). Consequently, it is hypothesized the drive to consume carbohydrate-laden meals may be linked to the need to restore depressed serotonin expression in SAD patients (Wurtman and Wurtman, 2018).

Sleep disruption is particularly prevalent in individuals that suffer from the recurrent seasonal depression that characterizes SAD, suggesting that SAD may be linked to disruption of circadian clock function (Lewy et al., 2006). Circadian and serotonin systems are extensively intertwined with reciprocal network connections such that disruption of circadian function may alter serotonin signaling pathways, and vice versa (reviewed in (Ciarleglio et al., 2011). It has been proposed that circadian delay may underlie SAD with affected individuals having phase-shifted circadian clocks relative to geophysical time (Sohn and Lam, 2005; Rohan et al., 2009). In support of this, polymorphisms in circadian-clock related genes are often associated with the prevalence of SAD and with timing of sleep and activity (Johansson et al., 2003; Roecklein et al., 2009, 2013).

Over the past decade, increased attention has been placed on the potential role of gut microbial remodeling in the pathophysiology of major depressive disorders (Foster and Neufeld, 2013; Evrensel and Ceylan, 2015). For example, gut microbiome remodeling has been implicated in the induction of depressive-like behaviors in some rodent studies – fecal microbiota transplantation of germ-free mice with microbiota derived from major depressive disorder patients resulted in more affective behaviors when compared to mice colonized with microbiota from healthy control individuals (Zheng et al., 2016). Remodeling of the gut microbiome may be an environmental factor contributing to the onset of major depressive disorders, though the role of the gut microbial community in contributing to SAD remains unexplored. Recent studies in hamsters and voles, however, reveal the gut microbiome can be sensitive to short photoperiods (Ren et al., 2020; Zhu et al., 2022). This remodeling may represent adaptive plasticity in species that occupy highly seasonal environments; whether the gut microbiome is sensitive to photoperiod in species that evolved in equatorial regions is unknown.

The development of animal models is a critical step for exploring the underlying pathophysiological mechanisms of SAD and for evaluating potential therapeutic approaches. A cornerstone of biomedical research using animal models is that the results from research on the model system reflect the human disease of interest. This may not be the case for nocturnal mouse and rat models of SAD, given that SAD is linked to circadian disruption and the effects of day-length on behavioral rhythms differ between diurnal and nocturnal species (Vetter, 2020). This has led to a call for the development of diurnal rodent models to investigate the molecular and neurological mechanisms that link circadian/sleep dysfunction, carbohydrate craving/weight gain, and seasonal depression in SAD (Yan et al., 2019; Shankar and Williams, 2021). The Nile grass rat, Arvichantis niloticus, has emerged as a potentially useful model of SAD, since exposure to short photoperiods induces affective-like behavioral changes that are associated with dysfunction in serotonin signaling systems (Leach et al., 2013a). Grass rats occur in equatorial regions, have diurnal patterns of activity, and do not show seasonal changes in their reproductive functions (Blanchong et al., 1999). In the present study, we investigate whether short photoperiods also disrupt diurnal rhythms, reduce sleep duration, and alter the gut microbiome and energy homeostasis in grass rats. Given prior work indicating that grass rats are susceptible to the spontaneous development of metabolic syndrome (Noda et al., 2010), we predicted short photoperiods would cause remodeling of the gut microbiome and augment metabolic dysbiosis.

Methods

Study species

We studied the effects of photoperiod and access to sucrose solution on sleep, liver steatosis, and the gut microbiome in 45 diurnal Nile grass rats split into two trials (24 in trial 1, and 21 in trial 2). These animals came from a founder colony of five males and four females from a breeder population at Michigan State University. All animals were housed under neutral photoperiods (12h Light: 12h Dark [12:12 LD]) throughout their lives prior to the initiation of the experiment. Fluorescent lighting was used (GE 73095; color 4100K) and light intensity was measured at ~1100 lux at 1m from floor. Across both trials, animals were an average of 136.4 ± 29.8 (SD) days old at the start of the experiment – animals were slightly younger in the first trial (111.7 ± 13.6 days old), compared to the second trial (164.7 ± 12.5 days old). All procedures used in the study were approved by the Institutional Animal Care and Use Committee (IACUC 1282484) of the University of Alaska Fairbanks.

Experimental design

We used a 2×2 experimental design, with animals subjected to either a short (4h Light: 20h Dark [4:20 LD]) or neutral (12:12 LD) photoperiod and provided with either access to 8% sucrose or only water. Room temperatures were always maintained at 22°C. Grass rats were given ad libitum food (Prolab RMH 2000 5P06 rodent chow from PMINutrition International, St. Louis, Missouri, USA) and ad libitum water access from a water bottle. Every three days, slip cages and bottles were changed, and animals were weighed. Prior to the experiment, animals were co-housed with littermates on a neutral photoperiod (lights on: 05:00–17:00). Animals were then moved into individual cages and provided two weeks to acclimate to their experimental rooms prior to altering photoperiod. After acclimation, we kept 23 rats (11 males, 12 females) on this neutral photoperiod, and shifted 22 rats (10 males, 12 females) to the short photoperiod treatment (lights on: 09:00–13:00). Four weeks later, half of the animals were given a second bottle with sucrose solution whereas the other half were given a second bottle of water (details in Sucrose treatment). At the end of the experiment, animals were euthanized (using CO₂), weighed, fat pads were extracted and weighed, liver samples were taken, and fecal pellets were extracted from the large intestine. Liver samples were fixed in 10% buffered formalin, trimmed, and then stained with a hematoxylin & eosin (H&E) stain. Fecal samples from the large intestine were immediately frozen on dry ice and then stored at −80°C for further analyses. Fecal samples were also collected directly from animal tubs (Trial 2 only) for microbiome analyses immediately prior to photoperiod treatment (0 weeks), after 4 weeks of photoperiod treatment, and at the end of the experiment after 7 weeks of photoperiod treatment and 3 weeks of sucrose treatment (Fig 1).

Fig 1.

Schematic of experimental design. We acclimatized grass rats to their experimental rooms for two weeks, under a 12:12 LD (Light:Dark) neutral photoperiod. Grass rats were fed ad libitum and initially had access to a single bottle of water. Half of the grass rats were maintained on this neutral photoperiod and the other half were shifted to a short photoperiod treatment (4:20 L:D) for the remainder of the experiment. Four weeks after shifting half of animals to a short photoperiod, animals were given a second water bottle that either contained sucrose solution (initially 2% but gradually increasing to 8% sucrose) or water (control). Fecal samples were collected after the 2-week acclimation (wk 0), after the first 4 weeks of photoperiod treatment (wk 4), and after 3 weeks of access to sucrose (wk 9). A fecal sample was also taken from the large intestine 9 weeks after the shift to short photoperiods after animals were euthanized for tissue sample collection.

Sucrose preference test

After four weeks of photoperiod treatment, all animals were given a sucrose preference test (SPT; (Serchov et al., 2016)) to measure anhedonic behavior which involved providing access to a bottle filled with 2% sucrose solution, in addition to their regular water bottle, for six days. We allowed animals to acclimate to the second bottle for 4 days and then measured their consumption of sucrose solution and water on days 5 and 6.

Sucrose treatment

Following the SPT, we provided half the animals in each photoperiod treatment with only access to water (i.e., two water bottles) whereas the other half were gradually shifted from the 2% sucrose solution to 8% sucrose solution, ramping up in 2% increments for two days per incremental increase (Fig. 1). In preliminary trials, we found that if we did not gradually ramp up the sucrose concentrations, and instead switched animals from water directly to 8% sucrose, they often avoided the high concentration sucrose solution. Animals remained on the 8% sucrose solution (or control water) for 3 weeks prior to terminal sampling (Fig 1).

Recording sleep and activity

Throughout the experiment, we characterized patterns of sleep vs. activity using a non-invasive Piezo Sleep Sensor System (Donohue et al., 2008; Signal Solutions LLC, Mang et al., 2014; Yaghouby et al., 2016). This system involves placing a thin dielectric or ‘piezoelectric’ sheet under the animal’s slip cage. In response to changes in surface pressure as the animal moves, this piezoelectric sheet generates a voltage signal. This signal was acquired using SleepStatLab software, which uses an algorithm to discriminate sleep from wake every 2 seconds based on periodic changes in pressure, generated by regular breathing observed during sleep, that are recorded from the piezoelectric floor. The PiezoSLEEP system has been independently validated in mice and rats using electroencephalograms (EEG) and electromyograms (EMG) (Mang et al., 2014; Topchiy et al., 2022). To date, no algorithm has been developed specifically for grass rats – therefore we applied the sleep algorithm developed for rats in the present study.

Liver steatosis

To assess whether any of the treatments were associated with fatty livers, we conducted liver histology with the H&E stained samples to measure macrovesicular steatosis. Imaging was done on a TissueFAXS 200 confocal slide scanner (Tissuegnostics, Germany) and detection was done by a Zyla 5.5 sCMOS camera (Andor). We used ImageJ (Schneider et al., 2012) to quantify the percent area of the liver that showed macrovesicular steatosis. Because we found significant micro and macrosteatosis in some animals (see results), we also stained liver samples for collagen accumulation (i.e., evidence of fibrosis) using Masson’s trichrome.

Fecal sampling and Microbiome analysis

For the second trial, fecal samples were collected at week 0 (prior to photoperiod treatment but after 2 weeks of acclimation to solo housing), at week 4 (after 4 weeks of photoperiod treatment but prior to SPT and sucrose treatment), and at week 8.5 (after 8.5 weeks of photoperiod treatment and 3 weeks of 8% sucrose treatment; Fig 1). At the end of the experiment, we also removed fecal samples from the large intestine following euthanasia. All fecal samples were frozen at −80°C until analysis. After homogenizing each fecal sample, we extracted DNA from 0.25 g of sample using the PowerFecal Pro kit (Qiagen, USA) according to the manufacturer’s instructions. The concentration and purity of the DNA was assessed using the NanoDrop 2000 (Thermo Scientific, USA). We amplified the V4 region of the 16S rRNA gene using dual-indexed 515F (Parada et al., 2016) and 806R (Apprill et al., 2015) PCR primers following the Earth Microbiome Project PCR protocol (https://earthmicrobiome.org/protocols-and-standards/16s/). PCR was performed in duplicate for each sample and pooled. We verified the PCR amplification products on a 1% agarose gel. We sent amplicons to the Institute of Arctic Biology Genomics Core Lab, where they were sequenced using v3 reagents on a MiSeq (Illumina, USA).

Data analyses

All data analysis was done in R v4.2.2, comparing both trial phases and their metrics as dependent variables, with significance set at 0.05. For data that contained continuous repeated measures, data was run using mixed effects models, with the R package, lme4 (Bates et al., 2015), including Individual as a random effect. For all models, we included photoperiod and sucrose access as fixed and interactive effects. Model residuals were examined for both normality and heteroscedasticity, with post-hoc tests run using the package, lmerTest (Kuznetsova et al., 2017), and the Kenward-Roger method for denominator degrees of freedom. When residuals appeared non-normal, the package bestNormalize (Peterson, 2021) was used to transform and rerun models; in cases where no transformation was appropriate, a generalized linear model with an ordered beta distribution was used, or a nonparametric Kruskal Wallis rank sum test was run, followed by a Dunn’s post-hoc test for comparisons.

Microbiome Analyses -

After obtaining Illumina microbiome data, we demultiplexed the samples using Mr_Demuxy v.1.2.0 (https://pypi.python.org/pypi/Mr_Demuxy/1.2.0), and then imported the paired-end reads into QIIME 2 (Bolyen et al., 2019) for further processing and analysis. We classified sample taxonomies using the SILVA v. 132 database (Quast et al., 2013) in QIIME 2. To avoid biases introduced by different sequencing depths, we rarefied the feature table at 9000 reads per sample. We calculated diversities using Shannon’s Index in QIIME 2. We used the R package ‘phyloseq’ (McMurdie and Holmes, 2013) to do a principle coordinates analysis (PCoA) using unweighted unifrac distances, which measure the phylogenetic distances between taxa using presence/absence data.

Results

Effects of Photoperiod on Sleep and Activity

Grass rats maintained on a 12:12 LD photoperiod were diurnal, being far more active during the photophase (light phase) and primarily inactive/sleeping during the scotophase (dark phase). However, the PiezoSleep system revealed anticipatory activity occurring immediately prior to lights on. In contrast, animals that were moved to a 4:20 LD photoperiod phase advanced their circadian clocks such that the daily onset of inactivity (and sleep) occurred coincident with lights off (Fig 2). Despite being on a 4:20 LD photoperiod, grass rats exhibited anticipatory activity onset approximately 13h prior to lights off, indicating that they maintained robust circadian rhythmicity but were unable to contract their active phase to match the new shorter photophase (Fig 2). In other words, grass rats given 4h of light maintained a 12h subjective day. However, shortly after their bout of anticipatory activity, grass rats on short photoperiods spent more time sleeping compared to individuals on neutral photoperiods (Fig 3). Animals on short photoperiods spent a longer percentage of time asleep during the subjective day (45.3 ± 15.6; mean ± SD) compared to animals on neutral photoperiods (29.7 ± 14.2; F_(1,44) = 27.19, p < .001), and grass rats on short photoperiods exhibited longer percentage of sleep bouts during the subjective day (246.3 ± 96.4) compared to animals on neutral photoperiods (181.7 ± 85.4; F_(1,43) = 12.84, p < .001), as estimated by the PiezoSleep System.

Fig 2.

Actograms showing an index of activity, assessed using the PiezoSleep system (3 most active, 0 at rest) for an animal on a neutral 12:12 LD photoperiod (Top) and an animal on a short (4:20 LD) photoperiod (Bottom) – the final 5 weeks of the experiment are shown. Each line represents 2 consecutive days of data with the x-axis showing time of day. Yellow bars at top indicate when lights were on whereas black bars indicate lights off. In both groups, animals exhibit anticipatory activity ~12–12.5h before lights off. For animals on neutral photoperiod, activity commences shortly before lights on, whereas for animals on short photoperiod, anticipatory activity occurs more than 8h before lights on. Activity data from the animal on short photoperiod (bottom) is delayed by 4h in this figure so that the daily activity onset is plotted to occur at the same time as the neutral photoperiod group (top).

Fig 3.

The average percentage of time spent asleep each hour (mean ± SE) 6–9 weeks after the short photoperiod group was initially exposed to 4:20 LD photoperiods. Sleep data from the animals on short photoperiod (brown) is delayed by 4 h in this figure so that the daily activity onset (and lights off) is plotted to occur at the same time as the neutral photoperiod group (blue). Although animals on short photoperiod exhibited similar anticipatory activity (and low sleep) ~13h before lights off, they were more likely to sleep during the subsequent 6h when they remained in darkness, relative to animals held on a neutral photoperiod.

Sucrose preference and consumption

After 4 weeks of exposure to a neutral or short photoperiod treatment, we did not find any evidence of anhedonia, as there were no differences in preference for 2% sucrose based on photoperiod (F_(1,20.0) = .13, p = .72) (Fig S1), nor were there significant effects of sex or evidence of a sex*photoperiod interaction (Table S1). After the SPT, half of the animals were gradually switched to an 8% sucrose solution, whereas the other half of animals were given a second bottle of water. Animals that were given a second water bottle drank significantly more from their original water bottle over the next three weeks (F_(1,297.0) = 62.1, p < .0001), with males and animals on short photoperiods exhibiting a stronger preference (Table S1). In contrast, animals with access to 8% sucrose strongly preferred to drink from the new bottle containing 8% sucrose solution (X² = 231.2, p < .0001); this preference was stronger in males but was not affected by photoperiod or the interaction between sex and photoperiod (Table S1). Bottle preference was not affected by photoperiod, regardless of whether the second bottle contained water (F_{(1, 19.013)} = 3.68, p = .070) or sucrose (F_{(1, 18.74)} = 3.33, p = .083). Thus, we failed to find evidence that carbohydrate craving (or at least consumption) was impacted by exposure to short photoperiods. Likewise, for the last two weeks of the experiment, the amount of time asleep during the subjective day was not affected by access to sucrose (F_(1,17) = 0.91, p = .353), by photoperiod treatment (F_(1,17) = 1.57, p = .227), or by an interaction between sucrose access and photoperiod (F_(1,17) = 4.27, p = .054).

Effects of photoperiod and sucrose consumption of body mass, fatpad mass and liver histology

Animals gained body mass across the experiment, increasing by an average of 22.6 ± 15.8% (SD) from the start to the end of the experiment (F_(1,43) = 172.8, p < .001) – exposure to neutral or short photoperiod conditions during the 4-week interval prior to sucrose exposure did not affect body mass gain (X² = .22, p = .64) (Fig S2). Body mass gain at the end of the experiment was not affected by photoperiod (X² = .93, p = .34) or access to sucrose (X² = .78, p = .38), or the interaction between photoperiod and sucrose access (X² = 7.26, p = .07; Fig S2). By the end of the experiment, some individuals were obviously obese, however, there was a no effect of photoperiod (F_(1,41) = 0.00, p = .98), sucrose (F_(1,41) = 1.77, p = .19) or their interaction (F_(1,41) = 1.99, p = .17) on fatpad mass (Fig S3). Examination of liver sections stained with H&E revealed that while some individuals had normal livers (Fig 5a), others showed evidence of non-alcoholic fatty liver disease (NAFLD), with both micro and macrosteatosis (Fig 5b). However, the percent area of the liver that was made up of macrosteatosis was not affected by photoperiod (F_(1,40.7) = 2.16, p = .15), by access to 8% sucrose solution (F_(1,40.7) = 0.72, p = .40), or by the interaction between photoperiod and access to sucrose (F_(1,40.7) = 1.06, p = .31) (Fig 5c). Additionally, we found that age did not affect percent macrosteatosis in the liver (Table S1). Staining of liver sections with Masson’s trichrome did not reveal any evidence of inflammation or fibrosis, suggesting that animals with NAFLD had not developed non-alcoholic steatohepatitis (NASH).

Fig 5.

Hematoxylin and eosin (H&E) stained histology images for (a) a normal grass rat liver section and (b) a grass rat liver with both macro- and micro-steatosis. Large white circles (where lipids have been washed away) is evidence of macrosteatosis, whereas greyish regions show much smaller white circles (many intracellular lipid droplets) characteristic of microsteatosis. (c) Percentage of macrosteatosis in liver samples across all treatment groups – no significant differences were detected across groups.

Microbiome analyses

PCoA analyses revealed microbiomes of fecal samples collected from home cages of grass rats or removed from the large intestine of grass rats after euthanasia were not altered upon exposure to high or low sucrose or varying photoperiods. Instead, individuals that shared the same parents (regardless of litter) had similar microbiomes (Fig 6), and microbiome composition changed little across the experiment (Fig S4). Further, Shannon Diversity Index for intestinal fecal samples was not affected by photoperiod (F_(1,39) = 2.602, p = .11), sugar (F_(1,39) = 0.04, p =.84) or their interaction (F_(1,39) = 2.94, p =.09) (Fig S5). Intestinal fecal microbiome diversity was not correlated with the percent area of liver macrosteatosis (F_(1,40.8) = 3.49, p = .07).

Fig 6.

Principal coordinates analyses (PCoAs) of the microbiome of fecal samples from cage floor (repeated across time; left panel) and from the large intestine after 8.5 weeks of photoperiod treatment and 4.5 weeks of sucrose treatment (right panel). Different photoperiod treatments are indicated by shape (triangle = neutral 12:12 LD; circle = short 4:20 LD), access to sucrose solution is indicated by fill (no fill = water only; fill = 8% sucrose), and color indicates parental pairs (all grass rats were obtained from 4 parental pairs). An individuals’ fecal microbiome changed little across time, regardless of exposure to short photoperiods or sucrose (left panel). The majority of variation in the gut microbiome was attributable to parental pairs with no significant effect of photoperiod or sucrose treatments (both panels).

Discussion

Diurnal rodent models have emerged as a tool for understanding the implications of photoperiod and light intensity on mood disorders, including SAD (reviewed in (Kronfeld-Schor and Einat, 2012; Yan et al., 2019; Shankar and Williams, 2021)). Prior work in diurnal grass rats, for example, revealed that dim light or short photoperiods can increase affective-like behaviors, as indicated by the forced swim test and saccharine preference test (Leach et al., 2013b). Further, these effects were accompanied by attenuated serotonergic function (Leach et al., 2013a, 2013b) and orexinergeic signaling (Deats et al., 2014) in brain regions involved in mood regulation. In the present study, we investigated whether short photoperiods also induce shifts in patterns of sleep, carbohydrate consumption, and metabolic dysbiosis in grass rats. Although daily patterns of sleep were altered by reductions in photoperiod, with animals sleeping more during dark conditions within their subjective day, we found no evidence for anhedonia in these animals (based on SPT). Further, diel rhythms remained robust with predictable increases in activity and reductions in sleep ~13h before lights off. This suggests grass rats lack the ability to contract the phase of the circadian cycle during which they are active, Similar to a prior study showing spontaneous development of metabolic syndrome in grass rats (Noda et al., 2010), some individuals in our study developed NAFLD with extensive macro- and micro-steatosis. Surprisingly, however, we found no evidence that short photoperiod or 3 weeks of access to high concentration sucrose solution affected the likelihood of developing NAFLD or altered the gut microbiome of grass rats. Instead, we found the gut microbiome was shaped by parental effects and changed little over the course of our study.

Diet, Photoperiod, and the gut microbiome

Remodeling of the gut microbiome has been proposed as an important factor contributing to the onset of major depressive disorders (Capuco et al., 2020) and may also contribute to metabolic dysbiosis (Agus et al., 2021), which has led to increased interest in delineating the factors that contribute to remodeling of the microbial community. Antibiotics, diet, and photoperiod have all been shown to alter the composition of the gut microbiome (Francino, 2016; Ren et al., 2020). For example, high sucrose diets have been shown to alter the gut microbiome and influence liver steatosis and/or lipids in mice and rats (Magnusson et al., 2015; Kong et al., 2019; Sun et al., 2021). Further, supplementing a normal diet with sugar water early in life causes remodeling of the gut microbiome in rats, though apparently without altering levels of obesity (Noble et al., 2017; De Marco et al., 2021). In contrast, we found no evidence that providing access to high concentration (8%) sucrose solution for three weeks altered the gut microbiome of grass rats and metabolic dysbiosis occurred independent of the gut microbiome. Notably, we found substantial individual variation in the composition of the gut microbiome, with grass rats in our study clustering into 3 distinct groups – clustering appeared to depend on the parents, suggesting strong parental inheritance of the gut microbiome. In addition to maternal inheritance of the gut microbiome shortly after birth (Daft et al., 2015), rodents engage in coprophagy such that both parents may be influencing the gut microbiome early in development (Bo et al., 2020). Our longitudinal analysis of the fecal microbiome indicated little change in composition across the study, regardless of treatment group. The lack of sensitivity we observed in grass rats contrasts with prior rodent studies - this may reflect differences between species or, alternatively, differences in the age of animals. We expect younger animals will have a gut microbiome that is more labile and sensitive to environmental factors.

Recent studies of temperate rodents reveal that photoperiod can also have substantial effects on the gut microbiome, even in adults. For example, hibernators such as the Siberian Hamster (Phodopus sungorus) and 13-lined ground squirrel (Ictidomys tridecemlineatus) undergo dramatic remodeling of the gut microbiome in response to changes in photoperiod (Carey et al., 2013; Ren et al., 2020). In ground squirrels, remodeling of the gut microbiome co-occurs with dramatic structural and functional remodeling of the gut (Kurtz et al., 2021) and it has now been shown that this remodeling likely facilitates gut microbiome-mediated urea nitrogen recycling during hibernation (Regan et al., 2022). Remodeling is not restricted to hibernators however, as Brandt’s voles (Lasiopodomys brandtii) also show pronounced changes in the gut microbiome in response to changing photoperiod (Zhu et al., 2022). These species are adapted to seasonal pulses in resource availability and exhibit pronounced seasonal changes in metabolism and reproduction in response to changes in photoperiod (Dardente et al., 2019). In contrast, Nile grass rats from equatorial populations have evolved in regions with relatively constant photoperiod (12:12 LD) throughout the year and, in captivity, reproductive condition is not affected by photoperiod (Nunes et al., 2002). Thus, lack of sensitivity of the gut microbial community to photoperiod may reflect a lack of evolved plasticity. A key challenge both in animal models and in humans is interpreting whether shifts in the gut microbiome are consequences of metabolic dysbiosis or represent adaptive shifts in response to environmental factors. While we found that remodeling of gut microbiome cannot be acutely induced by an artificial change in photoperiod, further study is needed to ascertain whether these animals display seasonal variations in their microbiome profile under natural free-living conditions. Regardless, we found no evidence that circadian disruption or shifts in the gut microbiome occur, and therefore these processes are unlikely to underly the development of metabolic syndrome in grass rats.

Does photoperiod alter affective-like behaviors in grass rats

Prior research in grass rats demonstrated that short photoperiods (8:16 LD) induce anhedonia, as evidenced by reduced consumption of saccharine solution, compared to neutral photoperiod controls (Leach et al., 2013b). Leach et al. (2013b) also found that animals on short photoperiods spent more time immobile during forced swim tests, which is frequently interpreted as evidence of behavioral despair (Castagné et al., 2010). Thus, prior studies support effects of photoperiod on affective-like behaviors. However, we found no effect of photoperiod on sucrose consumption rates, regardless of whether that was initial exposure to 2% sucrose, or eventual exposure to high 8% sucrose concentrations. There are several factors that may have contributed to the discrepancy between our study and the study of Leach et al. (2013b), including differences in the duration of photoperiods (4h vs. 8h), the use of sucrose vs. saccharine as a sweetener, and differences in when presence was compared (i.e., Leach et al. (2013b) used the first 24h as a base line and the preference was compared in the following 3 days whereas we compared days 5 & 6 after 4 days of acclimation). Further, animals in our study were older and single-housed (necessary for sleep-monitoring using the PiezoSleep system) which may have influenced mood and the sensitivity of mood to environmental cues.

Conclusions

Prior work in grass rats indicates that animals are sensitive to developing metabolic syndrome even when fed “normal” chow diets (Noda et al., 2010) and this situation is exacerbated by the feeding of a high fat diet (Ramanathan et al., 2022). In our study, a subset of animals in all treatment groups developed NAFLD, with high levels of macro- and/or microsteatosis. However, the propensity to develop NAFLD did not appear to be related to the gut microbiome, and both NAFLD and the gut microbiome were unaffected by 2 months of exposure to short photoperiods and ~3 weeks of access to high sucrose solutions. Animals that occupy temperate environments often respond to reduced photoperiods by undergoing pronounced physiological remodeling, including inactivation of the reproductive axis, the seasonal accumulation of adipose tissue stores, and changes in the composition of the gut microbiome. In contrast, the neuroendocrine pathway that links photoperiod to reproduction has apparently become disconnected in Nile grass rats from equatorial regions (Heideman and Bronson, 1990). Similarly, our results suggest that the gut microbiome and liver adiposity are also unaffected by photoperiod in these diurnal equatorial animals. Thus, while prior research suggests that short photoperiods induce affective behaviors in grass rats (Leach et al., 2013b), we find no evidence that it exacerbates metabolic dysbiosis in this species.

Fig 4.

The proportion of either water (left) or sucrose solution (8%; right) consumed from the second bottle added to the animal tub during the three weeks of access to 8% sucrose solution, relative to the amount consumed in the water bottle that has always been present in the tub. Regardless of photoperiod treatment, grass rats preferred consuming water from the original water bottle relative to the new water bottle (left). However, animals strongly preferred consuming sucrose solution from the new water bottle rather than water from the original water bottle (right). Photoperiod treatment did not have a significant effect. Points are for each day for each individual (max 9 points/individual); days where water bottles leaked are excluded.

Data accessibility

All the R code used to analyze data in this paper are available on github: https://github.com/ckdeal/Grass_Rats. Data used in this paper are available from Dryad: https://doi.org/10.5061/dryad.n2z34tn3g. Sequence data can be found at the following link: https://www.ncbi.nlm.nih.gov/sra/PRJNA1006454.