External light‐dark cycle shapes gut microbiota through intrinsically photosensitive retinal ganglion cells

Article27 April 2022free access Transparent process External light-dark cycle shapes gut microbiota through intrinsically photosensitive retinal ganglion cells Chi-Chan Lee Chi-Chan Lee Department of Life Science, National Taiwan University, Taipei, Taiwan Search for more papers by this author Feng Liang Feng Liang Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Data curation, Formal analysis Search for more papers by this author I-Chi Lee I-Chi Lee Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Data curation, Formal analysis Search for more papers by this author Tsung-Hao Lu Tsung-Hao Lu Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Data curation, Formal analysis Search for more papers by this author Yu-Yau Shan Yu-Yau Shan orcid.org/0000-0003-4157-9250 Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Data curation, Formal analysis Search for more papers by this author Chih-Fan Jeng Chih-Fan Jeng Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Formal analysis Search for more papers by this author Yan-Fang Zou Yan-Fang Zou orcid.org/0000-0002-3646-8591 Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Data curation Search for more papers by this author Hon-Tsen Yu Hon-Tsen Yu Department of Life Science, National Taiwan University, Taipei, Taiwan Genome and Systems Biology Degree Program, National Taiwan University and Academia Sinica, Taipei, Taiwan Contribution: Conceptualization, Methodology Search for more papers by this author Shih-Kuo Chen (Alen) Corresponding Author Shih-Kuo Chen (Alen) [email protected] orcid.org/0000-0002-7921-1358 Department of Life Science, National Taiwan University, Taipei, Taiwan Genome and Systems Biology Degree Program, National Taiwan University and Academia Sinica, Taipei, Taiwan Contribution: Conceptualization, Supervision, Funding acquisition, Writing - original draft, Project administration, Writing - review & editing Search for more papers by this author Chi-Chan Lee Chi-Chan Lee Department of Life Science, National Taiwan University, Taipei, Taiwan Search for more papers by this author Feng Liang Feng Liang Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Data curation, Formal analysis Search for more papers by this author I-Chi Lee I-Chi Lee Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Data curation, Formal analysis Search for more papers by this author Tsung-Hao Lu Tsung-Hao Lu Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Data curation, Formal analysis Search for more papers by this author Yu-Yau Shan Yu-Yau Shan orcid.org/0000-0003-4157-9250 Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Data curation, Formal analysis Search for more papers by this author Chih-Fan Jeng Chih-Fan Jeng Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Formal analysis Search for more papers by this author Yan-Fang Zou Yan-Fang Zou orcid.org/0000-0002-3646-8591 Department of Life Science, National Taiwan University, Taipei, Taiwan Contribution: Data curation Search for more papers by this author Hon-Tsen Yu Hon-Tsen Yu Department of Life Science, National Taiwan University, Taipei, Taiwan Genome and Systems Biology Degree Program, National Taiwan University and Academia Sinica, Taipei, Taiwan Contribution: Conceptualization, Methodology Search for more papers by this author Shih-Kuo Chen (Alen) Corresponding Author Shih-Kuo Chen (Alen) [email protected] orcid.org/0000-0002-7921-1358 Department of Life Science, National Taiwan University, Taipei, Taiwan Genome and Systems Biology Degree Program, National Taiwan University and Academia Sinica, Taipei, Taiwan Contribution: Conceptualization, Supervision, Funding acquisition, Writing - original draft, Project administration, Writing - review & editing Search for more papers by this author Author Information Chi-Chan Lee1, Feng Liang1, I-Chi Lee1, Tsung-Hao Lu1, Yu-Yau Shan1, Chih-Fan Jeng1, Yan-Fang Zou1, Hon-Tsen Yu1,2 and Shih-Kuo Chen (Alen) *,1,2 1Department of Life Science, National Taiwan University, Taipei, Taiwan 2Genome and Systems Biology Degree Program, National Taiwan University and Academia Sinica, Taipei, Taiwan *Corresponding author. Tel: +886 2 3366 1571; E-mail: [email protected] EMBO Reports (2022)23:e52316https://doi.org/10.15252/embr.202052316 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Abstract Gut microbiota are involved in many physiological functions such as metabolism, brain development, and neurodegenerative diseases. Many microbes in the digestive tract do not maintain a constant level of their relative abundance but show daily oscillations under normal conditions. Recent evidence indicates that chronic jetlag, constant darkness, or deletion of the circadian core gene can alter the composition of gut microbiota and dampen the daily oscillation of gut microbes. However, the neuronal circuit responsible for modulating gut microbiota remained unclear. Using genetic mouse models and 16s rRNA metagenomic analysis, we find that light-dark cycle information transmitted by the intrinsically photosensitive retinal ganglion cells (ipRGCs) is essential for daily oscillations of gut microbes under temporal restricted high-fat diet conditions. Furthermore, aberrant light exposure such as dim light at night (dLAN) can alter the composition, relative abundance, and daily oscillations of gut microbiota. Together, our results indicate that external light-dark cycle information can modulate gut microbiota in the direction from the brain to the gut via the sensory system. Synopsis In addition to the endogenous circadian clock of the host, the external light-dark cycle can modulate gut microbiota composition, diversity, and daily oscillation through intrinsically photosensitive retinal ganglion cells. Mice that lack intrinsically photosensitive retinal ganglion cells (ipRGCs) or melanopsin have a distinct gut microbiota composition from WT mice. Light dark cycle information regulates the daily oscillation of gut microbiota through the ipRGC-sympathetic nerve pathway. Dim light at night causes dysbiosis through sympathetic nerve-dependent and -independent pathways. Introduction Gut microbiota could influence the central nervous system through a proposed gut-brain axis to modulate neuronal development, function, and degeneration. It has been shown that germ-free mice have different hypothalamus metabolites compared with normal mice. In addition, germ-free or antibiotic-treated mice have different gene expression profiles in the CNS and lower anxiety levels compared with normal SPF mice (Sudo et al, 2004; Bercik et al, 2011; Diaz Heijtz et al, 2011; Neufeld et al, 2011). The development of the endocrine neuron such as oxytocin neurons is affected by the gut microbe during the postnatal period, which will influence the adult stage social interaction (Buffington et al, 2016). Finally, the gut microbiota will also influence the aggregation of alpha-synuclein in the gut, which could lead to Parkinson’s disease (Sampson et al, 2016). The composition of gut microbiota and the relative abundance of specific microbes could be influenced by the host’s genetic factors or feeding scheme. Interestingly the relative abundance of many gut microbes displays daily oscillation even under normal condition (Thaiss et al, 2014; Voigt et al, 2014; Zarrinpar et al, 2014; Leone et al, 2015). Although the direct physiological implication of the gut microbe daily oscillation is poorly understood, a recent report showed that arrhythmic gut microbe is associated with patients with type 2 diabetes (Reitmeier et al, 2020). Recent evidence showed that chronic jetlag, constant darkness, and reversal of the light-dark cycle could disrupt the daily oscillation of gut microbe (Thaiss et al, 2014; Voigt et al, 2014; Liang et al, 2015; Wu et al, 2018). In addition, mice with clock gene knockout also display arrhythmic gut microbiota (Thaiss et al, 2014; Wu et al, 2018). These studies suggest that both external and internal factors could influence the daily oscillation of gut microbe, and their interaction may play an important role in gut homeostasis and microbiota (Mukherji et al, 2013; Thaiss et al, 2014; Voigt et al, 2014; Liang et al, 2015; Wu et al, 2018; Godinho-Silva et al, 2019; Kuang et al, 2019). However, how external factors influence the gut microbiota is unknown. One candidate for modulation of daily gut microbe oscillation is the circadian clock system. This system is entrained into the daily light-dark cycle (LD) via input from the melanopsin-expressing, intrinsically photosensitive retinal ganglion cells (ipRGCs) to the suprachiasmatic nucleus (SCN) (Berson et al, 2002; Hattar et al, 2002, 2003; Provencio et al, 2002; Guler et al, 2008; Fernandez et al, 2016). In addition, ipRGCs also influence many other non-image-forming functions by projecting to brain regions in the thalamus and hypothalamus such as the olivary pretectal nucleus for pupillary light reflex (Lucas et al, 2003; Hattar et al, 2006; Baver et al, 2008; Chen et al, 2011). Genetic elimination of ipRGCs using the Opn4-DTA or Opn4-Cre; DTR mouse lines impairs the ability of mice to transmit external light-dark cycle information for circadian photoentrainment which causes these mice to “free run” under any kind of environmental light-dark cycle (Hatori et al, 2008; Prigge et al, 2016; Chew et al, 2017). On the other hand, knockout of the photopigment melanopsin (MKO) only produced a light detection phenotype for non-image-forming functions under high light intensity (Hattar et al, 2003; Lucas et al, 2003; Schmidt & Kofuji, 2010), which suggests that ipRGCs transmit signals to the brain by combining high luminance signals detected via melanopsin and low luminance signals originating from the canonical rod and cone photoreceptors. Recent evidence showed that ipRGCs can modulate many additional physiological functions such as emotion, hair regeneration, and body temperature (Fan et al, 2018; Fernandez et al, 2018; Rupp et al, 2019). Since ipRGCs provide environmental luminance signals for many non-image-forming visual functions, it is likely that they could also influence gut microbiota. Using 16s rRNA analysis and various kinds of ipRGC-related mutant mice, here we show that light-dark cycle information could drive the daily oscillation of gut microbes through the ipRGCs-sympathetic nerve circuit. Furthermore, aberrant light-dark cycles such as light exposure during the nighttime could cause dysbiosis and dampen the daily oscillation of gut microbes. Results Light/dark cycle information is important to drive daily oscillations of gut microbes To determine how the light-dark cycle may directly influence the composition of gut microbiota, we designed a specific dim light at night (dLAN) conditions to provide light exposure while minimizing the disruption of the central circadian clock. To control feeding time, we modified the dLAN condition by incorporating a temporal restricted feeding cycle with a high-fat diet to eliminate daytime feeding, a behavioral phenotype that could disrupt the circadian clock (Fonken et al, 2010). Mice housed under our modified dLAN conditions could only eat during the nighttime and have similar circadian behaviors and gene expression patterns compared with normal light-dark cycle (Fig EV1). The average nighttime food intake is similar between all experimental setups (Fig EV2). Together, our results indicated that the central clock remains entrained and shows daily oscillation similar to the normal LD cycle when mice were exposed to 25 lux of light during the subjective nighttime under the temporal restricted high-fat diet condition. Therefore, our modified dLAN conditions allowed us to test the influence of additional light exposure during the nighttime while keeping mice entrained to 24 h daily light-dark cycle. Click here to expand this figure. Figure EV1. Mice housed in dLAN condition display similar activity and clock gene expression patterns to LD condition Schematic representation of the experimental design for mice exposed to dim light at night (dLAN) or normal light/dark (LD). LD mice were housed under a 12h:12h LD cycle with 0 lux during the dark phase; dLAN mice were housed under 25 lux during the dark phase. Food availability was restricted to the dark phase for both conditions. Temporal activity profile of mice, plotted by the moving distance in the home cages per hour, under dim light at night (dLAN, red) or normal light/dark (LD, blue) conditions. n = 5 for each group. qPCR analysis of expression levels of the circadian clock genes Bmal1 and Per2 in the suprachiasmatic nucleus (SCN) and liver from control mice housed under conditions of dLAN (gray) and LD (black). Representative images of Per2 immuno-positive cells in the SCN at ZT4, ZT10, ZT16, and ZT22 from control mice housed under LD (upper) and dLAN (lower). Quantification of Per2 immuno-positive cell in the SCN. There is no significant difference in the Per2 positive cell number between mice housed under LD and dLAN conditions. Daily food intake measured in Kcal is similar between all experimental groups. n = 4–5 mice per group. Data information: No significant difference is reported using 2-way ANOVA. n = 4–5 for qPCR, n = 3 for IHC. Scale bar is 100 μm for (D). Data are presented as mean ± SEM. Download figure Download PowerPoint Click here to expand this figure. Figure EV2. Actogram of Opn4DTA/DTA mice Wheel-running activity of Opn4DTA/DTA mice immediately after final fecal sample collection. Opn4DTA/DTA mice, which activity onsets are at least 3 h in advance or delay to the light off, housed under LD or dLAN condition. Opn4DTA/DTA mice housed under dLAN condition with activity onsets less than 3 h in advance or delay to the light off. Data information: Yellow circles indicate predicted onset for fecal sample-collecting days. Red triangles indicate fecal sample-collecting times. Download figure Download PowerPoint To determine the impact of the light-dark cycle on gut microbes, we first compared the gut microbiota under normal LD conditions, under constant darkness (DD) with reduced light exposure, or under dLAN conditions with extra light exposure. After 2 weeks of normal LD cycle, constant darkness, or dLAN conditions, fecal samples were collected from WT mice at six time points throughout the day. We perform 16S rRNA-based next-generation sequencing and use the operational taxonomic unit (OTU) based JTK_CYCLE (Hughes et al, 2010) to analyze the daily gut microbial oscillation in mice housed under different light-dark cycles. For WT mice, many microbes display daily oscillations under LD conditions (Fig 1A and D). To compare the amount of oscillating gut microbe between different light-dark cycles, we added all microbes with both P-value and q-value (false-positive rate) smaller than 0.05 in JTK _CYCLE analysis to calculate the total percentage of gut microbe displayed daily oscillation (Fig 1C). Surprisingly, although the locomotor activity of mice and most circadian clock genes were entrained to 24 h cycle under dLAN conditions (Fig EV1), percentages of oscillating gut microbes were highly reduced in mice housed under dLAN conditions (Fig 1A, C and D). We also confirm that the daily oscillation of gut microbes is significantly reduced in mice housed under the DD condition, even with free-running endogenous clock (Fig 1B, C and D). The temporal restriction feeding with a high-fat diet is not sufficient to drive the full daily oscillation of gut microbe for mice housed under dLAN and constant darkness conditions. Together, our results suggest that the environmental light-dark cycle is one of the major factors driving the daily oscillation of gut microbe. Figure 1. Light-dark cycle is important to drive the daily oscillation of gut microbes A, B. Oscillating OTUs detected by the JTK_CYCLE (adjust P < 0.05, Benjamini–Hochberg q-value < 0.05) in WT mice housed under normal light-dark cycle (A, LD), dim light at night (A, dLAN), and constant darkness (B) conditions. Dash lines indicate the Benjamini–Hochberg q-value < 0.05. Graphs without dash line indicate that the q-values for all OTU were higher than 0.05. Both adjusted P-value and q-value were generated by JTK cycle. C. Summation of relative abundance from all oscillating OTUs from A and B. D. Relative abundance heat map of oscillating OTUs detected by the JTK_CYCLE. Each row represents 1 OTU across the day, and the graph is double plotted. Gray color indicates OTU not found. n = 4–5 for each time point. Data information: n = 4–5 mice. *P < 0.05 with the 1-way ANOVA Tukey post hoc test. All bar graphs are presented as mean ± SEM. Download figure Download PowerPoint Daily oscillation of gut microbes is driven by the light-dark cycle through ipRGCs It has been shown that ipRGC is the primary conduit to transmit light information for many non-image-forming visual functions such as circadian photoentrainment and pupillary light reflex in mammals. Although the free-running central clock could not fully drive the daily oscillation of gut microbes, ipRGCs and/or melanopsin signaling may be involved in transmitting light-dark cycle information to shape the gut microbiota. To test this hypothesis, we analyzed the daily oscillation of gut microbe using JTK_CYCLE from control, MKO, and ipRGC-eliminated (Opn4DTA/DTA) mice housed under normal LD cycle or dLAN conditions (Fig 2A–C). First, we compared the gut microbe oscillation between control and MKO mice. Interestingly, the pattern of daily oscillations in MKO differed from that in control mice under LD conditions, while the total oscillation percentage in MKO is still similar to control (Fig 2D and E). In control mice, peaks of oscillating OTUs were spread relatively evenly throughout the day (Fig 2F), whereas in MKO mice, peak times for many oscillating OTUs occurred between ZT8 and ZT12 (Fig 2G). These de novo microbe oscillations in MKO mice suggested that rod and/or cone signals might partially compensate for the loss of melanopsin. Next, we test whether ipRGC is the sole conduit to transmit melanopsin and rod/cone signal for gut microbe daily oscillation by comparing the rhythmicity between control and Opn4DTA/DTA mice. Because the running activities of ipRGC-eliminated (Opn4DTA/DTA) mice were not tied to the environmental light-dark cycle, we first collected fecal samples according to the light-dark cycle and immediately recorded their wheel-running activity afterward (Fig EV2). Therefore, JTK_CYCLE analysis for Opn4DTA/DTA mice was performed by arranging data points to match either circadian time (CT) predicted from the wheel-running activity (Fig EV2) or zeitgeber time (ZT) according to the light-dark cycle. Strikingly, in Opn4DTA/DTA mice, daily gut microbial oscillations were greatly attenuated (< 3%) under both LD and dLAN conditions when analyzed with either CT or ZT time (Fig 2H). This result indicates that ipRGC is essential to provide normal light-dark cycle information from melanopsin and rod/cone to drive the gut microbe daily oscillation. Figure 2. Light-dark cycle information transmitted by ipRGC is required for gut microbe oscillation A–C. Oscillating OTUs detected by the JTK_CYCLE (adjust P < 0.05, Benjamini–Hochberg q-value < 0.05) in control (A), Opn4DTA/DTA (B), and MKO (C) mice housed under LD and dLAN conditions. Dash lines indicate the Benjamini–Hochberg q-value < 0.05. Graphs without dash line indicate that the q-values for all OTU were higher than 0.05. Both adjusted P-value and q-value were generated by JTK cycle. D. Summation of relative abundance from all oscillating OTUs in control and MKO mice. E. Venn diagram of oscillating OTUs in control and MKO mice. F, G. Relative abundance heat map of oscillating OTUs detected by the JTK_CYCLE in LD or dLAN conditions from control (F) or MKO (G) mice. Each row represents 1 OTU across the day, and the graph is double plotted. H. Summation of relative abundance from all oscillating OTUs in Opn4DTA/DTA mice according to the circadian time (CT) of the mice, or the light/dark and food cycle (ZT). Data information: n = 4–6 mice. *P < 0.05, **P < 0.01 using the 2-way ANOVA Bonferroni post hoc test. All bar graphs are presented as mean ± SEM. Download figure Download PowerPoint Light signals transmitted by ipRGC influenced gut microbiota composition In addition to modulating the daily oscillation of particular gut microbes, we next asked whether ipRGCs could modulate the overall composition of gut microbiota. We used principal coordinate analysis (PCoA) to compare the gut microbiota composition between control, MKO, and Opn4DTA/DTA mice. Analysis of mean and variation, using ANOSIM (Fierer et al, 2010) and Adonis (PERMANOVA) (Anderson, 2008), respectively, showed that gut microbiota composition is significantly different between control and MKO mice (Fig 3A), and also significantly different between control and Opn4DTA/DTA mice (Fig 3B). The alpha diversity (richness) was significantly lower for MKO and Opn4DTA/DTA mice compared with control mice (Fig 3C), while beta diversity (variation between samples) from MKO and Opn4DTA/DTA mice was both significantly higher than control mice (Fig 3D). These results indicate that melanopsin signaling and ipRGCs are important factors in maintaining normal gut microbiota composition. Next, we compared the gut microbiota from control, MKO, and Opn4DTA/DTA mice housed under LD or dLAN condition. The PCoA analysis showed a clear separation of gut microbiota from control mice housed under LD versus dLAN conditions (Fig 3E), indicating nighttime light exposure could modulate overall gut microbiota in mice. In contrast, neither MKO nor DTA animals showed separation of gut microbiota when mice were housed in LD or dLAN condition (Fig 3F and G). Moreover, neither alpha nor beta diversity of microbiota differed between LD and dLAN conditions in MKO and Opn4DTA/DTA mice, though they were significantly different in controls (Fig 3C and D). Since Opn4DTA/DTA free run under any light-dark cycle, the activity time of mice was sometimes synchronized to the feeding schedule and then desynchronized to the feeding schedule during the course of the experiment. Interestingly, PCoA analysis resulted in a strong separation of gut microbiota between Opn4DTA/DTA mice with their circadian clock synchronized or not synchronized to the feeding schedule under any light-dark cycle (Fig 3H). These results showed that dyssynchronization between the feeding schedule and the endogenous circadian clock could modulate the gut microbiota in ipRGC-eliminated mice. However, aberrant light-dark cycle-induced dysbiosis is blocked in both melanopsin knockout and ipRGC-eliminated mice. Figure 3. dLAN altered composition of gut microbiota A. Principal coordinate analysis (PCoA) of gut microbiota using weighted UniFrac matrix from control and MKO mice housed under LD conditions. B. PCoA of gut microbiota from control and Opn4DTA/DTA mice housed under LD conditions. The gut microbe compositions are significantly different between control and ipRGC-manipulated mutant mice with both ANOSIM and Adonis analysis. C, D. Violin plots of the alpha diversity using Shannon index (C) and the beta diversity generated from the weighted UniFrac distance matrix (D) for gut microbiota from control, MKO, and Opn4DTA/DTA mice under LD and dLAN condition. There is no significant difference under dLAN condition. E–G. PCoA of gut microbiota from control, MKO, and Opn4DTA/DTA (with similar phase-angle difference to the light-dark cycle) mice housed under LD or dLAN conditions. Both ANOSIM and Adonis had significant differences for control mice housed under LD or dLAN conditions (E), but not for MKO (F) and Opn4DTA/DTA mice (G). H. PCoA of gut microbiota from Opn4DTA/DTA mice either synchronized or desynchronized to the light/dark and feeding cycle. Data information: n = 4–5 mice. For violin plots (C and D), dot lines are 25 and 75% quartile lines. The dashed line represents the mean between groups; *P < 0.01 using the 2-way ANOVA Bonferroni post hoc test; #P < 0.05 between same genotype using the student t-test; n.s. not significant. ANOSIM and ADONIS tests and their P-values were indicated in each PCoA graph. Download figure Download PowerPoint Next, we compared the relative abundance of each gut microbe under different light-dark cycles. The apparent gut microbe composition at the phylum level did not differ significantly among groups (Fig 4A). However, linear discriminant analysis at multiple phylogenetic levels showed that thirteen classifications, including the genus Rikenella, Odoribacter, and Lactobacillus, were significantly different in control mice housed under dLAN versus LD conditions (Fig 4B), whereas in MKO, only one classification had significant differences between dLAN and LD conditions (Fig 4C). Strikingly, there were 23 classifications with significant differences between dLAN and LD conditions in Opn4DTA/DTA mice. However, none of them were included in the 13 differential classifications in control mice housed under LD and dLAN conditions (Fig 4D). Since Opn4DTA/DTA mice were free running under our 12h:12h light-dark cycle, the difference in microbiota pattern from Opn4DTA/DTA mice may, again, be caused by dyssynchronization between the feeding schedule and the endogenous circadian clock. Together, our results suggested that light information through activation of melanopsin and/or transmitted by ipRGCs could influence the composition and diversity of gut microbiota. Figure 4. Enrichment of specific microbes from mice housed under dLAN and LD conditions Mean relative abundance of phylum from control (Con)-, melanopsin knockout (MKO)-, and ipRGC-eliminated (Opn4DTA/DTA) mice housed under conditions of normal light/dark (LD) or dim light at night (dLAN) (n = 4–5 from each group). The cladogram shows the difference in relative abundance of microbes from control mice housed under conditions of LD and dLAN. Microbes in red have significantly higher relative abundance in LD conditions. Microbes in blue have significantly higher relative abundance in dLAN conditions. The cladogram shows the difference in relative abundance of microbes from MKO mice under LD and dLAN conditions. The cladogram shows the difference in relative abundance of microbes from Opn4DTA/DTA mice under LD and dLAN conditions. For RF39, it is higher in the dLAN condition from control mice but higher in the LD condition from Opn4DTA/DTA mice. Download figure Download PowerPoint ipRGC drive light-evoked, daily oscillations of gut microbes via sympathetic nerves It has been shown that light can influence peripheral organs via the retinohypothalamic tract, independent of the circadian clock through autonomic circuits (Kiessling et al, 2014). Therefore, to determine the neural circuit responsible for the light-evoked modulation of gut microbiota, we tested whether the sympathetic nerve was involved in dLAN-induced dysbiosis. To eliminate sympathetic nerve function, WT mice were treated with 6-Hydroxydopamine (6-OHDA) (Thoenen & Tranzer, 1968). Body weight of mice decreased during the first week of 6-OHDA injection (Fig 5A), consistent with the successful elimination of sympathetic nerve function (Joseph et al, 1979). Furthermore, we did not observe any change in the number of dopaminergic amacrine cells with TH-immunostaining, suggesting that 6-OHDA may not penetrate across the blood-retinal barrier (Fig EV3). After body weight stabilized, mice were housed under either dLAN or LD conditions for 2 weeks. Although the gut microbiota oscillation in the PBS control group is much higher than the LD condition due to daily injection throughout the experimental period, the daily oscillation showed a trend of decline under dLAN compared with LD conditions (Fig 5B and D). The gut microbiota composition was also clustered into 2 groups (ANOSIM P = 0.054) (Fig 5E). Surprisingly, sympathetic nerve ablation significantly dampened the daily oscillations of gut microbes under both LD and dLAN conditions (Fig 5C and D), similar to our observations in ipRGC-ablated animals. These data suggest that light input driving daily oscillations occurs via the sympathetic nervous system. Interestingly, sympathetic nerve ablation failed to abolish the effects of dLAN-induced dysbiosis, as dLAN still altered the composition of gut microbiota under this manipulation (Fig 5F), suggesting that these influences occur via a sympathetic nerve-independent circuit. Figure 5. Elimination of sympathetic nerve by 6-OHDA inhibits daily oscillation of gut microbes A. After 1 week of initial injection (baseline), about 10% of transient body weight loss was observed in the 6-OHDA group (open circle) but not in the PBS injection group (closed circle) compared with pretreat time point. The body weight recovered after 1 week, although 100mg/kg of 6-OHDA was injected every 2 days throughout the experiment. LD = light-dark cycle, dLAN = dim light at night. B, C. OTU-based JTK_CYCLE analysis from PBS (B) and 6-OHDA (C) treated mice housed under LD and dLAN conditions. Dash lines indicate the Benjamini–Hochberg q-value