Hypocretin underlies the evolution of sleep loss in the Mexican cavefish

The duration of sleep varies dramatically between species, yet little is known about the genetic basis or evolutionary factors driving this variation in behavior. The Mexican cavefish, Astyanax mexicanus, exists as surface populations that inhabit rivers, and multiple cave populations with convergent evolution on sleep loss. The number of Hypocretin/Orexin (HCRT)-positive hypothalamic neurons is increased significantly in cavefish, and HCRT is upregulated at both the transcript and protein levels. Pharmacological or genetic inhibition of HCRT signaling increases sleep in cavefish, suggesting enhanced HCRT signaling underlies the evolution of sleep loss. Ablation of the lateral line or starvation, manipulations that selectively promote sleep in cavefish, inhibit hcrt expression in cavefish while having little effect on surface fish. These findings provide the first evidence of genetic and neuronal changes that contribute to the evolution of sleep loss, and support a conserved role for HCRT in sleep regulation.

(R 2 =0.229) in Pachón cavefish, consistent with the notion that increased HCRT signaling is 8 associated with sleep loss (Fig 5G). A weak correlation between BoTx-expressing HCRT 9 neurons and sleep duration (R 2 =0.091) was also observed in surface fish Fig 5G, consistent  10 with the notion that HCRT may also regulates sleep in surface fish, though to a lesser extent 11 than in cavefish populations. quantified hcrt mRNA and protein levels in adult cave and surface fish following gentamicin 2 treatment. Quantitative PCR analysis revealed that hcrt expression was significantly reduced in 3 cavefish treated with gentamicin to levels similar to untreated surface fish (Fig. 6E). We 4 observed a non-significant decrease in hcrt expression following gentamicin treatment in 5 surface fish, but this effect was much smaller than that for cavefish. Using a HCRT-specific 6 antibody and immunohistochemistry, we found that lateral line ablation does not significantly 7 affect the number of HCRT-positive hypothalamic neurons in surface or cavefish, but does 8 significantly reduce the level HCRT protein within each cell in cavefish but not in surface fish 9 ( Fig. 6F-K). Together, these findings suggest that sensory input from the lateral line promotes 10 sleep and hcrt expression in Pachón cavefish, providing a link between sensory input and 11 transcriptional regulation of a wake-promoting factor. 12 In addition to its potent role in sleep regulation, hcrt promotes food consumption in fish and HCRT levels per cell in surface fish (Fig. 7B). These results indicate that the starvation 25 modulates HCRT levels, rather than the number of cells that produce HCRT. The acute 1 regulation of HCRT by feeding state and lateral-line dependent sensory input demonstrates a 2 unique link between these neuronal systems and those mediating sleep/wake cycles. 3

Discussion 1
Cavefish are a unique model for investigating neural and genetic regulation of sleep, particularly 2 from an evolutionary vantage point. Robust phenotypic differences in sleep have been observed 3 in multiple populations of cavefish, but our findings provide the first evidence for a mechanism 4 that underlies sleep variation between surface and cave fish. Alignment of hcrt sequences 5 derived from surface and Pachón cavefish indicate that there are no differences in the protein 6 sequence between the two morphs. Our findings do, however, reveal dramatic differences in 7 hcrt expression and neuron number between surface fish and cavefish. These findings suggest 8 that functional differences between surface fish and cavefish likely occur at the level of changes 9 in genomic enhancers or neuronal connectivity, which affect hcrt expression. Because our 10 findings also reveal an increased number of HCRT-positive neurons in early development, it is 11 also likely that developmental differences between the brains of surface and cavefish underlie 12 differences in hcrt function. Examination of cell body number in 6 dpf fry reveals increased 13 HCRT-positive neurons in cavefish, indicating HCRT differences are present during early 14 development. In agreement with these findings, broad anatomical differences in forebrain 15 structure have previously been documented between surface fish and cavefish including an 16 expanded hypothalamus (Menuet et al., 2007). Therefore, it is likely that developmentally-17 derived differences in the number of HCRT-positive neurons and modified hypothalamic neural 18 circuitry contribute to sleep loss in cavefish. 19 Multiple lines of evidence presented here support a robust role for evolved differences in HCRT 20 function in the evolution of sleep loss in Pachón cavefish. Pharamacological inhibition of HCRT 21 signaling, targeted knockdown of HCRT using morpholinos, or genetic inhibition of HCRT 22 neurons promote sleep in Pachón cavefish without significantly affecting sleep in surface fish. 23 While manipulations that inhibit HCRT signaling have potent affects on cavefish sleep, it does 24 not rule out a possible role for HCRT modulation of sleep in surface fish. All three 25 manipulations employed are likely to only partially disrupt HCRTR signaling, and it is possible 1 that complete inhibition of HCRTR signaling would increase sleep in surface fish. Indeed, Hcrt 2 is highly conserved and has been shown to promote wakefulness in species including in 3 animals ranging from zebrafish to mammals (Mieda et al., 2004;Prober et al., 2006). The 4 finding presented here extend studies in mammalian and zebrafish models, and suggest that 5 regulation of HCRT signaling may be subject to evolutionary pressure, and implicate it as a 'hot-6 spot' for variation in sleep throughout the animal kingdom. 7 While the neural processes regulating HCRT activity are not fully understood, growing evidence 8 suggests these neurons integrate sleep-wake regulation with responses to sensory stimuli 9 indicating that lateral line input is a potent regulator of hcrt production in cavefish. These 20 discoveries suggest that evolution of sensory systems can affect central brain processes that 21 regulate behavior, and provide further support that HCRT neurons integrate sensory stimuli to 22 modulate sleep and arousal. 23 While a full understanding of the neural circuitry regulating HCRT neurons has not been 24 determined, these neurons project to numerous areas implicated in behavioral regulation, 25 including the periventricular hypothalamus, raphe, and thalamic nuclei (Panula, 2010). Evidence 1 suggests that the wake-promoting role of HCRT neurons is dependent on norepinephrine 2 (Carter et al., 2012; Singh et al., 2015), and optogenetic activation of HCRT neurons activates 3 the locus coeruleus (Singh et al., 2015), raising the possibility that activation of this arousal 4 pathway is enhanced in Pachón cavefish. We previously demonstrated that treating cavefish 5 with the β-adrenergic inhibitor propranolol restores sleep in cavefish without affecting sleep in 6 surface fish (Duboué et al., 2012), similar to findings observed in this study in fish treated with 7 the HCRTR inhibitor TCS0X229. Therefore, it is possible that differences in norepinephrine 8 signaling contribute to sleep loss in cavefish. Further investigation of the synergistic effects of 9 norepinephrine and hcrt in surface and cavefish, and the effects of their pre-supposed 10 interaction on feeding-and sensory-mediated hcrt production, will be critical in our 11 understanding of how sleep changes can be driven by alterations in the environment. 12 In addition to its role in sleep-wake regulation, hcrt neurons regulate feeding and metabolic 13 function, raising the possibility that HCRT neurons are integrators of sleep and metabolic state. 14 Previous findings reveal that injection of HCRT peptide increases food consumption in cavefish, 15 suggesting the consummatory behavior induced by HCRT in mammals is conserved in A. . Future studies will reveal if enhanced HCRT function represents 7 a conserved mechanism for sleep loss, or that sleep loss in other fish populations is HCRT 8 independent. 9

Fish maintenance and rearing 2
Animal husbandry was carried out as previously described (Borowsky, 2008b) and all protocols 3 were approved by the IACUC Florida Atlantic University. Fish were housed in the Florida 4 Atlantic University core facilities at 21°C ± 1°C water temperature throughout rearing for 5 behavior experiments (Borowsky, 2008b). Lights were kept on a 14:10 hr light-dark cycle that 6 remained constant throughout the animal's lifetime. Light intensity was kept between 25-40 Lux 7 for both rearing and behavior experiments. All fish used for experiments were raised to 8 adulthood and housed in standard 18-37 L tanks. Adult fish were fed a mixture diet of black 9 worms to satiation twice daily at zeitgeber time (ZT) 2 and ZT12, (Aquatic Foods, Fresno, CA,) 10 and standard flake fish food during periods when fish were not being used for behavior 11 experiments or breeding (Tetramine Pro). 12 13

Sleep behavior 14
Adult fish were recorded in standard conditions in 10 L tanks with custom-designed partitions 15 that allowed for five fish (2 L/fish) to be individually housed in each tank as previously described 16 platform. Larvae were allowed to acclimate for 24 hours before starting behavioral recordings. 25 Videos were recorded using Virtualdub, a video-capturing software (Version 1.10.4) and were 1 subsequently processed using Ethovision XT 9.0 (Noldus, IT). Water temperature and chemistry 2 were monitored throughout recordings, and maintained at standard conditions in all cases. at 1-2 cell stage using a pulled borosilicate capillary with a Warner PLI-A100 picoinjector. 23 Survival of all embryos was monitored every 6 h for the first 96 h of development until behavior 24 was recorded over the next 24 h. hcrt:Gal4; uas:BoTxBLC-GFP injections were carried out as 25 follows: Separate plasmids containing the Gal4 and the BoTx transgenes were coinjected into 1-1 4 cell stage embryos using a pulled borosilicate capillary with Warner PLI-A100 picoinjector at 2 concentration of 25 ng/µL. Tol2 mRNA was coinjected in the cocktail at a concentration of 25 3 ng/µL. uas:BoTx-BLC-GFP was injected alone at 25 ng/µL as a negative control. All injected fish 4 were raised to 25 dpf in standard conditions, when behavioral recordings were carried out. 5 Brains from all fish recorded for behavior were dissected and processed for 6 immunohistochemistry in order to quantify by GFP the number of cells expressing hcrt:Gal4; 7 uas:BoTx-BLC-GFP . fish were bathed in gentamicin for 24 h. Following the treatment, a complete water change was 13 administered and behavior was again recorded for 24 h. Fish treated with gentamicin were 14 housed in separate tanks for at least 1 month after treatment in order to avoid contamination. 15 Lateral line re-growth was measured with DASPEI staining two weeks following ablation to 16 confirm that there were no long-term effects from the ablation treatments. 17

Sequence analysis 19
To compare Hcrt sequences in different species, we aligned the Hcrt protein sequences of A. Bio Rad CFX96 with a C1000 thermal cycler: 95.0ºC for 3 min followed by a plate read at 17 95.0ºC for 10 s to 53.3ºC for 30 s followed by a plate read repeated 39 times. All samples were 18 compiled into Bio Rad CFX manager gene study (version 3.1) to account for inter-run 19 calibration. All samples were normalized to one (relative to surface fish controls) Housekeeping 20 genes were validated using Bio Rad CFX manager and found M values of 0.82 for rpl13α, and 21 0.57 for gapdh, falling well within the acceptable range for quality reference genes 22 then placed in Vectashield with DAPI until mounted in 2% low melt agarose (Sigma) for imaging. 17 All samples were imaged in 2 µm sections and are presented as the Z-stack projection through 18 the entire brain. For quantification of HCRT levels, all hypothalamic slices were imaged in 2 µm 19 sections, merged into a single Z-stack using maximum intensity projections (Nikon Elements), 20 and the total brain fluorescence was determined by creating individual ROIs for all soma 21 expressing HCRT, background intensity was subtracted in order correct for non-specific 22 fluorescence. All imaging analysis was carried out using Nikon Elements (v. 4.50). 23 24

Statistics 25
Two-way ANOVA tests were carried out to test the effects of pharmacological and starvation 1 paradigms among different groups and populations on behavior. Each was modeled as a 2 function of genotype (Surface and Pachón) and genotype by treatment interaction (TCS, 3 gentamicin, or starvation, respectively). Significance for all tests was set at p<0.05. When the 4 ANOVA test detected significance, the Holm-Sidak multiple comparison post-test was carried 5 out to correct for the number of comparisons. For comparison of two baseline groups, non-6 parametric t-tests were carried out to test for significance. Each experiment was repeated 7 independently at least three times. All replicates were biological replicates run independently 8 from one another. No data was excluded, and no statistical outliers were removed. All statistical 9 analyses were carried out using SPSS (IBM, 22.0) or InStat software (GraphPad 6.0). Power 10 analyses were performed to ensure that we had sufficient N to detect significant differences at a 11 minimum of 80% power at the 0.05 threshold using Graphpad InStat.

Competing Interests 21
There are no competing interests associated with this work.