Sleep Deprivation During Early-Adult Development Results in Long-Lasting Learning Deficits in Adult Drosophila

Abstract

Study Objectives:

Multiple lines of evidence indicate that sleep is important for the developing brain, although little is known about which cellular and molecular pathways are affected. Thus, the aim of this study was to determine whether the early adult life of Drosophila, which is associated with high amounts of sleep and critical periods of brain plasticity, could be used as a model to identify developmental processes that require sleep.

Subjects:

Wild type Canton-S Drosophila melanogaster.

Design; Intervention:

Flies were sleep deprived on their first full day of adult life and allowed to recover undisturbed for at least 3 days. The animals were then tested for short-term memory and response-inhibition using aversive phototaxis suppression (APS). Components of dopamine signaling were further evaluated using mRNA profiling, immunohistochemistry, and pharmacological treatments.

Measurements and Results:

Flies exposed to acute sleep deprivation on their first day of life showed impairments in short-term memory and response inhibition that persisted for at least 6 days. These impairments in adult performance were reversed by dopamine agonists, suggesting that the deficits were a consequence of reduced dopamine signaling. However, sleep deprivation did not impact dopaminergic neurons as measured by their number or by the levels of dopamine, pale (tyrosine hydroxylase), dopadecarboxylase, and the Dopamine transporter. However, dopamine pathways were impacted as measured by increased transcript levels of the dopamine receptors D2R and dDA1. Importantly, blocking signaling through the dDA1 receptor in animals that were sleep deprived during their critical developmental window prevented subsequent adult learning impairments.

Conclusions:

These data indicate that sleep plays an important and phylogenetically conserved role in the developing brain.

Citation:

Seugnet L; Suzuki Y; Donlea JM; Gottschalk L; Shaw PJ. Sleep deprivation during early-adult development results in long-lasting learning deficits in adult drosophila. SLEEP 2011;34(2):137-146.

SLEEP AMOUNTS ARE HIGH DURING BRAIN DEVELOPMENT SUGGESTING AN IMPORTANT ROLE FOR SLEEP DURING PERIODS OF NEURONAL PLASTICITY.^1,2 Interestingly, while the effects of sleep deprivation on cognitive behavior are fully reversible in healthy adults, deficits may persist when sleep loss occurs in children,^3–6 or during critical periods of brain development.^7,8 Much of the experimental work on sleep and development has focused on the visual system.^9,10 However, evidence suggests that sleep may also be important for the development of dopamine (DA) circuits in the prefrontal cortex that are responsible for short-term memory and response inhibition.¹¹

In the developing brain, improvements in response inhibition and short-term memory are believed to parallel DA synapse formation.^12–14 Response inhibition is the ability to inhibit prepotent responses that are no longer appropriate (i.e., actions having priority over other response tendencies) and, in humans, is frequently evaluated with go–no go and stop-signal tasks.^15–17 In both rodents and non-human primates, reducing DA or blocking the D1 receptor disrupts response inhibition and short-term memory.^18,19 Similarly, children and adults with the Val¹⁵⁸Met polymorphism in catechol-O-methyltransferase exhibit both enhanced DA catabolism and deficits in response inhibition and short-term memory compared to those with an allele that is associated with higher levels of synaptic DA.^20,21 Moreover, children with sleep disordered breathing display on average decreased IQ, deficits in executive function and signs of neuronal injury that may not resolve with treatment.^3,22 Together these data suggest that sleep may play an important role in regulating the development of neuronal circuits including those involved in behavioral flexibility. Nonetheless, the mechanisms by which sleep loss during development contribute to long-lasting impairments in adults remain unknown.

To address this issue we turned to the model organism Drosophila melanogaster. We chose the fly because it has been effective in elucidating basic mechanisms of brain development, sleep and learning.^23–28 Specifically, flies, like humans, exhibit ontogenetic variations in brain plasticity^23,27,28 that are associated with increased sleep time.²⁴ In addition, sleep deprivation in mature adult flies (5-7 days old) results in impairments in short-term memory and response inhibition that are rapidly reversed following a 2-h nap indicating that there are no long-term consequences of acute sleep loss in the adult.²⁵ Here, we investigated the consequences of sleep deprivation during the fly's first full day of adult life, when sleep and plasticity are high.^26,27 We show that, in contrast to mature adults, acute sleep deprivation during this critical period of fly development is associated with long-lasting negative consequences on learning.

METHODS

Fly Stocks, Sleep Monitoring, and Sleep Deprivation

Canton-S (Cs) flies were obtained from the Bloomington Drosophila stock center. Flies were cultured at 25°C, 50% to 60% humidity, in 12 h:12 h Light-Dark cycle, on a standard food containing yeast, dark corn syrup, molasses, dextrose, and agar. Newly eclosed adult flies were collected from culture vials daily under CO₂ anesthesia. Three to five hours after collection, flies were individually placed into 65-mm glass tubes (5-mm diameter) and sleep was evaluated using the Trikinetics activity monitoring system (www.Trikinetics.com) as previously described.²⁴ Flies were sleep deprived using the SNAP (Sleep Nullifying APparatus, or SNAP), an automated sleep deprivation apparatus that has been found to keep flies awake without nonspecifically activating stress responses.²⁹ Differences in sleep time were assessed using either a Student t-test or analyses of variance (ANOVAs), which were followed by planned pair-wise comparisons with a Tukey correction.

Short-Term Memory and Response Inhibition

Aversive phototaxic suppression (APS) was performed as previously described.^25,30 Flies were individually tested in a T-maze, where they are allowed to choose between a dark and a lighted vial. Choices of the lighted vial were associated with an aversive stimulus provided by a filter paper soaked with a quinine solution placed in the vial. The test consisted of 16 trials through the maze, during which flies learn to make more frequent choices of the dark vial (photonegative choices). The performance score was the % of photonegative choices made in the last block of 4 trials. At least 10 flies were evaluated for each condition. For each experiment, learning was evaluated by the same experimenter, who was blind to genotype and condition. All flies were tested in the morning between ZT0 and ZT4. Score differences between control and experimental groups were assessed using a Student t-test. Phototaxis index: phototaxis was evaluated in the T-maze without filter paper. The average proportion of choices of the lighted vial during 10 trials was calculated for each individual fly. The phototaxis index (PI) was the average of the scores obtained for ≥ 5 flies ± SEM. Quinine/humidity sensitivity: Sensitivity to quinine/humidity was evaluated as in Seugnet et al. 2008: each fly was individually placed in a 14 cm transparent cylindrical tube covered with filter paper. The quinine/humidity sensitivity index (referred to as quinine sensitivity index or QSI) was determined by calculating the time in seconds that the fly spent on the dry side of the tube when the other side had been wetted with quinine, during a 5-min period.

Courtship

Male flies were sleep deprived starting at ZT10 (sleep deprived group) for 24 h or kept undisturbed in Trikinetics tubes (control group). After 3 days of recovery, control and sleep deprived males were exposed to virgin females and videotaped for 10 minutes under white light. The courtship index (CI) was calculated by dividing the time spent courting (the sum of the lengths of all of the courtship bouts) by the total length of the test (10 min).

Social Enrichment

Upon eclosion, male flies were housed in groups of 30/vial to ensure that early social experience was maintained across experimental groups. Flies were sleep deprived for 24 h at age-0 and age-4 days or remained unperturbed to serve as untreated controls. Sleep deprivation was conducted while flies resided in vials in groups of 30 as previously described.³¹ At age 6 days, male flies were divided into a socially isolated group, in which individuals were housed in 65-mm glass tubes, and a socially enriched group, consisting of 35-40 male flies housed in a single vial. After 5 days of social enrichment/isolation, flies were placed into clean 65-mm glass tubes, and sleep was recorded for 3 days, using the Trikinetics activity monitoring system.³² Mean Δ Daytime Sleep was generated by quantifying daytime sleep for socially isolated and socially enriched flies for 3 full days. The mean daytime sleep for socially isolated siblings was subtracted from the amount of daytime sleep for each individual socially enriched fly for each of 3 days.³³ The Difference in Δ Daytime Sleep between experimental groups was evaluated using a one-way ANOVA.

Drug Sensitivity

All drugs were suspended in melted (40°C) vehicle. To test for D1-Like receptor sensitivity, flies sleep deprived on eclosion day were allowed to recover for 3 days before being transferred to new tubes containing either vehicle alone (1% agar 1% sucrose), L-DOPA (1 mg/mL), or the D1 agonist SKF-82958 (0.5 mg/mL). These doses of L-DOPA and SKF-82958 do not maximally increase wake time, allowing us to detect a graded response. Flies were transferred to drug between zeitgeber time ZT0-2, and sleep was evaluated for 24 h. For learning, flies were fed L-DOPA 24 h before the learning test (5 mg/mL in 1%agar 1%sucrose) or the D1 agonist SKF-82958 (3 mg/mL) for 2 h before the learning test. To prevent D1-like receptor activation, flies were fed the D1 antagonist SCH 23390 (1 mg/mL) during sleep deprivation. The doses of L-DOPA, SKF-82958, and SCH 23390 were selected for the learning studies based upon previous reports.²⁵

Whole Brain Immunohistochemistry

Fly brains were dissected in cold PBS, fixed for 20 minutes in a 4% paraformaldehyde phosphate buffered saline (PBS), solution, washed in 0.3% triton X100 PBS (PBS-T), and blocked for ≥ 45 min in 3% Goat Serum PBS-T. Mouse anti-Tyrosine Hydroxylase (Immunostar) antibodies were used at 1:50 and detected with Alexa-Fluor 488 goat anti-mouse (molecular probes). Brains were mounted in hard-set Vectashield and imaged using a Fluoview confocal microscope (Olympus).

QPCR

Total RNA was isolated from groups of 20 fly heads and processed for cDNA synthesis and QPCR as previously described.^25,29 Flies were frozen at lights-on (08:00, zeitgeber time ZT-0). Expression values for RP49 were used to normalize results between groups.

Evaluation of DA levels by HPLC

For each condition, 4-5 independent sets of 20 flies were frozen at −80°C at lights on (08:00, zeitgeber time ZT-0). Heads were then separated from the frozen bodies. Samples were sent to Dr. Raymond F Johnson, Neurochemistry Core Lab, Nashville, for HPLC analyses.

RESULTS

24 h of Sleep Deprivation in Adult (Age 6 days) and Newly Eclosed Flies

In young mammals, sleep disruption may result in long-lasting cognitive impairments. To determine whether this may be true in flies, we evaluated short-term memory and response inhibition in female adult (age 6 days) and newly eclosed (age 0 days) wild-type Cs flies, using aversive phototaxic suppression (APS). In this task, flies learn to avoid light that is paired with an aversive stimulus (quinine/humidity); a high score indicates learning. Both adult and newly eclosed flies were deprived of sleep for 24 h and allowed to recover unperturbed for 3 days. As previously reported,^34,35 adult Cs flies display a robust increase in sleep above untreated controls that persists for 3 days (Figure 1A, B). However, adult female Cs flies show no evidence of long-term cognitive impairments following 3 days of recovery, as measured by performance in the APS (Figure 1C). These results are consistent with our previous observation that performance is rapidly restored by a single night of sleep following 22 h of sleep deprivation in adult flies.²⁵

Figure 1.

Persistent learning impairments. A: Six-day old female Cs flies were sleep deprived for 24 h (n = 14, white) and allowed to recover unperturbed for 3 days; comparisons are made to untreated siblings placed into Trikinetics tubes at the same age (n = 14, black). Sleep is expressed in min/hour, a 2 (Condition: Total Sleep Deprivation (TSD) vs. control) × 4 (Day) × 24 (Hour) repeated measures ANOVA revealed a significant Condition × Day × Hour interaction (F_69,1794 = 3.27, P = 3.62 E-9). B: % Sleep recovered was calculated for each individual sleep deprived fly as a ratio of the minutes of sleep gained above the group mean of the controls during the corresponding recovery period divided by the total min of sleep lost vs. the group mean for the sleep obtained by controls during the 24 h of sleep deprivation period. C: Performance in the APS was not modified in 6 day old flies that were sleep deprived for 24 h and allowed to recover for 3 days (n = 7) compared to their untreated siblings (n = 10; t-test P = 0.40, n.s. = non-significant). D: Flies aged 3-5 hours after eclosion were sleep deprived for 24 h and allowed to recover undisturbed for 3 days (n = 17) or placed into Trikinetics tubes to serve as controls (n = 20); a 2 (Condition: TSD vs control) × 4 (Day) × 24 (Hour) repeated measures ANOVA revealed a significant Condition × Day × Hour interaction (F_69,2415 = 25.03, P = 1.0 E-15). E: 24 h of sleep deprivation in 3-5 h old flies did not alter the % Sleep recovered compared to mature adults (t-test, P = 0.34, compare with B). F: Performance in the APS was significantly impaired when 24 h of sleep deprivation in 3-5 h old flies was followed by 3 days of recovery, (n = 10) compared to their untreated siblings (n = 10; t-test P = 0.04).

Newly eclosed female Cs flies also responded to 24 h of sleep deprivation with sleep rebound that persisted for 3 days (Figure 1D, E). In contrast to older flies, however, performance in the APS was significantly reduced even after 3 days of recovery (Figure 1F). Sleep deprivation during the first day of adult life did not alter the time required for the fly to complete the 16 trials (TCT), the photosensitivity index ([PI] percentage of photopositive choices in 10 trials in the absence of quinine) nor the quinine sensitivity index ([QSI] time in seconds flies reside on the non-quinine side of a chamber) when the flies became mature adults, indicating that the impairment in response inhibition and short-term memory were not due to alterations in motivation or sensory thresholds (Supplemental Table S1). Finally, neither sleep time nor sleep architecture were significantly different from untreated siblings on the day learning was evaluated (Supplemental Figure S1).

We have previously shown that total sleep time is greater in newly eclosed Cs flies than in adults.²⁴ Indeed, 24 h of sleep deprivation in young flies results in a loss of ∼1200 min, while the same amount of sleep deprivation in adults produces a loss of ∼700 min. To determine whether the long-term learning deficits are due to the absolute amount of sleep loss and/or residual sleepiness, we conducted 2 experiments. First, we sleep deprived newly eclosed Cs flies for 12 h during their first dark period and allowed them to recover unperturbed for 6 days. We chose to extend the recovery period from 3 to 6 days to ensure that sleep had returned to baseline levels. Young flies deprived of sleep for 12 h lost 686 ± 9 min of sleep. Importantly, sleep had returned to baseline levels by the time flies were evaluated in the APS (Supplemental Figure S2). Consistent with the return of sleep to baseline levels, we found that Amylase, a known biomarker of sleepiness in flies, rapidly returned to baseline and was not statistically different from their untreated siblings (P = 0.69, n = 4 replicates; data not shown). We have previously shown that, in flies, Amylase levels are only elevated following waking conditions that are associated with increased sleep homeostasis and are not induced by stress.³⁶ Although sleep and Amylase mRNA levels did not differ from controls, flies that had been sleep deprived for only 12 h on their first day of life showed persistent learning impairments when tested 6 days later (sleep deprived = 0.35 ± 0.03 vs. untreated control = 0.48 ± 0.03; P < 0.05, Student t-test).

Secondly, we asked whether long-term learning impairments could be induced in 6-day old adult flies that were exposed to longer durations of sleep loss. Adult flies were sleep deprived for 48 h and lost 1668 ± 7 minutes of sleep. To maximize sleepiness, flies were only allowed to recover for 2 h before being tested in the APS; we have previously shown that 2 h of recovery was sufficient to restore learning in adult flies following 12 h of sleep loss.²⁵ During the 2-h recovery period, sleep deprived flies slept significantly longer than untreated controls (Figure 2A). Although the flies began to discharge homeostatic drive, they were only able to recover ∼5% of the sleep that was lost within the allotted time of recovery and thus were likely to have had high sleep drive when tested in the APS (Figure 2B). Despite the increased amount of sleep loss and the reduced opportunity to recover, adult flies achieved learning scores typically achieved by untreated flies following a night of sleep (Figure 2C, compare with controls in Figure 1C, F). These data support our previous results demonstrating that performance decrements in the APS cannot be attributed to increased sleep drive.²⁵ Importantly, these data indicate that the consequences of sleep loss in newly eclosed flies are dramatically different from those observed in the adult.

Figure 2.

48 h of sleep deprivation in 6 day old flies. A: Sleep was significantly increased between 08:00 and 10:00 (ZT0-2) compared to circadian matched untreated controls (n = 22) t-test, P = 1.76E-13. B: % Sleep recovered was calculated as a ratio of the minutes of sleep gained above baseline during the 2 h of recovery divided by the total min of sleep lost during 48 h of sleep deprivation. C: Performance in the APS was significantly impaired following 48 h of waking (n = 9) but was restored to baseline levels following a 2-h nap (n = 7), t-test, P = 0.0004. TSD refers to Total sleep deprivation.

Previous studies have shown that plasticity in the Drosophila brain persists for the first several days of adult life.^23,27,28 To determine whether persistent learning impairments might be useful for identifying a critical period of fly development, we deprived flies of sleep on the day they eclosed (Day 0), their first full day of adult life (Day 1), and on Days 2, 3, and 4. In each instance, flies were released into recovery on the evening of the following day, allowed to recover unperturbed for 3 days. Thus, flies were tested for learning on Days 5-9. As seen in Figure 3, flies sleep deprived on Day 1 and on Day 2 exhibited long-lasting learning impairments similar to those found when deprivation was conducted on Day 0.

Figure 3.

Sleep loss in young flies. Flies were deprived of sleep for 24 h on the day indicated by the x-axis, released into recovery on the evening of the following day and allowed to recover unperturbed for 3 days before being tested in the APS. Data were evaluated by one-way ANOVA followed by planned comparisons (F_5,64 = 1.7, P = 0.14; *P < 0.05).

Sleep Deprivation and Social Experience

To determine whether sleep deprivation on the day of eclosion would result in long-term modifications to other adaptive behaviors, we evaluated male courtship and whether flies would maintain their ability to increase sleep following exposure to an enriched social environment. We chose these behaviors because, like performance in the APS, courtship behaviors and the response to social enrichment are modified by dopamine signaling.^{25,32,37–39} Courtship behavior is typically quantified in male flies. Thus, we also evaluated the effects of sleep loss on both courtship and the response to an enriched social environment in male flies. As seen in Figure 4A, naïve courtship of virgin females is not modified when 6-day-old male flies are sleep deprived for 24 h and allowed to recover for 3 days unperturbed, consistent with our previous results.²⁵ However, when sleep loss occurs on the first day of life, male flies exhibit significantly reduced courtship of virgin females, even after 3 days of recovery (Figure 4B). Note that at the time courtship was examined, sleep in flies that had been sleep deprived on the day they eclosed was not statistically different from controls (Supplementary Figure S3).

Figure 4.

Persistent changes in adaptive behavior. A: No change in courtship behavior is observed in 6-day-old male flies that were sleep deprived for 24 h (n = 11) and allowed to recover unperturbed for 3 days compared to age matched controls (n = 13, *t-test, P = 0.2) B: The courtship index of 3-5 h old flies sleep deprived for 24 h and allowed to recover unperturbed for 3 days (n = 17) was significantly reduced vs. untreated-control siblings (Control, n = 19). (*t-test, P = 0.03). C: Untreated male flies housed in groups of 35-40 male siblings for 5 days beginning at age 6-days (n = 64, black) display a significant increase in daytime sleep compared to their isolated siblings (Δ Daytime Sleep); 24 h of sleep deprivation at age 4-days (n = 54, white) does not alter Δ Daytime Sleep, while 24 h of sleep loss in 3-5 h old flies significantly attenuates the response to social enrichment comparisons (n = 64, gray) (One-way F_2,181 = 5.4, P = 0.005; *P < 0.05, modified Bonferroni Test). TSD refers to Total sleep deprivation.

When flies are exposed to an enriched social environment for 5 days they display an increase in sleep that is dependent upon genes involved in synaptic plasticity.^32,33 Thus, 6-day-old male flies were exposed for 5 days to either an enriched social environment consisting of 35-40 male siblings or a socially impoverished environment in which flies were individually housed in 65-mm glass tubes. As previously described, socially enriched male flies display an increase in sleep compared to their isolated siblings^32,33 (Figure 4C, black). Due to the length of the enrichment protocol, we chose to evaluate the effects of sleep loss in the mature adult at Day 4. This is the earliest age where no long-term deficits in APS are observed following sleep deprivation (Figure 3). Indeed, no change in the response to social enrichment was observed when 4-day-old male flies were sleep deprived for 24 h and exposed to social isolation or enrichment (Figure 4C, white). We then deprived flies of sleep for 24 h on the day that they eclosed and allowed them to recover unperturbed before being placed into either social enrichment or isolation when they were 6 days old; sleep was not statistically different between controls and sleep deprived flies at this age (total sleep: 924 ± 34 min vs. 972 ± 48 min, t-test, P = 0.4) As seen in Figure 4C (gray), 24 h of sleep deprivation on the day of eclosion significantly attenuated the increase in sleep that occurs following 5 days of social enrichment. Together, these data suggest that sleep deprivation during a specific stage of development may result in long-term changes in adaptive behaviors that are responsive to dopamine signaling.

Sleep Deprivation and Dopamine

If the learning deficits and changes in sexual behavior are due to changes in dopamine signaling, then flies sleep deprived on their first day of life should show a reduced response to the wake-promoting properties of dopamine agonists. As seen in Figure 5A, when flies that were sleep deprived on their first day of life were fed L-DOPA (3,4-dihydroxyphénylalanine) after 3 days of recovery, sleep was only modestly affected in comparison to untreated siblings. In contrast, sleep was not significantly altered following L-DOPA administration after 3 days of recovery when flies were sleep deprived at 6 days of age (Figure 5A). Similar but more modest results were found when flies were administered the D1 agonist SKF 82958: flies sleep deprived on eclosion day lost −38% ± 5% n = 35 of baseline sleep upon feeding 0.5 mg/ml of the agonist while controls lost −64% ± 5% n = 40 of baseline sleep, P < 0.005, Student t-test. These results suggest that dopamine signaling through a D1-like receptor is defective for at least several days after sleep deprivation. An alternative interpretation is that flies that are sleep deprived on the first day of life are more sleepy and thus the dopamine agonists are not strong enough to overcome their sleep drive. Given that the sleep rebound between young and mature flies was similar, and that co-administration of the D1 antagonist during sleep loss can prevent the negative effects of sleep deprivation on learning (see below) we favor the interpretation that signaling through a D1-like receptor is adversely affected.

Figure 5.

Long term changes in DA signaling. Young (3-5 h) and mature (6-day-old) flies were sleep deprived for 24 h, allowed to recover for 3 days, and compared to untreated control siblings. A: % nighttime sleep lost in flies fed L-DOPA (1 mg/mL, n > 26) 3 days after being sleep deprived for 24 h at day 0 (left) or at day 6 (right) compared to their untreated siblings. A 2 (Young vs. Mature) × 2 (Control vs. Sleep Deprived) ANOVA revealed an Age × Condition interaction (F_1,103 = 7.3, P = 0.007; *P < 0.05 modified Bonferroni correction). B: number of DA neurons per cluster as evaluated using anti-Tyrosine Hydroxylase whole brain immunostaining. At least 12 clusters were analyzed for each condition. C: representative anti-Tyrosine Hydroxylase whole brain immunostaining showing the DA neuron clusters in an untreated control (left) and in a sleep deprived fly (right). The inset show an enlarged view of the PPM1,2 cluster, with individual cell bodies numbered. D: expression of the DA receptors dDA1 and D2R is increased 3 days after sleep deprivation in contrast to other DA related genes: pale (tyrosine hydroxylase), Dat (dopamine transporter), Damb (dopamine receptor in mushroom bodies). Transcript levels are presented as the difference in expression between control and sleep deprived divided by the expression value in control flies (% change). QPCR data using 20 whole heads for each condition (n = 3 replicates of 20 heads); heads were collected on recovery day 4 (for each gene, % change in sleep deprived flies is compared to untreated age-matched siblings using a one sample t-test; P values are indicated above each bar, *P < 0.05).

Sleep deprivation during development has been associated with an increase in programmed cell death in the brain.⁴⁰ Thus to determine whether sleep deprivation on the first day of life is deleterious to dopaminergic neurons or to postsynaptic dopamine receptor signaling we evaluated the number of dopaminergic neurons. As shown in Figure 5B-C, brains of flies sleep deprived at eclosion present the normal complement of dopaminergic neurons in all six clusters examined at the time of test. No obvious morphological abnormalities were detected in these neurons. To further evaluate the integrity of dopaminergic neurons, we quantified whole head dopamine levels using HPLC. After 3 days of recovery, flies that had been sleep deprived on their first day of life showed similar amounts of dopamine (63 ± 9 pg/head; n = 4 replicates of 20 heads) compared to their untreated siblings (57 ± 6 pg/head; n = 5 replicates of 20 heads; P = 0.41 Student t-test). No significant differences in DA levels were found when examining brains using HPLC (Data not shown). Moreover, genes expressed in dopaminergic neurons that are associated with the synthesis and handling of dopamine such as pale (coding for tyrosine hydroxylase), dopadecarboxylase (Ddc), and the Dopamine transporter (DAT) are transcribed at similar levels in previously sleep deprived and control flies (Figure 5D). In contrast, the D1-like Dopamine receptor dDA1, and the D2-like receptor D2R are expressed at significantly higher levels for several days after flies were allowed to recover from sleep deprivation on their first day of life (Figure 5D). Figure 5D shows that sleep deprivation at 6 days of age did not induce a long term increase in dDA1 receptor mRNA levels. We have previously shown that dDA1 mRNA is down regulated following sleep deprivation in mature adults and that the dDA1 receptor is required for learning in this APS paradigm.²⁵ The D1-like dopamine receptor in Mushroom Bodies (DAMB) and D2R do not play a role in the APS²⁵ and thus were not investigated further. Whether D2R plays a role in the other phenotypic modifications that we have seen, such as courtship or social enrichment will require further investigation. Together these data are reminiscent of the differential neuronal programming found between neonatal and adult mammals exposed to dopamine⁴¹ and suggest that flies sleep deprived at eclosion have acquired a long term deficit in dDA1 signaling.

Pharmacological Rescue of Adult Learning Impairments

If sleep deprivation at the time of eclosion results in deficits in dopamine signaling later in adulthood, then it should be possible to reverse learning impairments in adults by acutely administering a dopamine agonist. As seen in Figure 6A, flies that had been sleep deprived on their first day of life and fed L-DOPA prior to the evaluation of learning 3 days later displayed normal performance. Similarly, the D1 agonist SKF82958 is also effective in restoring learning in the previously sleep deprived adult when administered for 2 hours before the learning test (Figure 6A-B). SKF82958 can activate Drosophila D1 receptors,⁴² and its specificity has been confirmed genetically.²⁵ Thus the learning impairments can be rescued in adults by activation of the D1-like receptors.

Figure 6.

Rescue and prevention of learning impairments. A-C: Bottom, sleep deprivation and test schedule; top, performance. 3-5 hour old flies were sleep deprived (TSD: Total sleep deprivation) and allowed to recover for 3 days before being tested for learning. A-B: rescue of learning impairments using DA agonists. Flies were fed vehicle (1% agar plus 1% sucrose), L-DOPA (5 mg/mL, A) or the D1 agonist SKF82958 (3 mg/mL, B) before the learning test. C: flies were fed the D1 antagonist SCH23390 (TSD + D1 antagonist) or vehicle (TSD) during sleep deprivation, then transferred to regular food until being tested for learning. At least 10 flies were tested for each condition. (*t-test, P < 0.05)

Pharmacological Prevention of Persistent Learning Deficits

If the activation of the D1-like receptors during waking in young flies results in a change in receptor signaling, then it should be possible to prevent learning deficits by blocking the activation of the D1-like receptor during sleep deprivation on the day of eclosion. Thus, on the day flies eclosed, they were collected, placed into tubes with food containing the D1 antagonist SCH23390, and sleep deprived for 24 h. Following deprivation, the flies were placed back onto normal food and allowed to recover unperturbed for 3 days. As seen in Figure 6C, administering the D1 antagonist only during sleep deprivation on the first day of life completely blocked the adult learning deficits. In contrast, vehicle fed controls were learning impaired when tested as adults (Figure 6C, black). Since SCH23390 has been shown to block Drosophila D1-like receptors,⁴² this result indicates that blocking D1-like receptor activation during sleep deprivation is sufficient to prevent learning impairments.

DISCUSSION

Sleep amounts are highest during postnatal periods associated with rapid brain development. This observation led to the hypothesis that sleep plays a crucial role in neonatal plasticity.¹ Subsequent results obtained from sleep depriving young animals during critical periods of brain development have generally supported a role for sleep in the plasticity of the developing brain.^9,10 For example, REM sleep deprivation early in life, but not during adolescence, results in long-lasting deficits in the stability of hippocampal circuits in rodents.^7,43 Similarly, the synaptic remodeling of the visual cortex induced by monocular deprivation is blocked by acute sleep deprivation in kittens.² Interestingly, flies also display a critical period of brain plasticity that is modified by visual input,^23,27,28 and that is associated with increased amounts of sleep²⁴ (supplemental Figure S4). Using a learning task that involves a visual component, we report that preventing flies from sleeping on their first full day of life, a critical period during which brain plasticity is high, results in long-lasting cognitive deficits that persist into maturity. Together these data support an important and phylogenetically conserved role for sleep in brain development and neuronal plasticity.

It should be noted that a question common to all sleep deprivation studies is whether an observed outcome is due to sleep loss or confounding variables related to the methods used to keep the animals awake.⁴⁴ Since both sleep duration and sleep consolidation are high throughout the entire first day of development, it is impossible to expose flies to a control stimulus without also disrupting sleep. Moreover, most stressors disrupt sleep making it particularly difficult to disentangle the unique effects of sleep loss or stress. As a consequence, we cannot formally exclude the possibility that the learning deficits observed in the adult are due to confounding variables associated with the apparatus. However, we believe that it is unlikely that the learning deficits observed in adults are solely due to stress during development. First, we have previously shown that learning deficits in the APS are a consequence of sleep loss and that neither mechanical perturbations nor stress in adult flies result in even transient changes in performance.^25,45 Second, the sleep nullifying apparatus used to keep animals awake does not appear to activate stress-response genes such as those involved in metabolic stress (SNF4a, Hif1), chemical stress (mpk2), and humoral stress (turandot).²⁹ Third, we have shown that some stressors, such as heat stress, protect flies from the negative effects of sleep deprivation indicating that some stressors may be beneficial.²⁹ Finally, we have shown that flies that spontaneously exhibit fragmented sleep throughout their lives, including during development, are learning impaired in the absence of any stress inducing experimental intervention.²⁵ Nonetheless, given the difficulty in excluding the nonspecific role of the intervention used to keep the animals awake,¹⁰ one must use caution when interpreting results from all studies that disrupt sleep during development.

Our data demonstrate that sleep loss has different outcomes when it occurs during the first full day of life from outcomes when sleep loss occurs in the mature adult. For example, sleep loss in the mature adult results in a down-regulation in D1-like receptor mRNA and learning impairments that are rapidly reversed by a short nap (2 h). In contrast, sleep loss that occurs on the first full day of life results in an upregulation of both D1-like and D2-like receptors, a reduced response to the wake-promoting properties of dopamine agonists, and long-term learning impairments that persist for at least 6 days. At first glance, the changes in learning may appear to be minor (∼25% reduction); however, the changes in performance are well within the range of effect sizes observed following sleep loss in humans and rodents across a number of cognitive domains.^46–49 In addition to changes in short-term memory and response inhibition, sleep deprivation on the first day of life results in long-term reductions in naïve male courtship, while sleep loss in mature adults does not adversely affect courtship. Moreover, sleep loss on the day of eclosion also attenuates the increase in sleep that is normally observed following social enrichment. These later results suggest that sleep deprivation early in development may have a broader impact on the brain than is measured by learning alone.

What are the causes underlying the long term learning deficits induced by early sleep deprivation? Interestingly, sleep deprivation on the first day of adult life did not result in long-term changes in sleep time, sleep architecture or locomotor activity. Thus, it is unlikely that the learning deficits observed later in adulthood were due to alterations in baseline sleep. Similarly, we did not detect changes in either the number of dopaminergic neurons or genes involved in dopamine synthesis and handling, suggesting that the primary target of early sleep loss may not be the dopaminergic neurons themselves. However, expression of the D1-like receptor dDA1, which is required for APS, is significantly increased after early sleep deprivation, whereas it is reduced when learning is impaired following sleep deprivation in the adult.²⁵ These results are reminiscent of the differential neuronal programming reported between neonatal and adult mammals exposed to dopamine.⁴¹ Neuronal programming refers to the observation that interventions that alter neurotransmitter signaling during critical periods of development can result in permanent changes in the structure and function of these neurotransmitter systems throughout adult life.⁵⁰ For example, prenatal exposure to cocaine results in permanent changes in aspects of DA signaling, including a reduction in second messenger coupling.^51,52 Similarly, long-term changes in the number and sensitivity of dopamine receptors are observed following the administration of haloperidol⁵³ or D1 and D2 receptor antagonists during critical periods of development.^54,55 Together, these data suggested to us that early sleep loss may result in lower dopamine signaling later in adulthood.

To test this hypothesis, we administered dopamine agonists to adult flies with learning deficits and were able to rescue short-term memory and response inhibition. In addition, we blocked D1-receptor activation during sleep deprivation to determine whether it would be possible to prevent changes in dopamine receptor sensitivity later in adulthood. Blocking D1-receptors during this critical window completely prevented long-term learning deficits when the flies were tested later as adults. Interestingly, cocaine treated animals have lower D1 dependent G-protein signaling activity and this deficit can be blocked by co-treating the animals with the D1 antagonist SCH23390.^56,57 Together these observations suggest the possibility that sleep loss in the developing brain may disassociate dopamine receptor levels from their functionality. However, and in contrast to studies in which DA agents are administered prenatally, we did not observe changes in the levels of the endogenous ligand as measured by whole-brain DA levels. One explanation might be that sleep deprivation alters DA release in specific circuits and that these local changes would not be sufficient to alter whole-brain DA levels. Alternatively, the plasticity of D1 signaling during this critical period may be downstream of receptor activation. Clearly, dopamine signaling is complex and thus future work will be necessary to evaluate alternative explanations and to identify the precise underlying mechanisms that may mediate the effect of sleep deprivation in this context.

Previous studies have shown that Drosophila exhibit a peak of neuronal plasticity that occurs in the first few days of life.^23,27,28 For example, eliminating visual stimuli in young animals alters the structure of several neuropils involved in visual processing including the lamina, medulla, and lobula plate.^26–28 Importantly, the plastic changes that are observed by modifying visual stimuli are restricted to the first 4-5 days after eclosion, suggesting the existence of a critical developmental time window for plasticity.²⁷ However, a critical period for brain plasticity may not exist for all circuits that are modified by changes in environmental or social interactions. For example, the size of the mushroom body calyces can be modified by social experience 16 days after eclosion.⁵⁸ In this regard, it is worth noting that the mushroom bodies are not involved in changes in sleep following social enrichment.³³ In any event, the existence of a critical period for some, but not all, experience-dependent stimuli suggests that plasticity may be controlled by separate mechanisms. In the early visual system, a critical period may be helpful to optimize the ability of the brain to receive and process visual stimuli. On the other hand, plastic events that can encode memories throughout adult life appear to exhibit experience-dependent morphological changes for a much longer period of time. This kind of temporal division can also be observed in mammalian models. For example, while ocular dominance plasticity in the visual cortex can be robustly observed during a critical period early in life,⁵⁹ structural changes in other brain regions including the hippocampus can be observed much later.⁶⁰ (For a more thorough discussion of the role of environmental or social interactions on brain plasticity in vertebrate and invertebrate models see Donlea.⁶¹)

In conclusion, the data reported here demonstrate that disrupting sleep during a critical window of plasticity results in long-lasting changes in short-term memory and behavioral flexibility and that these deficits persist well into adulthood. Together these results indicate that the power of Drosophila genetics may open new ways of investigating the role of sleep during periods of high neuronal plasticity.

DISCLOSURE STATEMENT

This was not an industry supported study. The authors have indicated no financial conflicts of interest.

Table S1.

Control metrics in sleep deprived and control flies

Condition	TCT	PI	QSI
	mean ± SEM	mean ± SEM	mean ± SEM
Control	13.7 ± 0.5	78% ± 6%	233 ± 30
Sleep deprived age 0-days, 6 days recovery	15.0 ± 0.7	82% ± 7%	258 ± 14
Sleep deprived age 0-days, 3 days recovery	14.3 ± 1.0	86% ± 4%	265 ± 11
Sleep deprived age 0-days with D1 antagonist treatment, 3 days recovery	12.3 ± 1.1	90% ± 3%	269 ± 11

TCT refers to Time to complete test; PI, Phototaxic index (n = 5); QSI, quinine sensitivity index (n = 5). None of the control metrics were statistically different from the untreated control values.

Figure S1

Sleep time was not significantly different between flies sleep deprived of sleep for 24 h on Day 0 and untreated siblings on the day of testing in the APS. A: Female flies aged 3-5 hours after eclosion were sleep deprived for 24 h and allowed to recover undisturbed for 3 days until being tested for learning in the APS (Test). Box indicates total sleep deprivation (TSD). B: Daily sleep in min/hour at age 4-days in sleep deprived flies (n = 61), compared to age-matched untreated controls (n = 60). Black box indicates lights off. No differences were observed for C, total daily sleep (P = 0.43 student t-test); D, average sleep bout duration during the day (lights on) (P = 0.40 student t-test); E, average sleep bout duration during the night (lights off). (P = 0.24 student t-test) Error bars represent s.e.m.

Figure S2

Sleep time was not significantly different between flies sleep deprived of sleep for 12 h on Day 0 and untreated siblings on the day of testing in the APS. A: Flies aged 3-5 hours after eclosion were sleep deprived for 12 h and allowed to recover undisturbed for 6 days until being tested for learning (Test). Box indicates total sleep deprivation (TSD). B: Daily sleep in min/hour at age 7-days in sleep deprived flies (n = 31), compared to untreated flies put in trikinetics tubes at the same age (n = 34). No differences were observed for C, total daily sleep (P = 0.65 student's t-test); D, average sleep bout duration during the day (lights on, n.s: non significant, P = 0.27 student's t-test); E, average sleep bout duration during the night (lights off) (P = 0.75 student's t-test). Error bars represent s.e.m.

Figure S3

Sleep time was not significantly different between male flies sleep deprived of sleep for 24 h on Day 0 and untreated siblings on the day of testing for courtship. A: Male flies aged 3-5 hours after eclosion were sleep deprived for 24 h and allowed to recover undisturbed for 3 days until being tested (Test). Box indicates total sleep deprivation (TSD). B: Daily sleep in min/hour at age 4-days in sleep deprived flies (n = 27), compared to untreated flies put in Trikinetics tubes at the same age (n = 22). Black box indicates lights off. No differences were observed for C, total daily sleep (P = 0.25 student's t-test); D, average sleep bout duration during the day (lights on) (P = 0.54 student's t-test); E, average sleep bout duration during the night (lights off) (P = 0.86 student's t-test). Error bars represent s.e.m.

Figure S4

Ontogeny of sleep. A: Daytime sleep in Cs female flies beginning with their first full day of adult life (n = 22); repeated measures ANOVA revealed a significant main effect for Day (F_3,63 = 31.4, P = 1.5E-12); *P < 0.05, modified Bonferroni Test. B: Maximum sleep bout duration, a measure of sleep consolidation, revealed a significant main effect for Day (n = 22); repeated measures ANOVA (F_3,63 = 12.2, P = 1.98E-6); *P < 0.05, modified Bonferroni Test.