Day–night cycles and the sleep-promoting factor, Sleepless, affect stem cell activity in the Drosophila testis

Significance

Daily rhythms in light coordinate many biological functions over the 24-h day, facilitating the adaptation of organisms to the environment. A crucial task accomplished routinely by all multicellular organisms is the maintenance of tissue homeostasis through the activity of adult stem cells, which generate new cells when required. Using a classic Drosophila stem cell niche, we show that the frequency of stem cell divisions is rhythmic in a day–night cycle. In addition, our data, which show that stem cell divisions are increased in a short-sleeping mutant, suggest that mechanisms regulating sleep–wake rhythms may also influence stem cell activity. Our findings provide links between environmental cycles, behavioral states, and maintenance of tissue integrity by stem cells.

Abstract

Adult stem cells maintain tissue integrity and function by renewing cellular content of the organism through regulated mitotic divisions. Previous studies showed that stem cell activity is affected by local, systemic, and environmental cues. Here, we explore a role of environmental day–night cycles in modulating cell cycle progression in populations of adult stem cells. Using a classic stem cell system, the Drosophila spermatogonial stem cell niche, we reveal daily rhythms in division frequencies of germ-line and somatic stem cells that act cooperatively to produce male gametes. We also examine whether behavioral sleep–wake cycles, which are driven by the environmental day–night cycles, regulate stem cell function. We find that flies lacking the sleep-promoting factor Sleepless, which maintains normal sleep in Drosophila, have increased germ-line stem cell (GSC) division rates, and this effect is mediated, in part, through a GABAergic signaling pathway. We suggest that alterations in sleep can influence the daily dynamics of GSC divisions.

In a constantly changing environment, the transitions between day and night represent the sole nearly invariable component of our living habitat and have influenced the evolution of life since its origin (1). Environmental rhythms in light, either directly or through endogenous circadian clocks, coordinate many biological functions, including gene expression, enzyme activity, and various cellular and behavioral processes (2). One of the most pronounced 24-h cycles, the rhythm in sleep and wakefulness, reflects neuronal activity in the brain and also affects the organism’s systemic milieu, evoking physiological responses in many peripheral tissues (3). In this study, we address possible roles of environmental rhythms and the behavioral status of the organism in regulating cellular homeostasis in adult tissue.

Cell renewal is essential for maintaining tissue function over the lifetime of the organism (4). To replace damaged and aging cellular content, comparatively small populations of adult stem cells reside within specialized microenvironments called stem cell niches, where they undergo repeated mitotic divisions, generating new stem and differentiated cells (5). Although the rate of stem cell division is sustained at a tissue and cell type-specific level, an abrupt decrease in the number of cells, caused by physical injury or massive cell death, can initiate a quick response within stem cell populations, leading to increased division activity and thus, enhanced cell production (6). Behavioral factors and environmental influences are also known to affect the frequency of cell divisions, adjusting tissue homeostasis to the changing conditions inside and outside the organism (7). Twenty-four–hour rhythms of division have been documented in several mammalian tissues, reflecting dynamic activity in complex populations of dividing cells consisting of stem cells and their differentiating progeny (8–15). Altered behavioral states, such as insufficient sleep, pregnancy, and physical exercise, induce mitotic activity in the brain and other stem cell-supported tissues (4, 7). However, in most of these studies, stem cells, in particular, were not singled out for analysis, leaving open the question of whether these master regulators are subject to environmental and behavioral influences. To assess this question, we take advantage of a classic Drosophila stem cell system, the spermatogonial stem cell niche, which supports two easily identifiable and thoroughly characterized populations of adult stem cells (16).

Located in the apical part of the testis, this niche is formed by a tight cluster of somatic support cells, called the hub, surrounded by germ-line stem cells (GSCs) and cyst progenitor cells (CPCs) (Fig. 1A). Hub cells produce short-range signaling molecules that activate downstream signaling pathways specifically in adjacent stem cells but not in more remotely located cells, thereby confining the stem cell domain in the testis (17). Thus, in contrast to more complex stem cell niches that maintain mammalian tissues, both GSCs and CPCs can be unequivocally identified in vivo and in fixed whole-mount testes based on their position next to the hub and by using simple lineage-specific markers (17) (Fig. 1). GSCs and CPCs self-renew while also generating differentiating cell progeny: gonialblasts and somatic cyst cells, respectively. Gonialblasts enclosed by cyst cell pairs form cysts, the units of sperm development (Fig. 1A). Within cysts, early germ cells continue to amplify, producing 2–16 cell clusters called spermatogonial cells before undergoing meiosis and maturation (16). Molecular communication between the germ line and somatic cell partners ensures proper progression through spermatogenesis (18).

Fig. 1.

The spermatogonial stem cell niche and early stages of spermatogenesis in the fly testis. (A) Schematic representation of the niche composed of somatic hub cells (light blue), 5–10 GSCs (peach), and approximately two times as many CPCs (blue). GSCs and CPCs produce differentiating gonialblasts (GBs; yellow) and cyst cells (gray) that form cysts. GBs undergo four additional mitotic divisions to become spermatogonial cells before proceeding into meiosis. (B and C) Single confocal sections show (B, arrowheads) duplicating GSCs and (C, arrow) CPCs. Hub cells are labeled with antibodies against protein Armadillo (membrane green; asterisks). GSCs are positive for a germ cell-specific marker Vasa (red) and positioned immediately next to the hub. CPCs contact hub cells and are Vasa-negative. Antibodies against the phosphorylated form of Histone H3 label mitotic chromatin (nuclear green). DNA counterstain is in blue.

Here, we show that, in the presence of environmental day–night cycles, both stem cell populations occupying the Drosophila spermatogonial stem cell niche show daily rhythms in division frequencies that do not persist in constant darkness and thus, do not seem to constitute free-running circadian rhythms. Because sleep–wake rhythms can be driven by environmental cycles, we further address the effect of sleep duration on stem cell activity. Using a combination of genetic and pharmacological assays, we find that loss of the sleep-promoting factor Sleepless (SSS) stimulates GSC division rates in the testis. At least some of the effects of SSS on the GSCs are mediated by reduced GABA levels, which also contribute to the short sleep phenotype. Based on these results, we suggest that some sleep-regulating pathways influence the rate of stem cell division in the fly.

Results

Diurnal Rhythms of Mitotic Activity in the Testis Niche.

To examine daily division dynamics in populations of male GSCs and CPCs, we dissected testes at 4-h time intervals throughout a complete 12:12-h light–dark (LD) cycle beginning at the transition from dark to light or Zeitgeber time 0 (ZT0). To identify relevant cell types in the fixed tissue, we used a set of molecular markers for labeling hub and germ cells: anti-Armadillo (Fig. 1 B and C, membrane green indicated by the asterisk) and anti-Vasa (Fig. 2, red) antibodies, respectively. In contrast to differentiating spermatogonial cysts, both stem cell types in the testis are physically attached to the hub (17); therefore, we were able to easily locate GSCs and CPCs based on their contact with hub cells and the presence of Vasa epitopes in the GSC but not the CPC lineage. In addition, the identification of CPCs was aided by the fact that somatic stem cells are the only dividing somatic cells in the apical testis, because their differentiating progeny do not proliferate. To detect stem cells undergoing division, we used an additional marker that labels mitotic chromatin, antibodies against the phosphorylated form of histone H3 (Fig. 1 B and C, nuclear green), and carefully counted dividing GSCs and CPCs in each testis.

Fig. 2.

Rhythms of stem cell division in the testis. (A and B) Stem cell division dynamics over the course of a 12:12-h LD cycle. (A) GSCs and (B) CPCs show diurnal rhythms in mitotic activity with increased frequencies around the day–night transition. ZT is defined by the environmental signal, with ZT0 indicating lights on and ZT12 indicating lights off. The division rates at ZT0 are plotted two times to better illustrate the rhythm. Significant differences are detected between time points ZT0 and ZT12 for both GSCs and CPCs (Table S1). *P < 0.05. (C and D) Stem cell division rhythms do not persist in constant darkness. Circadian time corresponds to ZT points of the preceding LD cycles. (C) Under constant conditions, GSC division rates fluctuate but do not show a 24-h rhythm (Table S2). (D) The frequency of CPC divisions does not change under DD conditions (Table S2). Error bars represent SEMs.

We found that mitotic frequencies in both stem cell populations change over the course of a day–night cycle, with a broad peak at dusk (ZT12) and decreased rates around dawn (ZT0) (Fig. 2 A and B and Table S1). The differences between the minimum (at ZT0) and maximum (at ZT12) division rates are 1.3- and 1.4-fold for GSCs and CPCs, respectively. We also compared the number of stem cells undergoing DNA replication at ZT0 and ZT12 by incubating testes with the thymidine analog BrdU. We did not detect differences in the rate of BrdU labeling between these time points: the average number of BrdU-positive GSCs at both time points was 0.9 ± 0.03 and average numbers of labeled CPCs were 2.8 ± 0.02 and 2.9 ± 0.02, respectively (N ∼ 300). These data suggest that environmental rhythms specifically influence the mitotic transition of the stem cell cycle.

Rhythmic activity in the testis niche suggested that the fly circadian clock might be involved in regulating stem cell divisions. Although synchronized to the periodic environment, rhythms controlled by the circadian clock can persist for prolonged periods of time, even when the cycling of light or other environmental signals is abolished (i.e., in a constant environment) (19). In contrast, those biological cycles that are not supported by the endogenous clock mechanism dampen rapidly in the dark [12:12-h dark–dark (DD) cycle]. To test if the cycles in stem cell activity persist in DD, we transferred male flies from an LD cycle into the dark and kept them under a DD regimen for 3 consecutive days before examining division dynamics in the testis. As shown in Fig. 2C and Table S2, GSC divisions continued to fluctuate in the dark but did not exhibit 24-h rhythms (compare with Fig. 2A). Likewise, CPC activity was not rhythmic in DD (Fig. 2D and Table S2). We also noticed that overall levels of GSC and CPC divisions were lower in the dark compared with the rates under the LD regime. The increased activity in LD could have arisen from a slight elevation in temperature that typically occurs with light and has been associated with higher stem cell division (20). Alternatively, light itself may have oxidizing effects (21), which increase stem cell activity (Discussion). Thus, our data indicate that rhythms of stem cell division in the testis are sustained in the presence of LD cycles but do not free run in constant darkness.

Sleep-Promoting Factor SSS Affects Mitotic Activity of Male GSCs.

Daily dynamics of GSC and CPC division rates suggest that stem cell activity in the testis is differentially affected during the day and at night. This rhythm of stem cell division could be generated by a direct effect of light, or it could depend on other physiological processes that are rhythmic in day–night cycles. For instance, the sleep–wake cycle, typically regulated by the circadian clock, is driven to be rhythmic by LD cycles, even in clockless animals (22). Fruit flies awaken shortly before dawn and become engaged in various behaviors throughout the day. After a typical evening burst in activity, the entire fly population transitions to sleep for the night. Because sleep cycles influence not only brain function but also many peripheral organs (3), we sought to determine if sleep regulates stem cell activity.

We first assayed the effects of acute sleep deprivation on stem cell divisions. We kept flies awake at night through mechanical stimulation, which has been previously described (23), and measured cell divisions after this treatment. GSC activity increased after 24-h stimulation; comparison of 0.09 ± 0.02 (n = 159) with 0.16 ± 0.03 (n = 186) dividing GSCs per testis in nondeprived and sleep-deprived males, respectively, suggests that reduced sleep may affect stem cell activity. Because it is difficult to achieve chronic deprivation (over multiple nights) with mechanical stimuli, we next assayed genetic mutants with decreased sleep duration. We determined the frequency of GSC divisions in four mutant or transgenic fly lines that have reduced sleep, including flies containing a P-element insertion in the sleepless gene (sss^P1) (23), a mutation in the dopamine transporter gene DAT^fmn (24), and a deficiency in the shaker gene locus [sh(Df)] (25) and flies carrying Gal4/UAS transgenic elements for ectopic expression of a constitutively active form of protein kinase A (PKA) in neurons (26). Only sss^P1 flies showed a significant change in GSC division rates, specifically an increase compared with control (Fig. 3A, compare 3 with 3′ and Table S3). Elevated levels of GSC activity were also detected in both control and experimental groups of flies that carried the transgene for constitutive PKA production (Fig. 3A, compare 4 with 4′). This increase in stem cell frequency is likely because of an unexpected side effect of the transgenic insertion in the genome (all fly lines used for this experiment were maintained in the same genetic background) and could have masked a potential sleep-related change in stem cell divisions.

Fig. 3.

SSS modulates GSC activity in the testis niche. (A) Sleep duration was genetically reduced in males carrying a deficiency in the (1) shaker gene locus [sh(Df)], loss-of-function mutations in the (2) DAT (DAT^fum) and (3) sss (sss^P1) genes, and (4) using the transgenes for drug-induced PKA expression in the brain (PKA*). Compared with corresponding control groups of flies (iso31 males in 1′, 2′, and 3′ and drug-free transgenic PKA* flies in 4′), GSC division rates are increased in sss^P1 males but not other short-sleeping mutants. **P < 0.005. (B) Numbers of BrdU-positive cells are increased in sss^P1 mutants in both GSC and CPC populations. **P < 0.005. (C) Neuronal rescue of SSS function restores WT GSC activity in sss^P1 flies. GSC division rates in heterozygous sss^P1 mutants and rescued flies, elav-Gal4/Y; sss^P1/+ and elav-Gal4/Y; sss^P1; UAS-sss, respectively, that have normal sleep are lower than the rates in short-sleeping homozygous sss^P1 males (elav-Gal4/Y; sss^P1). *P < 0.05; **P < 0.005. (D) LD time course of GSC division rates in sss^P1 mutants (solid blue) and WT control males (striped blue). Error bars represent SEMs.

To further examine the effect of the sss^P1 mutation on GSC activity, we labeled replicating cells in the testes with BrdU. This method also indicated an increase in the number of cycling GSCs in sss^P1 mutants: 1.4 ± 0.03 BrdU-positive GSCs were detected in sss^P1 mutants compared with 0.7 ± 0.02 in iso31 flies (χ² = 41.3; P = 0; degree of freedom = 1) (Fig. 3B). The total number of GSCs in sss^P1 flies was not different from the number of GSCs in WT testes (5.5 ± 1.2 vs. 5.4 ± 1.3, respectively). Interestingly, CPC divisions also showed an increase in sss mutants: compare 1.2 ± 0.03 (n = 296) with 1.8 ± 0.03 (n = 198) BrdU-positive CPCs per testis in iso31 and sss^P1 males, respectively (χ² = 10; P = 0.001; degree of freedom = 1) (Fig. 3B).

Next, we verified that the increased division rates in the sss^P1 flies were caused by the sss^P1 insertion and not background mutations. Thus, we introduced a WT sss transgene into these flies that effectively rescues the sleep phenotype (27); sss^P1 flies carrying the sss transgene exhibited WT GSC activity (Fig. 3C and Table S4), indicating that the effect on GSC divisions was caused by SSS function. Importantly, the rescuing transgene is expressed in neurons using the promotor of the embryonic lethal abnormal vision gene (elav), which suggests that the effect of SSS on GSC divisions is not cell-autonomous. Given that sleep is also restored in these transgenic flies, these data support the idea that behavioral state is related to stem cell activity in the testis.

The daily pattern of division frequencies in sss^P1 flies indicates that the number of GSC divisions is increased at all time points in the cycle (Fig. 3D and Table S5). Interestingly, rhythmic variations in GSC activity are abolished in these mutants compared with iso31 background control flies. The apparent loss of rhythms may be caused by the overall increased rate of divisions. The other mutants tested here do not have sleep phenotypes as extreme as the reduction of sleep in sss mutants (23–26), which may account for the maintenance of normal stem cell activity in these flies. However, it is possible that some sleep-regulating pathways also regulate stem cell activity, whereas others do not.

Decrease in GABA Signaling Mediates GSC Activation in sss^P1 Mutants.

We next sought to address the mechanism by which sss affects stem cell divisions. We recently found that sss^P1 flies show reduced levels of gamma-aminobutyric acid (GABA) in the brain, which are caused by increased levels of the enzyme that breaks down GABA: GABA transaminase (GABAT) (28). Interestingly, GABAergic signaling in the brain has previously been implicated in the regulation of sleep in both flies and mammals (29). Decreasing the activity of GABAT using a loss-of-function mutation in the gabat gene gabat^PL00338 (gabat^PL) suppresses the sleep phenotype of sss mutants, indicating that GABAT is relevant for this phenotype (28). We sought to determine if the increase in GSCs division rates in sss^P1 males was also attributable to changes in GABA signaling. Thus, we increased GABA levels in sss mutants either genetically by introducing a gabat mutation or pharmacologically by treating the flies with an agonist of a GABA_A receptor (isoguvacine). As expected, the frequency of GSC divisions was significantly higher in sss^P1 mutants compared with control WT flies and isoguvacine-fed WT males, whereas gabat^PL mutants showed normal division activity (Fig. 4 and Table S6). Notably, GSC divisions were partially down-regulated in sss^P1 flies when we stimulated GABA_A receptors with isoguvacine or genetically decreased the expression of GABAT (Fig. 4 and Table S6). These data suggest that GSC function in the fly testis is modulated by loss of SSS, and altered GABA signaling downstream of the sss mutation mediates this regulation.

Fig. 4.

The GABA signaling pathway partially mediates GSC activation in sss^P1 males. GABA was increased in sss^P1 mutants by inactivating the gabat gene (sss^P1; gabat^PL) or using a GABA_A receptor agonist isoguvacine (IG). Although sss^P1 mutants show elevated division frequency compared with both untreated and IG-fed WT flies, activation of GABA signaling partially reduces GSC division rates in sss mutants. *P < 0.05; **P < 0.005. Error bars represent SEMs.

Discussion

Rhythms of Division in the Spermatogonial Stem Cell Niche.

Stem cell function is essential for renewing cellular content of the organism during aging and tissue repair (4–6). The frequency of divisions in adult stem cell populations depends on tissue-specific requirements for cell production and can be modulated in response to environmental and behavioral stimuli (7). Both excessive and insufficient cell supplies perturb tissue homeostasis and may compromise organ function, causing serious health conditions. In this study, we aimed to understand how periodic changes in the organism’s external environment modulate stem cell activity and tissue renewal.

Previous reports have shown that the day–night cycle affects cell cycle progression in a variety of unicellular and multicellular organisms and produces 24-h mitotic rhythms in many tissues, including rapidly dividing epithelia and the hematopoietic system (8, 10) and the more slowly renewing mammalian brain (9, 11, 12, 15). However, in these studies, a possible temporal coordination of stem cell divisions remained largely unexplored, because typically, all mitotic cells, not specifically stem cells, were analyzed. Positioned at the top of the cell hierarchy, stem cell populations undergo relatively infrequent mitotic divisions, producing further dividing restricted progenitor cells or transiently amplifying differentiating cells (5, 6). Thus, these populations represent only a portion of all replicating cells in the organism, making their identification and purification from intact tissues quite an elaborate task. Mouse hair follicle stem cells contain an endogenous circadian clock, which orchestrates a global transcriptional response during the anagen (hair follicle growth) phase, inducing stem cell migration and mitotic activation (14). However, the question of whether hair follicle stem cells undergo mitotic divisions preferentially at certain times of the day has not been addressed. To our knowledge, our work here and another recent study (30) are the only ones to assay stem cell division over the course of the day–night cycle. We took advantage of a well-characterized Drosophila stem cell niche, which maintains spermatogenesis in the adult fly and consists of two easily identified stem cell populations (male GSCs and CPCs) (Fig. 1A) (17), to address temporal regulation of tissue homeostasis. We also determined whether the fly behavioral state, undergoing daily transitions between wakefulness and sleep, perturbs normal stem cell activity in the testis niche.

We found significant diurnal variations in division frequencies in male GSCs and CPCs; specifically, mitotic counts increased during the day and decreased at night (Fig. 2 A and B and Table S1). Interestingly, mitotic activity in the GSC and CPC populations was coordinated throughout the 24-h cycle, which may be important to ensure an adequate supply of differentiating germ and somatic cyst cells in the formation of cysts, the initiating step in sperm development (Fig. 1A) (18). However, it is important to note that, although mitotic activity was rhythmic, DNA replication, measured through BrdU incorporation, was not. We suggest that mitosis in these cells is gated by the circadian clock, much like eclosion is in flies (31). Eclosion, which is the hatching of adult flies from their pupal cases, is rhythmic in Drosophila, whereas the preceding developmental events are not; basically, developmentally ready flies are held in pupae until the appropriate time of day when the gate opens. Similar gating in this system, perhaps at the mitotic transition, would account for rhythmic mitosis, even when other aspects of the cell cycle are not rhythmic.

Many, but not all, rhythms are driven by endogenous circadian clocks (1, 2). Consistent with previous findings (32), we did not detect expression of the core circadian clock genes period and timeless in the testis niche or stem cells, which suggests that there is no intrinsic clock function in these cells. Interestingly, murine spermatogonial stem cells also lack cell-autonomous circadian clocks (33), which is in contrast to clock-containing hair follicle, hematopoietic, and intestinal stem cells (14, 34–36). It is possible that central or peripheral circadian clocks located in the brain, other tissues (2), or even other cells of the testes drive rhythms in the GSCs/CPCs, although the absence of mitotic rhythms in constant darkness argues against this possibility (Fig. 2 C and D). However, many peripheral clocks dampen in constant darkness, and therefore, we cannot conclusively dismiss circadian regulation. Yet, it is clear that environmental LD cycles can drive rhythms of stem cell activity. Interestingly, the production of several components of the light transduction pathway in the testis (37, 38) suggests that testis cells may receive direct light inputs.

Loss of SSS Impacts GSC Divisions in the Testis Niche.

Day–night cycles drive rhythms in many physiological and behavioral processes, particularly sleep. We found that mitotic activity of GSCs and CPCs increases during the day, when flies are mostly awake, whereas prolonged nighttime sleep coincides with reduced division rates (Fig. 2 A and B). Similarly, specific populations of mouse epithelial cells preferentially undergo mitotic divisions during the active phase, although in the mouse, this phase corresponds to nighttime hours, and show decreased mitotic rates in the morning during the transition to sleep (8). Earlier studies have also shown that sleep deprivation affects adult neurogenesis in the mammalian brain (39, 40) and may impact cell turnover in other nonneuronal tissues (41). Based on these observations, we proposed that the behavioral status of the organism influences cell cycle progression in the testis niche. To test this hypothesis, we examined stem cell activity in several mutant lines that show disrupted sleep. Our results show that males carrying a loss-of-function mutation in the gene sleepless, sss^P1, exhibit significantly elevated division rates of GSCs throughout a 24-h cycle (Fig. 3D). In other short-sleeping mutants, GSC activity was not altered (Fig. 3A), presumably because of their less severe sleep phenotypes (23–26). Alternatively, SSS may modulate stem cell divisions through mechanisms unrelated to its function in sleep, or the sleep-regulating pathways affected in sss flies may be different from the pathways altered in other sleep mutants (discussed below). However, we also noted increased stem cell divisions after sleep deprivation using mechanical stimuli.

Rhythms of sleep and wake persist, even in the absence of day to night transitions (for instance, under DD conditions); however, rhythmic variations in stem cell division rates seem to dampen when flies are kept in a constant environment (Fig. 2 C and D). Although GSC activity shows pronounced fluctuations in constant dark, the dynamics of these changes, at least on the fourth day in DD, are inconsistent with a 24-h rhythm. CPC divisions also do not display a 24-h rhythm in DD and indeed, are invariant over the course of the day. Currently, we do not know why GSC division rates fluctuate in DD, but it is possible that these cells, more so than CPCs, are sensitive to small perturbations in the internal and external environments.

GABA signaling is known to have a sleep-promoting role in both mammals and flies (29); sss^P1 flies show reduced GABA levels, which when increased, largely restore sleep duration in these mutants (28). We found that increasing GABA signaling in sss^P1 males slightly reduces GSC activity in the testis, although it does not restore the frequency of GSC divisions to the levels detected in WT flies (Fig. 4) or flies with SSS that is rescued specifically in neurons (Fig. 3C). Given that changes in GABA signaling affect both sleep and the stem cell phenotype, it is likely that there is some mechanistic overlap between these processes. However, differential rescue of the two phenotypes by increases in GABA signaling argues against a direct effect of sleep loss on stem cell divisions. Thus, we cannot conclude that sleep amount and stem cell activity are causally linked, but we can say that some sleep-regulating pathways (e.g., GABA) also affect stem cell division. Because GABA is a very well-known sleep modulator and also the target of sleep-enhancing drugs, we suggest that problems with sleep can often be accompanied by changes in stem cell function. Lack of a stem cell phenotype in other sleep mutants may reflect the involvement of biochemical pathways other than GABA.

Because GABA only partially restores WT levels of GSC activity, other mechanisms might also be involved in the stem cell phenotype in sss mutants. In Drosophila, the expression of genes involved in the immune response is increased after sleep deprivation, and immune gene function contributes to regulating the amount of daily sleep (42, 43). It is possible that, in sss^P1 flies, the activation of GSC divisions results from an increased immune response caused by extended wake. Although the regulation of stem cell function by the immunological status of the organism is not yet completely understood, mitotic activity in stem cell populations is affected by bacterial infection and inflammation in both mammalian and fly tissues (44–47). Alternatively, oxidative stress could contribute to the altered stem cell activity in sss mutants. Sleep-deprived flies exhibit elevated levels of reactive oxygen species in the brain and possibly other tissues (3), and oxidative stress was shown to affect stem cell niches by stimulating stem cell division (48, 49) and inducing differentiation (50). Although reactive oxygen species or altered immune function have not been linked to the sleep phenotype of sss mutants, GABA signaling is shown to mediate reduced sleep levels in these flies, and our study has implicated its involvement in increased stem cell divisions, supporting a mechanistic link between sleep and the regulation of stem cells.

Materials and Methods

Drosophila Stocks and Circadian Regimens.

Flies were raised at room temperature on standard cornmeal/molasses medium. Before dissections, 0- to 3-d-old males were synchronized to a 12:12-h LD cycle for 5 d and then shifted to a DD regimen when required using circadian incubators (DigiTherm CircKinetics; Tritech Research). We used the following fly lines: WT Canton-S (Fig. 2) and iso31 strains (controls and genetic background in Figs. 3 and 4) (Bloomington Stock Center), sss^P1 (23), DAT^fumin (24), shaker gene locus (25), gabat^PL (provided by W.-F.C.), elav-GeneSwitch (51), elav-Gal4 (Bloomington Stock Center), UAS-mc* (52), and UAS-sss (27) strains. To induce neuronal PKA activity, elav-GeneSwitch/Y; UAS-mc* males were fed RU486-containing food (500 µM in ethanol) for 2 d. To boost GABAergic signaling, iso31 and sss^P1 males were fed 3 mM isoguvacine (G002; Sigma) for 3 d.

Immunostainings.

To determine GSC activity, testes were dissected and fixed in 4% (vol/vol) formaldehyde (20 min), blocked in PBS, 0.1% Triton X-100, 5% (m/vol) BSA, and 2% (m/vol) normal donkey sera, and incubated in primary and secondary antibodies. We used the following primary antibodies: mouse anti-Armadillo (1:200; N2 7A1; DSHB), rabbit antiphospho-Histone H3 (1:1,000; 06–570; Millipore), goat anti-Vasa (dN-13; 1:200; sc-26875; Santa Cruz Biotechnology), and rabbit anti-BrdU (1:100; 600–401-C29; Rockland Immunochemicals) antibodies combined with secondary anti-mouse and anti-rabbit Alexa₄₈₈ and anti-goat Alexa₅₆₈ fluorescent antibodies (1:400; Invitrogen). For BrdU treatment, testes were incubated in 10 µM BrdU (B5002; Sigma) solution in PBS for 30 min. DNA was counterstained with To-Pro-3 iodide (1:1,000; T3605; Invitrogen). Data analysis was done using a Leica confocal system DM6000. Stem cell divisions were analyzed using multiple confocal sections throughout the testis apical area.

Statistical Analysis.

We calculated stem cell division rates for each genotype or experimental condition based on the numbers of dividing stem cells in individual stem cell niches. The significance of differences between samples was evaluated using the Pearson χ² test of goodness of fit and independence. Because most niches/testes had either no or a single dividing cell, we used the total number of testes with or without dividing stem cells as the input for χ² tests. Error bars on all figures represent SEMs of measurement.