Opposing action of Hedgehog and Insulin signaling balances proliferation and autophagy to determine Follicle Stem Cell Lifespan

SUMMARY

Egg production declines with age in many species, a process linked with stem cell loss. Diet-dependent signaling has emerged as critical for stem cell maintenance during aging. Follicle Stem Cells (FSCs) in the Drosophila ovary are exquisitely responsive to diet-induced signals including Hedgehog (Hh) and Insulin (IIS), entering quiescence in the absence of nutrients and initiating proliferation rapidly upon feeding. Although highly proliferative FSCs generally exhibit extended lifespan, we find that constitutive Hh signaling drives FSC loss and premature sterility despite high proliferative rates. This occurs due to Hh-mediated induction of autophagy in FSCs via a Ptc-dependent, Smo-independent mechanism. Hh-dependent autophagy increases during aging, triggering FSC loss and consequent reproductive arrest. IIS is necessary and sufficient to suppress Hh-induced autophagy, promoting a stable proliferative state. These results suggest that opposing action of diet-responsive IIS and Hh signals determine reproductive lifespan by modulating the proliferation-autophagy balance in FSCs during aging.

In Brief (50 words or fewer)

Singh, et al shed light on how the diet-regulated growth factors Hedgehog and Insulin maintain Drosophila Follicle Stem Cells during aging. Hedgehog induces autophagy in FSCs via Patched, creating an untenable stem cell state that results in early sterility. Insulin suppreses Hh-induced autophagy, restoring proliferative balance and extending FSC lifespan.

INTRODUCTION

Unlike the decay of inanimate objects, living systems retain the capacity to self-renew by virtue of long-lived stem cells that can repopulate lost or defective tissues in aging organisms. Stem cells respond to cues from their local environments, called niches, and to systemic signals that depend on external environmental influence, such as nutrition. Aging is associated with genetic and metabolic alterations in stem cells that impact their ability to compete for niche occupancy and/or produce functional daughter cells (Schultz and Sinclair, 2016). These alterations determine the lifespan of individual stem cells. Acquisition of detrimental changes reduces stem cell longevity, and beneficial changes promote expansion and long-term maintenance of clonally-derived populations (Schultz and Sinclair, 2016; Signer and Morrison, 2013). Importantly, interventions that promote stem cell longevity improve tissue function and health over time, emphasizing the importance of stem cell regulatory mechanisms for counteracting age-associated symptoms (Signer and Morrison, 2013).

One mechanism that controls stem cell lifespan is diet. Dietary restriction in the form of reduced calories, feeding and fasting cycles, or reduced intake of specific nutrients is the best-studied method for increasing organismal longevity (Longo et al., 2015). Diet impacts key hallmarks of cellular aging including diminished nutrient sensing, genomic instability, altered proteostasis, age-associated epigenetic changes, mitochondrial dysfunction, senescence, and altered cell-cell communication (Longo et al., 2015; Lopez-Otin et al., 2013). The same hallmarks of organismal aging diminish stem cell lifespan by reducing self-renewal capacity, niche occupancy, survival, or ability to generate differentiated daughter cells (Lopez-Otin et al., 2013; Signer and Morrison, 2013). Dietary restriction ameliorates accumulation of these hallmarks, enhancing the longevity of both stem cells and the organism in parallel (de Cabo et al., 2014). In contrast, excess nutrients or continuous feeding can reduce stem cell lifespan and impact organismal longevity (Longo et al., 2015; Signer and Morrison, 2013).

The effects of dietary changes, both positive and negative, are mediated by molecular sensors that translate nutritional inputs to control signaling pathways that influence stem cell function during aging. Key among these is the insulin-IGF signaling (IIS) cascade. Reduced IIS promotes longevity in most stem cell populations and organisms including worms, flies, mice, and humans (Lopez-Otin et al., 2013; Shim et al., 2013; Signer and Morrison, 2013). Conversely, constant or high levels of IIS signaling are associated with reduced organismal lifespan, accelerated stem cell loss, and increased incidence of age-related diseases (van Heemst, 2010). Consistent with this general principle, dietary restriction promotes organismal longevity in Drosophila melanogaster females, an effect that is mimicked by reduced IIS signaling (Grandison et al., 2009). Unlike other systems, however, reduced IIS is counterproductive for egg production (fecundity) in flies (Grandison et al., 2009; Hsu and Drummond-Barbosa, 2009; Ikeya et al., 2002; LaFever and Drummond-Barbosa, 2005). Fecundity defects caused by reduced IIS can be partially reversed by enhancing the diet with key nutrients (Grandison et al., 2009; Hsu and Drummond-Barbosa, 2009), suggesting that a precise combination of nutrients promotes successful reproduction and delays aging.

The impact of diet on egg production is mediated in part by the response of ovarian stem cells to nutrient-dependent signals (Shim et al., 2013). Drosophila ovarian stem cells reside in a structure called a germarium located at the anterior-most end of each of 16 ovarian units (ovarioles) that comprise each ovary. Germline Stem Cells (GSCs) undergo asymmetric divisions to self-renew the stem cell population and simultaneously generate a daughter cell (cystoblast) that divides four times to produce a cyst of 16-cells, including the developing oocyte (King, 1970). Germline cysts are encapsulated in the center of the germarium by the daughters of epithelial Follicle Stem Cells (FSCs) (Margolis and Spradling, 1995), resulting in the formation of a cuboidal epithelium surrounding each developing cyst. The resulting follicles, called egg chambers, develop through 14 stages to produce mature eggs. Ovarian stem cell function is exquisitely regulated by diet, with arrest of GSC and FSC proliferation during periods of nutrient restriction (Drummond-Barbosa and Spradling, 2001; Hartman et al., 2013). This non-proliferative, quiescent stem cell state is reversible, with rapid transition of both stem cell populations to proliferation and robust egg production within hours of feeding flies a rich diet of yeast. A diet lacking protein and lipids or reduced IIS can diminish GSC number and function with age (Drummond-Barbosa and Spradling, 2001; Hartman et al., 2013; Hsu and Drummond-Barbosa, 2009). This suggests that nutritional imbalance can impact female fecundity by driving GSC loss over time.

Altered function of FSCs also impacts egg development. FSCs, like GSCs, are highly sensitive to dietary changes, with proliferative arrest in the absence of sufficient nutrients (Drummond-Barbosa and Spradling, 2001; Hartman et al., 2013; Hartman et al., 2010). This arrest depends on two critical pathways. First, FSCs are responsive to dietary cholesterol, which stimulates FSC proliferation after a period of nutrient restriction via activation of the Hedgehog (Hh) signaling pathway (Hartman et al., 2013). In addition to Hh, IIS has been implicated in FSC regulation. A reduced IIS response in apical cells suppresses proliferation of both GSCs and FSCs in a Hh-independent manner (Hartman et al., 2013). IIS may also contribute to FSC function autonomously, as IIS effectors have been implicated in FSC maintenance or proliferation (LaFever and Drummond-Barbosa, 2005; LaFever et al., 2010; Wang et al., 2012). Here, we present evidence that cooperative action of Hh and IIS pathways cell-autonomously promotes FSC longevity via regulation of autophagy, with defects that increase autophagy reducing FSC lifespan. We find that simultaneous activation of autophagy and proliferation drives FSC loss from the niche, resulting in age-dependent arrest of egg production. Our results suggest that the precise balance of nutrients that induce IIS and Hh signaling (protein/carbohydrate and cholesterol, respectively) determines FSC lifespan, thus providing a mechanistic explanation for why a balanced diet promotes optimal fecundity over time.

RESULTS

Constitutively Active Hh Signaling Leads to Decline in FSC Lifetime

Classical lineage tracing studies define FSCs as the anterior-most cells in the follicular epithelium that have low expression of the epithelial marker FasIII and a defined position at the Region 2A/2B border (Figure 1A) (Margolis and Spradling, 1995; Nystul and Spradling, 2007, 2010; Zhang and Kalderon, 2001). Recent work suggests the existence of three pools of cells with FSC characteristics, arranged around the surface of the germarium in adjacent rings at the Region 2A/2B border (Reilein et al., 2017). Our studies focus on classically defined FSCs (Layer 2 FSCs), which reside one cell diameter anterior to FasIII-expressing follicle cells, as their quiescent and proliferative states are controlled by Hh in a feeding-dependent manner (Hartman et al., 2013). FSCs in nutrient-restricted flies do not divide, but cholesterol ingestion triggers Hh release and induction of proliferation. Reduced Hh signaling in fed flies prevents FSC proliferation, and constitutive Hh release bypasses quiescence, even in nutrient-restricted flies (Hartman et al., 2013).

Figure 1. Excess Hh Reduces FSC Lifespan.

(A) Germarium schematic. Germline Stem Cells (GSCs) (dark blue) reside anterior, adjacent to apical cells (green) and produce 16-cell germline cysts, including the oocyte (yellow). FSCs (red) reside in region 2A/2B, and produce follicle cells (FSCs, pink). Germ cell/follicle cell units form egg chambers.

(B) Fraction of dividing FSCs (PH3 positive) per genotype quantified for week 1. *, p= 0.03 relative to w¹¹¹⁸ (WT). n= number of germaria counted per genotype represented in all figures. Error bars = Standard Error of the Mean (SEM), actual values in Table S1.

(C) Average days till oogenesis arrest (no viable embryos) per genotype. *, p<= 0.03 relative to WT is shown. 5 flies/genotype/experiment, (WT, SEM=3.8), (boi^e, SEM=3.3), UAS-stg.N16 (w;String, SEM=3.8,). Error bars = SEM.

(D)Eggs laid/fly/week/genotype for weeks 1-6 for fed WT, boi^e and 109-30 Gal4-UAS-string flies. *, p< 0.05 at 2, 3 and 4 weeks (boi^e vs WT). Experiments were performed in triplicate (5 flies/genotype/experiment). Error bars=Standard Deviation (SD) (WT, week (wk) 1: SD=3.9, wk 2: SD= 10.3, wk3: SD= 5.1, wk4: SD= 5.5, wk5: SD= 2.9, wk6: SD= 0.5), (boi^e, wk 1: SD=3.4, wk 2: SD= 3.5, wk3: SD= 5.7, wk4: SD= 1.7, wk5: SD= 0.1, wk6: SD= 0.1).

(E) Eggs laid/fly/week/genotype scored for week 1 through week 3 for fed hsHh gof flies, kept at temperatures indicated (5 flies/genotype/experiment).

(F) Loss of GFP-labeled FSCs over a 4 week timecourse: Control (RFPGal8019Aflp; 109-30Gal4;UAS-GFP^nls) (wk 1: n=308, SEM=0.028, wk 2: n= 331, SEM= 0.027, wk3: n= 158, SEM= 0.039, wk4: n= 153, SEM= 0.04), Hh-gof mutants (RFPGal8019Aflp; 109-30Gal4/hsHh; UAS-GFP^nls) (wk 1: n=402, SEM=0.016, wk 2: n= 246, SEM= 0.023, wk3: n= 40, SEM= 0.053, wk4: n= 95, SEM= 0.015) Clones were induced on day 0. *, p< 0.0004 at weeks 1-4. Error bars = SEM.

(G) GFP-labeled FSCs in Controls (RFPGal8019Aflp; 109-30Gal4;UAS-GFP^nls) (top) or Hh-gof mutants (RFPGal8019Aflp; 109-30Gal4/hsHh; UAS-GFP^nls) (bottom) 1-4 weeks after lineage marking. Follicle cells (Fas III, red), nuclei, (Draq5, blue), and GFP-labeled FSCs and progeny (green) are shown. FSC is indicated (dotted yellow triangle). Follicle cells are located posterior to Region 2A/2B (white bracket).

Recent work suggests that proliferation is a key determinant of FSC longevity. More proliferative FSCs compete effectively for niche occupancy by displacing less proliferative cells from the niche and promoting their differentiation (Nystul and Spradling, 2010; Wang and Kalderon, 2009). We found that ectopic expression of String, the fly homolog of the cell cycle regulator CDC25 (Edgar and O'Farrell, 1989), exclusively in FSCs and early pre-follicle cells (UAS-String/109-30 Gal4) elevated FSC proliferation rates, dramatically extended fertile lifetime, and increased egg numbers relative to wild-type (WT) females (Figure 1B-D, Table S1). Thus, proliferation extends FSC longevity in parallel with fecundity, supporting a model in which FSC lifespan is determinative for timing reproductive arrest.

Based on these observations, we anticipated that mutants with constitutive Hh signaling and FSC proliferation levels similar to those observed upon ectopic String expression (Figure 1B), would yield a similar effect on fecundity. We first analyzed boi mutants, which lack expression of the Hh sequestering protein Boi in apical cells and constitutively release Hh ligand to drive FSC proliferation (Hartman et al., 2010). Rather than enhancing fecundity, boi mutation resulted in fewer eggs and fecundity arrest nearly two weeks earlier than WT flies (Figure 1C, D). Temperature-induced expression of Hh ligand yielded a similar effect (hs-hh/+, Figure 1E). Lineage labeled FSCs in mutants with constitutive Hh (Flp FRT 19A Gal80/Flp FRT 19A; 109-30 Gal4, hs-hh/UAS-GFP (Ingham, 1993)) were rapidly lost (Figure 1F). FSCs were not lost due to cell death, based on the absence of cleaved caspase 3 and TUNEL staining, and no difference in incorporation of propidium iodide in live WT versus boi mutant FSCs (data not shown). To determine the fate of FSCs in flies with constitutive Hh signaling, we lineage labeled a subset of dividing ovarian cells with GFP and measured their retention in the germarium over a 4-week timecourse (Hartman et al., 2015; Margolis and Spradling, 1995). Clones initially generated in differentiating follicle cells incorporate into egg chambers and exit the germarium within 5 days after labeling. Marked FSCs, in contrast, are maintained, continuing to produce marked daughters until they are displaced from the niche by more competitive FSCs (Margolis and Spradling, 1995). Thus, FSC lifespan under different genetic or environmental conditions can be determined based on the presence of marked FSCs at the Region 2A/2B border relative to GFP-labeled cells that were recently displaced from the niche.

In WT germaria, GFP-labeled FSCs and follicle cells were observed throughout a 4-week timecourse, consistent with maintenance of self-renewing FSCs during aging (Margolis and Spradling, 1995), Figure 1G). In contrast, GFP-labeled FSCs were rarely observed in gain-of-function hh mutants 2 weeks after the initial marking. Labeled FSCs were displaced from the niche throughout the timecourse, appearing instead as GFP-expressing follicle cells that incorporated into egg chambers (Figure 1G). Together, these results demonstrate that increased String expression extends FSC lifespan and fecundity whereas increased Hh signaling leads to FSC loss and early sterility. The contradiction between shortened FSC lifespan/early sterility in constitutive Hh mutants with highly proliferative FSCs suggests that additional Hh-dependent mechanisms must modify FSC lifespan independently of proliferation.

Autophagy Pathway Components are Linked to FSC Lifetime

To identify Hh-dependent mechanisms that control FSC lifespan, we took advantage of the observation that boi mutant females arrest egg laying 2 weeks prior to WT flies (Figure 1D). We reasoned that reduced expression of genes necessary for the Hh-dependent fecundity arrest should suppress the phenotype, resulting in restoration of fertility. Using this premise, we conducted a screen in which RNAi was expressed in FSCs and early follicle cells in boi mutant females, and the timing to fecundity arrest was measured. Among the genes that suppressed boi mutant fecundity arrest were four components of the core autophagy machinery (Figure 2A). Autophagy is a lysosome-dependent degradation pathway that has been implicated in aging, both at the organismal and stem cell levels (Russell et al., 2014). Autophagy is induced in order to maintain the health and viability of cells by clearing damaged proteins or organelles and recycling critical cytoplasmic materials for energy production or anabolic processes (He and Klionsky, 2009). Notably, autophagy is particularly critical under conditions that reduce stem cell longevity (Russell et al., 2014). Autophagy proceeds through sequential recruitment of core proteins that promote initiation, nucleation of a specialized structure called a phagophore, expansion of phagophores into autophagosomes, and finally fusion with the lysosome and cargo degradation (Figure 2B, (He and Klionsky, 2009)). Critical components of each step were identified in our screen including Atg1/Ulk1 (initiation), Atg6/Beclin-1 (nucleation), Atg9 (phagophore elongation), and Atg8b/LC3 (conjugation and expansion into autophagosomes) (Figure 2B). Moreover, kinases including Liver Kinase B1 (LKB1) and Protein Kinase A (PKA) that control autophagy induction (Russell et al., 2014) were identified as suppressors of boi mutant fecundity arrest (Figure 2C).

Figure 2. Autophagy Genes Suppress boi mutant Fecundity Arrest.

(A) Average number of days until oogenesis arrest (no viable embryos) per genotype. Experiments were performed in triplicate (5 flies/genotype/experiment). WT (SEM=3.2), boi^e (SEM=3.7) boi^e; Atg1^RNAi (SEM=1.7), boi^e; Atg6^RNAi (SEM=0.2), boi^e; Atg9^RNAi (SEM=1.7), boi^e; Atg8b^RNAi (SEM=0.8). Error bars = SEM.

(B) Autophagy proceeds through 4 steps: 1) Initiation (Atg1 complex), 2) Phagophore nucleation (Atg6 complex), 3) Expansion into autophagosomes (Atg9 complex), 4) Fusion of autophagosomes to lysosome and degradation of engulfed components (Atg8 complex). Genes associated with each step were identified in our screen (highlighted with matched colors in A).

(C) Average number of days until oogenesis arrest (no viable embryos) per genotype. Experiments were performed in triplicate (5 flies/genotype/experiment). WT (SEM= 3.8), boi^e (SEM= 3.3) boi^e; PKA-C1^RNAi (SEM=0), boi^e; LKB1^RNAi (SEM=2.02). Error bars=SEM.

Autophagy can be detected in fly cells using Lysotracker (Ltr), a dye that binds to autophagic vesicles after they fuse with the lysosome (DeVorkin and Gorski, 2014). WT germaria isolated from young flies (<7 days old) rarely exhibited Ltr staining, except under starvation conditions when FSCs stained positively (Figure 3A-C, Table S2, (Barth et al., 2011)). Strikingly, a high percentage of germaria isolated from fed or starved boi mutants exhibited strong Ltr staining of FSCs (Figure 3A’, B’, C’, Table S2). Similarly, expression of Hh ligand in apical cells promoted autophagy induction (Figure 3D, E, K). Boi was required in apical cells for autophagy induction, as reducing expression using RNAi in apical cells (AC- boi^RNAi (boi^RNAi/+; bab-Gal4/+)), but not FSCs (FSC- boi^RNAi (boi^RNAi/109-30-Gal4)) promoted increased Ltr staining in FSCs (Figure 3K). Occasionally, cells anterior to FSCs exhibited Ltr staining in boi mutants (Figure 3A”, B’). These cells may represent the posterior-most escort cells that comprise the FSC niche (Margolis and Spradling, 1995) or newly defined Layer 3 FSCs (Reilein et al., 2017). Ltr staining also was observed in rare germ cells (Figure 3B”). We confirmed the correlation between Ltr staining and autophagy by observing that Atg8a/LC3-GFP, a highly specific marker of autophagosome membranes that condenses into vesicles within the cytoplasm of cells undergoing autophagy (Nagy et al., 2015) co-localized with Ltr in FSCs in 32% of boi mutant germaria, relative to 14% in WT (Figure 3F-H). Autophagy leads to lysosomal-mediated degradation of many proteins, including ubiquitinated cargo associated with Ref(2)P/p62, which is itself an autophagic target (Lőrincz et al., 2017). Ref(2)P levels increase when autophagy is inhibited (e.g. in fed flies) and decrease upon autophagy activation (e.g. under starvation conditions) (Figure 3I, J, (Lőrincz et al., 2017)). Consistent with a function for boi in promoting autophagy, Ref(2)P levels were dramatically lower in fed boi mutants relative to wild-type flies (Figure 3I, J).

Figure 3. Autophagy Induction in Hh Gain-of-Function mutants.

(A,B,D) Autophagy quantified in FSCs using lysotracker (ltr, red) in germaria with labeled follicle cells (Fas III, green) and nuclei (blue) (A-A’) Starved WT and boi^e mutants (A”) boi^e (starved) germarium with ltr staining in an escort cell/layer 3 FSC. (B-B’) Autophagy is induced in fed boi^e mutant flies at ~6-fold higher levels than WT. (B”) ltr staining in germ cells anterior to Layer 2 FSCs.

(C) Fraction of germaria with lysotracker (ltr) positive FSCs in starved or fed conditions. *, p< 0.0001 vs. WT (starve). **, p< 0.0001 vs. WT (fed). Error bars = SEM.

(D) Constitutive Hh induces ltr at the permissive temperature (29°C).

(E) Fraction of hs-hh mutant germaria with ltr staining in FSCs during fed conditions. *, p= 0.001 vs. hs-Hh vs. controls at 18°C. Error bars = SEM.

(F,G) dAtg8a-GFP condenses into vesicles that co-stain with ltr (F’,G’) in boi^e mutant FSCs, but not w; 109-30; UAS-dAtg8a-GFP. (F”,G”) Triple labeling for Atg8a-GFP (green), ltr (red) and follicle cells (blue). FSC in G” indicated by yellow arrow.

(H) Fraction of germaria with ltr and GFP co-staining in Layer 2 FSCs. *, P= 0.0004 vs. w; 109-30;UAS-dAtg8a-GFP. Error bars = SEM.

(I) Ref(2)P/p62 protein levels in starved or fed WT and boi^e mutants. ẞ-actin was used as a loading control.

(J) Quantitation of Ref(2)P levels in WT and boi^e mutants. *, p= 0.03 vs. WT (fed). Experiments were done in triplicate. Error bars = SEM, (WT, starve, SEM=0.004) (boi^e, starve, SEM=0.009) (WT, fed, SEM=0.126) (boi^e, f ed, SEM=0.026).

(K) Boi and Hh function in apical cells. Reduced boi (babGal4; UAS-boi^RNAi) or expression of Hh (babGal4; UAS-Hh-N) in apical cells (labeled “AC- boi^RNAi), but not FSCs (boi (w; 109-30; UAS-boi^RNAi, labeled “FSC- boi^RNAi”), w; Hh-N (109-30tubGal80^ts; UAS-Hh-N)) induced autophagy. Reduced expression of autophagy genes suppressed boi-induced autophagy. Fraction of germaria with ltr positive FSCs is indicated. *, p=<0.0001 vs. WT and **, p= 0.001 vs. boi^e,***, p=< 0.0001 vs. boi^e. Error bars = SEM.

(L). Activated AMPK (UAS-AMPKα^T184D) or overexpressed Atg1 (UAS-Atg1) induce autophagy, indicated by increased ltr positive FSCs versus WT (*, p< 0.0001 vs. WT). boi^e mutant flies expressing RNAi targeting LKB1, AMPK and PKA-C1^B10 under control of 109-30 Gal4 (boi^e; 109-30Gal4/+; RNAi/+) exhibit reduced autophagy relative to boi mutants alone (**, P< 0.0001 vs. boi^e). Error bars = SEM.

(M) Reduced Atg5 expression suppressed FSC loss due to constitutive Hh signaling.FSC loss over a 4-week timecourse for flies bearing GFP-labeled FSCs from Controls (RFPGal8019Aflp; 109-30-Gal4; UAS-GFP^nls) (wk 1: n=308, SEM=0.028, wk 2: n= 331, SEM= 0.027, wk3: n= 158, SEM= 0.039, wk4: n= 153, SEM= 0.04), Hh-gof mutants (RFPGal8019Aflp; 109-30Gal4/hsHh; UAS-GFP^nls) (wk 1: n=402, SEM=0.016, wk 2: n= 246, SEM= 0.023, wk3: n= 40, SEM= 0.053, wk4: n= 95, SEM= 0.015), Hh-gof-Atg5^RNAi (RFPGal8019Aflp; 109-30Gal4/hsHh; UAS-GFP^nls/UAS-Atg5^RNAi) (wk 1: n=219, SEM=0.033, wk 2: n= 58, SEM= 0.064, wk3: n= 117, SEM= 0.045, wk4: n= 256, SEM= 0.03), or activated AMPK mutants (109-30Gal4/hsHH; UAS-GFP^nls/ UAS-AMPK^T184D) (wk 1: n=117, SEM=0.037, wk 2: n= 144, SEM= 0.022, wk3: n= 137, SEM= 0.01) was measured. *, p=< 0.0004 at 1-4 weeks and *, p=< 0.00001 at 1-3 weeks vs Hh-gof. Error bars = SEM.

See also Table S2.

Importantly, the increased Ltr staining observed in boi mutants was suppressed by reducing expression of core autophagy genes in FSCs (boi^e; 109-30-Gal4/+; Atg^RNAi/+, Figure 3K, Table S2). Similarly, reduced expression of LKB1 and PKA or AMPK, a kinase that triggers autophagy induction via direct phosphorylation of Atg1/ULK-1 (Russell et al., 2014), suppressed Ltr staining in FSCs in boi mutants (Figure 3L, Table S2). The impact of reducing core autophagy gene expression on FSC lifespan was striking. Whereas GFP-labeled FSCs in ovaries with constitutive Hh signaling were rapidly lost, reduced expression of Atg5 restored FSC lifespan to WT levels (Figure 3M). Conversely, ectopic induction of autophagy upon overexpression of Atg1/ULK-1 or a constitutively activated form of AMPK (AMPK^T184D) (Mirouse et al., 2007b) in FSCs was sufficient to induce autophagy (Figure 3L, Table S2). Constitutive activation of AMPK also led to rapid loss of FSCs from the niche in a manner similar to that observed for Hh gof mutants (Figure 3M).

Hh-mediated Autophagy is Ptc-dependent and Smo-independent

The Hh signaling pathway that controls FSC proliferation has been well studied. Hh ligand is produced by apical cells and controlled by a Boi-mediated sequestration and release mechanism (Hartman et al., 2013; Hartman et al., 2010). In FSCs, Hh relieves the inhibitory activity of Patched (Ptc) on Smoothened (Smo), resulting in activation of Cubitus Interruptus (Ci), a Gli-family transcription factor that promotes expression of cell cycle regulatory genes (Duman-Scheel et al., 2002; Forbes et al., 1996; Wang and Kalderon, 2009; Zhang and Kalderon, 2001). In addition to elevated FSC mitosis, boi mutants exhibit excess production of follicle cell daughters that accumulate between egg chambers (Figure 4A,B,F, (Hartman et al., 2010)), and elevated autophagy (Figure 4A, F). Reduced Hh expression in apical cells suppressed boi-induced FSC proliferation and autophagy (Figure 5A, (Hartman et al., 2010)), demonstrating ligand dependence. Reducing Smo in FSCs suppressed feeding-stimulated proliferation in WT flies and the elevated proliferation observed in boi mutants (Figure 4A, D, Table S2, (Hartman et al., 2013; Hartman et al., 2010). We initially hypothesized that increased FSC proliferation in boi mutants might induce autophagy to promote recycling of cellular materials during proliferative stress. Surprisingly, however, reducing Smo or Ci levels in FSCs had no effect on autophagy in boi mutants under conditions where proliferation was dramatically suppressed (Figure 4A, G, Tables S1,2). During aging in WT flies, autophagy was low when FSCs were highly proliferative and steadily increased as proliferation began to drop (Figure 4I,K Tables S1, 2). Proliferation was high throughout the timecourse in boi mutants, with autophagy increasing coordinately with fecundity arrest (Figure 1D, 4J,L Tables S1,2). Importantly, String expression in FSCs increased proliferation, but did not induce autophagy (Figure 4A,B, Tables S1,2). In fact, String expression in boi mutants suppressed autophagy and promoted proliferation, suggesting that highly proliferative FSCs are less prone to autophagy induction (Figure 4A,H). Together, these results suggest that proliferation and autophagy are controlled by separate Hh-dependent molecular mechanisms.

Figure 4. Autophagy Induction is Age-dependent and Smo-independent.

(A) Smo and String control FSC proliferation, but not autophagy. Fraction of germaria with ltr positive FSCs (black bars, left primary y-axis) or mitotic FSCs (PH3 positive, represented by gray bars, right secondary y-axis) per germarium is shown for one-week old flies. (109-30 Gal4/UAS-Smo^RNAi, 109-30 Gal4/UAS-String, respectively). Error bars = SEM.

(B-B”) Germaria of indicated genotypes stained for ltr (red), follicle cells (Fas III, green) and nuclei (Draq5, blue). boi^e mutants (B’) and flies expressing String in FSCs (B”, 109-30 Gal4/UAS-string) exhibit long stalks between follicles (white brackets) (B).

(C-H) Representative images for indicated genotypes depicting autophagy (ltr, (red), follicle cells (Fas III, green) and nuclei or germ cells (Draq5 or Vasa as indicated, blue)).

(C’-H’) Representative images for indicated genotypes depicting proliferation (PH3 (red), fcs (Fas III, green) and nuclei (Draq5, blue)).

(I) Autophagy increased in WT flies over a 4 week timecourse; proliferation decreased. Fraction of germaria with ltr positive FSCs (black bars, left y-axis) or dividing FSCs (PH3 positive) (gray bars, right y-axis). Error bars = SEM.

(J) boi^e mutants exhibit elevated autophagy and proliferation throughout the timecourse. Fraction of germaria with ltr positive FSCs (black bars, left y-axis) or dividing FSCs (PH3 positive) (gray bars, right y-axis). Error bars = SEM.

(K-K’, L-L’) Representative images for WT (K-K’) or boi^e mutants (L-L’) depicting autophagy and proliferation at week 3 (ltr (red) or PH3 (red), follicle cells (Fas III, green) and nuclei (Draq5, blue)).

See also Table S2.

Figure 5. Autophagy Induction Occurs through Patched and is Age-Dependent.

(A) Reduced ptc expression in boi^e mutants suppressed autophagy. Fraction of germaria with ltr staining in FSCs was quantified for indicated genotypes. *, P< 0.0001 vs. WT. **, P=< 0.0001 vs. boi^e. Error bars = SEM.

(B) Ptc is necessary and sufficient for autophagy induction. Fraction of germaria with ltr staining in FSCs is shown for indicated genotypes at restrictive (18°C) or permissive temperatures (29°C). *, P< 0.0001 vs. 109-30tubGal80^ts at 29°C and **, P= 0.001 (UAS-Smo^C320A vs ptc^RNAi; UAS-Smo^C320A) at 29°C. Indicated transgenes were expressed using 109-30 Gal4. Error bars = SEM.

(C-C”) (D-D”) Flies of indicated genotypes were fed for one week and stained for ltr, (red), follicle cells (Fas III, green) and nuclei or germ cells as indicated (Draq5 or Vasa, blue). Stalks are indicated (white brackets).

(E) Fraction of germaria with ltr staining in FSCs quantified for indicated genotypes. *, P< 0.0001 vs UAS-Ptc/. Error bars = SEM.

(F) Ref(2)P/p62 protein levels were measured in lysates collected from indicated genotypes. ẞ-actin is a loading control.

(G) Ltr positive FSCs were scored over a 4-week timecourse in fed flies of indicated genotypes. Fraction of germaria with ltr positive FSCs is shown. For activation of 109-30tubGal80^ts, flies were aged at 29°C. *, P<= 0.0006 relative to WT is shown at week 3 or week 4. Error bars = SEM.

See also Table S2.

Whereas Smo and Ci were dispensable, Ptc was necessary for autophagy induction in FSCs. Autophagy was significantly suppressed in boi mutants with reduced ptc expression, including in flies homozygous for a hypomorphic ptc mutant (boi^e; ptc^tuf1/ptc^tuf1) or expressing ptc-targeted RNAi in FSCs (boi^e; ptc^RNAi/109-30-Gal4) (Figure 5A, Table S2). The effects of ptc mutation on boi-mediated autophagy induction were further enhanced by reducing smo expression in FSCs (Figure 5A, Table S2). Since reducing smo expression alone had no effect on autophagy (Figure 4A, 5A, Table S2), this suggests that constitutive Smo/Ci signaling in ptc mutants partially rescues ptc mutant phenotypes. We tested this further by expressing a constitutively activated form of Smo (109-30tubGal80^ts/+; UAS-Smo^320A/+ (Carroll et al., 2012)). Smo^320A expression increased follicle cell numbers between stalks (Figure 5C), indicating excess FSC proliferation. Autophagy also was induced (Figure 5B). FSCs expressing Smo^320A with reduced ptc exhibited enhanced proliferation (Figure 5C’). In contrast, Smo^320A–induced autophagy was suppressed to wild-type levels by reducing ptc (UAS-ptc^RNAi/109-30 Gal4; UAS-Smo^320A/+, Figure 5B). Most likely, Smo activity drives Ci-dependent ptc transcription (Li et al., 2014), resulting in Ptc-dependent autophagy induction in hypomorphic ptc mutants.

Ptc expression also was sufficient to induce autophagy. Flies expressing temperature-induced WT Ptc in FSCs (w; 109-30 tubGal80^ts/UAS-ptc) exhibited minimal autophagy induction at the restrictive temperature. In contrast, Ptc induced autophagy robustly at the permissive temperature (Figure 5B,F). No increase in numbers of follicle cells between egg chambers was observed (Figure 5D), indicating that overexpressed Ptc maintained Smo in an inhibited state. Expression of a Ptc mutant in which the second extracellular loop is deleted, a region that includes a Hh binding domain (UAS-ptc^Δloop2/109-30tubGal80^ts (Casali and Struhl, 2004)), induced autophagy, albeit at a reduced level relative to wild-type Ptc (Figure 5B). Consistent with previous reports, Ptc^Δloop2 retained the ability to suppress Smo, as indicated by normal stalk lengths (Figure 5D”). A likely explanation is that deletion of Loop 2 induces a conformational change in Ptc that mimics Hh binding to activate the autophagy pathway, while maintaining the ability of Ptc to inhibit Smo.

Our results suggest that two domains of Ptc contribute to autophagy induction in FSCs. First, the C-terminal tail of Ptc is necessary for autophagy induction. FSCs overexpressing a form of Ptc that lacks the C-terminus (Ptc^1130X, (Johnson et al., 2000)) exhibited autophagy levels that were half those observed for WT Ptc and were indistinguishable from control flies raised at the restrictive temperature ((109-30 Gal4 tubGal80^ts/CyO), Figure 5B). This mutant did not inhibit Smo, as indicated by increased stalk length (Figure 5D’). The sterol sensing domain (SSD) of Ptc (Johnson et al., 2002; Martín et al.; Strutt et al., 2001) also was required for boi-mediated autophagy induction. boi mutants bearing heteroallelic ptc mutations that result in expression of the SSD mutant Ptc^D584N from the endogenous ptc promoter (boi; ptc^S2(D584N)/ptc^tuf1) exhibited excess follicle cell accumulation, supporting known dominant negative effects of Ptc^D584N on Smo activity (Figure 5C”, (Vied and Kalderon, 2009)). In contrast, Ptc^D584N suppressed autophagy induction in boi mutants (boi; ptc^S2(D584N)/ptc^tuf1, Figure 5A). Together, these data demonstrate that 1) Smo activation occurs independently of autophagy induction and 2) the C-terminal tail and SSD of Ptc contribute to autophagy induction in FSCs.

Our original screen uncovered core autophagy genes, LKB1, and PKA as suppressors of boi-induced fecundity arrest (Figure 2A,C). Reduced expression of each gene also suppressed Ptc-induced autophagy (Figure 5E,F). Conversely, increased expression of the autophagy inhibitory kinase TOR suppressed Ptc-driven autophagy induction (Figure 5E,F). Thus, Ptc induces autophagy by activating a PKA-LKB1 kinase cascade upstream of the core autophagy regulators that is opposed by increased expression of the inhibitory kinase TOR.

Autophagy Increases with Age and is Ptc-dependent

Autophagy induction in FSCs derived from WT flies increased steadily over a 4-week timecourse, demonstrating an age-dependence (Figure 5G, Table S1). This effect was abrogated by reducing ptc expression throughout the animal (w; ptc^tufl/ptc^tuf1) or just in FSCs (boi^e; UAS-ptc^RNAi/109-30-Gal4) (Figure 5G, Table S2). Similarly, autophagy was suppressed at three weeks of age when a dominant negative form of PKA (PKA^DN) was expressed in FSCs throughout the timecourse (Figure 5G, Table S2). Together, these results suggest that the same signaling cassette that controls the autophagy response to constitutive Hh is utilized to control autophagy during aging in WT flies.

Insulin Signaling Suppresses Hh-dependent Autophagy in FSCs

Based on our results, Hh should induce proliferation and autophagy simultaneously in FSCs. However, autophagy was not observed in FSCs in young, well-fed, WT flies (Figure 1), suggesting the existence of a signal that suppresses Hh-induced autophagy. A critical role for amino acids/protein in extending reproductive longevity in female flies (Grandison et al., 2009) suggests that signals induced by protein ingestion might complement cholesterol-dependent Hh signaling to suppress autophagy without affecting proliferation, resulting in improved FSC lifespan and delayed fecundity arrest. Insulin/Insulin-like growth factor Signaling (IIS) is an attractive candidate for several reasons. First, IIS is a primary pathway induced by dietary protein (Buch et al., 2008). Secondly, IIS has been implicated in reproductive aging, with increased production of Insulin-like peptides rescuing age-dependent fecundity arrest (Tatar, 2010). Finally, IIS is a known suppressor of autophagy, acting to initiate a Rheb GTPase-dependent signaling cascade that maintains TOR activity (He and Klionsky, 2009).

IIS was necessary for autophagy suppression under conditions of normal Hh signaling, as reduced expression in FSCs of the Insulin Receptor (InR), Akt, or TOR relieved autophagy inhibition (Figure 6A,C Table S2). Conversely, IIS was sufficient to suppress autophagy. Expression of WT TOR or activated forms of InR, Akt, or Rheb (Patel et al., 2003) reduced autophagy in FSCs in boi mutants (Figure 6B,C Table S2). Modulation of IIS levels also impacted autophagy levels in flies with WT levels of Hh signaling. Activated forms of InR, Akt, and overexpression of TOR reduced autophagy in FSCs isolated from 3-week old WT females (Figure 6D,E). Enhanced expression of dIlp2, the primary IIS ligand implicated in fecundity regulation (Burn et al., 2015; Hsu and Drummond-Barbosa, 2009; LaFever and Drummond-Barbosa, 2005), during the third week of life effectively suppressed autophagy induction (Figure 6D,E). Most likely, the normal age-dependent decline of dIlp2 (Alic et al., 2011; Flatt et al., 2008; Kwon et al., 2015; Morris et al., 2012) was rescued by this approach.

Figure 6. Insulin Signaling Suppresses Hh-induced Autophagy Induction.

(A, B) Fraction of germaria with ltr positive FSCs, comparing WT versus mutants expressing indicated RNAis (A, 109-30 Gal4/+; RNAi/+, * P= 0.0001 vs. WT. Error bars = SEM) or boi^e versus boi^e mutants overexpressing Insulin Receptor (InR^A1), AKT (mys-AKT), TOR or Rheb (UAS-Rheb.Pa) (B, boi^e; 109-30 Gal4/+; RNAi/+), * P=< 0.0001 vs. boi^e). Error bars = SEM.

(C) Representative images for panel (A and B) flies (ltr, (red), follicle cells (Fas III, green) and germ cells (Vasa, blue).

(D) Fraction of germaria with ltr positive FSCs, for 3 week old WT vs. mutants expressing dIlp2 (hs-dIlp2) under heat shock control (maintained at 29°C), Ga l4-induced activated InR (InR^A1), myristoylated Akt, or Tor. (109-30 Gal4/+; transgene/+), *, P<= 0.0002 vs. WT. Error bars = SEM.

(E) Representative images for panel D, ltr (red), follicle cells (Fas III, green) and germ cells (Vasa, blue).

(F) Hh activates two pathways in FSCs. Smo-Ci activation drives FSC proliferation and Ptc induces autophagy downstream of Hh ligand. We propose that Ptc activates a PKA-LKB1-AMPK kinase cascade to induce autophagy and that IIS inhibits autophagy through activation of Rheb/TOR. Collectively, Hh and IIS determine the autophagy-proliferation balance in FSCs.

See also Table S2.

DISCUSSION

The critical role of diet in determining organismal lifespan and fertility is well documented and seemingly universal. Conserved signaling pathways including IIS and AMPK mediate this response, in part through stem cell regulation (Longo et al., 2015; Lopez-Otin et al., 2013; Shim et al., 2013; Signer and Morrison, 2013). Here, we find that opposing action of the diet-responsive growth factors dIlp2 and Hh modulates autophagy and proliferation within somatic epithelial stem cells to determine their lifespan. Our results are consistent with a model in which feeding stimulates activation of the Hh and IIS pathways in FSCs simultaneously, maintaining FSCs in a competitive, proliferative state that promotes niche retention and robust egg production (Figure 6F and Graphical Abstract). Shifts in this equilibrium caused by increasing Hh signaling or decreasing IIS lead to autophagy induction in proliferative FSCs, an untenable dual state that reduces FSC lifespan by promoting differentiation (Graphical Abstract).

Importantly, we find that this same mechanism is critical for determination of FSC lifespan during aging. Whereas autophagy is not prevalent in FSCs in young females, it is dramatically upregulated in middle-aged (3-4 week) females (Figure 4D), a timepoint that immediately precedes fecundity arrest (Figure 1). Since cholesterol acts directly as a signal transduction molecule to induce Hh release, it is likely that similar levels of Hh pathway activity are maintained through aging if the diet composition remains the same. In contrast, IIS is influenced by multiple stimuli that alter systemic and local levels of ligand as well as mechanisms that limit IIS signaling within tissues (Graham and Pick, 2017). Previous work demonstrated development of IIS resistance in aging flies, a process mediated by increased expression of IMPL2, which inhibits IIS activation via ligand sequestration (Alic et al., 2011; Kwon et al., 2015; Morris et al., 2012). The resulting reduction in IIS and egg production with age can be overcome by ectopic expression of Insulin-like peptides (LaFever and Drummond-Barbosa, 2005) or by feeding a diet with appropriate levels of amino acids (Grandison et al., 2009) that likely act by boosting ligand levels to overwhelm the sequestration mechanism. Adding to these prior observations, we found that increasing dIlp2 or activating the IIS pathway via expression of InR, Akt, or TOR suppressed Hh-induced and age-dependent autophagy (Figure 6). Most likely, the diminishing levels of IIS over time dictate the relative equilibrium between autophagy and proliferation in FSCs to determine their lifespan (Graphical Abstract).

A surprising finding of our study is that Hh-mediated induction of autophagy is Smo- and Ci-independent. Smo and Ci are necessary for FSC proliferation (Figure 4, (Forbes et al., 1996; Hartman et al., 2010; Zhang and Kalderon, 2000, 2001), but not autophagy (Figure 4A,B, Table S2). Instead, Ptc and PKA are positive regulators of the autophagy induction pathway (Figures 3,5). Both proteins are well-studied negative regulators of Ci-dependent events, with PKA phosphorylating Ci to promote a repressive cleavage event, and Ptc suppressing Smo activity in the absence of Hh ligand (Lee et al., 2016). Recent work has demonstrated more complex roles for these proteins in Hh signal transduction, however. In the presence of Hh, PKA is required to directly phosphorylate Smo, promoting Ci activation (Lee et al., 2016; Li et al., 2014). Here, we find a second positive effector role for PKA, as its catalytic subunit is necessary for Hh-induced autophagy (Figure 3L). Known functions for PKA as a positive regulator of LKB1 activity (Mirouse et al., 2007b) suggest that PKA may link Hh to the core autophagy pathway via LKB1 activation.

Our results support a positive signaling role for Ptc in autophagy induction. Ptc overexpression also can induce autophagy in fly fat body cells (Jimenez-Sanchez et al., 2012), supporting the idea that Ptc-mediated autophagy is a conserved function. A major distinction between the two processes is the dependence of the signaling on Smo-Ci signaling. Ptc-mediated autophagy induction in the fat body relies on inhibition of Smo-Ci signaling, a mechanism that is dispensable in FSCs (Figures 3,5). Our data are consistent with a model in which the Ptc C-terminal tail and SSD contribute to autophagy induction, perhaps via direct regulation of PKA activity. Alternatively, Ptc may regulate autophagy indirectly. In the developing wing, Ptc promotes translocation of the RNA splicing protein Sex lethal (Sxl) to the nucleus, regulating gene expression independently of Smo (Horabin et al., 2003). If conserved in FSCs, this mechanism might influence Ptc-mediated autophagy induction via translational regulation. Interestingly, Ptc orthologs exist in organisms that lack smo in the genome (e.g. C. elegans and Monosiga brevicollis (Burglin and Kuwabara, 2010; King et al., 2008)), strongly implying Smo-independent functions for Ptc in development.

Previous work has shown that Ptc can function as a dependence receptor in mammalian cells, triggering apoptosis in the absence of Hh ligand (Mille et al., 2009; Thibert et al., 2003). In contrast, Hh ligand is both necessary and sufficient for autophagy induction in FSCs based on several criteria. First, increasing expression of Hh in apical cells in young flies induces autophagy in FSCs (Figures 1, 3), a phenotype that correlates with accelerated FSC loss during aging (Figure 1). Secondly, autophagy is induced at high levels in boi mutants, which release Hh constitutively (Figures 1-3, (Hartman et al., 2010)). The effects of boi mutation on autophagy are suppressed by reducing Hh expression in apical cells (Figure 5A), demonstrating that the observed effects depend on boi-mediated Hh release. Thirdly, the accelerated FSC loss caused by increased Hh ligand is suppressed back to a wild-type timecourse by reducing expression of a core autophagy component (Figure 3M). This strongly supports the notion that excess Hh ligand induces autophagy to elicit the observed effects. Finally, overexpression of a mutant form of Ptc that cannot bind directly to Hh (Ptc^Δloop2) induces autophagy in FSCs, but not as well as wild-type Ptc (Figure 5B). This suggests that ligand binding may induce a conformational change that is partially mimicked by the deletion, resulting in Ptc-mediated autophagy induction. Thus, Ptc has an active, ligand-dependent signaling role in autophagy induction in FSCs, rather than a passive role as a dependence receptor. It will be interesting to learn whether autophagy induction is an ancient function for Ptc that promotes Hh-ligand-dependent control of developmental events in many cell types, including in organisms that lack Smo.

STAR METHODS

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Alana M. O’Reilly (alana.oreilly@fccc.edu).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Fly Strains and Preparation

All fly stocks were raised on standard food available from the in-house facility at Fox Chase Cancer Center that contained water (42 liters), agar (296 grams), cornmeal (2000 grams), yeast (824 grams), molasses (2000 ml), tegosept (450 ml) and propionic acid (160 ml). Flies were maintained at standard 25°C, additional fly stocks were maintained at 18°C temperature-controlled incubator s. For temperature permissive experiments flies were either heat shocked at 37°C in a water bath and/or placed at 27 or 29°C temperature controlled mini-incubators. All flies were kept in yeasted vials (except during starvation conditions) with males at all times. Young female flies (age week 1) were used in all except the aging experiments as indicated because Follicle Stem Cells only exist in females.

The following stocks were obtained from the Bloomington Drosophila Stock Center (Bloomington, IN): w¹¹¹⁸ (WT); y1 w*;P(GawB)109-30/CyO (RRID:BDSC_7023); boi^e01708 (boi^e) were generated by Exelixis and maintained by the Harvard stock center, boi^e is a loss-of-function allele and expresses only 0.2% of the wild-type full-length boi transcript in the fly ovary (Hartman et al., 2010). p(UAS-Atg8a.GFP)3 (RRID:BDSC_51656, (Scott et al., 2004)); ptc^tuf.1 ltd¹ (RRID:DGGR_105908, (Phillips et al., 1990)); p(neoFRT)42D ptc^S2/cyo (RRID:BDSC_6332, (Simcox et al., 1989)); Pka-C1^B10 (RRID:BDSC_32018, (Lane and Kalderon, 1993)); Pka^DN (Pan and Rubin, 1995); p(w[+mC]=UAS-stg.N) 16/cyo, (RRID:DGGR_107872); p(UAS-AMPKalpha.T184D)2/Cyo (RRID:BDSC_32110, (Mirouse et al., 2007a)); UAS-ptc^1130X (RRID:BDSC_44612, (Johnson et al., 2000)); UAS-ptc (RRID:BDSC_5817, (Johnson et al., 1995)); p(UAS-Rheb.Pa) (RRID:BDSC_9689, (Patel et al., 2003)); p(UAS-InR^A1325D) (RRID:BDSC_8263, (Parks, 2004)); p(UAS-Tor.WT), (RRID:BDSC_7012, (Hennig and Neufeld, 2002)); p(UAS-myr-Akt1.V), (RRID:BDSC_50758, (Birnbaum, 2013)); UAS-hhN-GFP (Hartman et al., 2013); hs-dilp2 (RRID:BDSC_37472, (Brogiolo et al., 2001; Rulifson et al., 2002)); P(UAS-unc-51/Atg1.WT) (RRID:BDSC_60734, (Toda et al., 2008)); UAS-ptc^Δloop2 (gift from G. Struhl, (Casali and Struhl, 2004)); UAS-attB-smo^C320A (gift from S. Ogden, (Carroll et al., 2012) and P{hh^hs.PI} (FBal0031488).

The following RNAi lines were used: Atg1p(TRiP.JF02273)attP2, (RRID:BDSC_26731); Atg2 p(TRiP.JF02786)attP2, (RRID:BDSC_27706); Atg5 p(TRiP.HMS01244)attP2, (RRID:BDSC_34899); Atg6 p(TRiP.JF02897)attP2, (RRID:BDSC_28060); Atg7 p(TRiP. HMS01358)attP2 (RRID:BDSC_34369); Atg8B p(TRiP.JF02706)attP2, (RRID:BDSC_27554); Atg9 p(TRiP.JF02891)attP2, (RRID:BDSC_28055); and Atg13 p(TRiP. HMS02028)attP2, (RRID:BDSC_40861); smo (p(UAS-smo^RNAi)2, (RRID:BDSC_24472); ci p(TRiP.JF01715) attP2, (RRID:BDSC_28984); hh P(TRiP. JF01804) attP2, (RRID:BDSC_25794); lkb1 P(TRiP. HMS01351) attP2, (RRID:BDSC_34362); SNF4Agamma/ AMP protein kinase P(TRiP. GL00252) attP2, (RRID:BDSC_35341); PKA-C1 P(TRiP. JF01218) and ptc p(TRiP.HMC03872), (RRID:BDSC_55686); InR p(TRiP. JF01482) attP2, (RRID:BDSC_31037); tor p(TRiP.HMS00735)attP2 (RRID:BDSC_32941, (Ni et al., 2011)); akt1 p(TRiP. HM04007) attP2, (RRID:BDSC_31701); rheb p(TRiP. HMS00923) attP2 (RRID:BDSC_33966); boi (boi^GD3060) (RRID:FlyBase_FBst0460064).

Expression of transgenes or RNAi in Layer 2 FSCs or apical cells was achieved using 109-30Gal4 (P{GawB}109-30) or babGal4 (P{GawB}bab1{Pgal4-2}) respectively. Heat-induced expression of transgenes was achieved using w*;P(GawB)109-30, tubGal80^ts in combination with indicated transgenes. Flies were raised at the restrictive temperature (18°C) or permissive temperature (29°C) to inactiva te Gal80 and activate Gal4 driven expression. Mosaic analysis with repressible cell marker (MARCM) stocks were generated by crossing Ub-RFP, Gal80 19AFRT Flp122/Y; UAS-nls-GFP; UAS-transgene/TM6b males to 19AFRT; 109-30-Gal4/CyO females (Hartman et al., 2015). Constitutive activation of the hh gene (Ub-RFP, Gal80 19AFRT Flp122/19AFRT; UAS-nls-GFP/109-30-Gal4; hs-hh/+) and mitotic clones were induced in the ovary of the adult progeny via heat shock at respective time points for 2 h at 37°C. After the heat shock, female flies were kept at 25°C in yeasted vials with males. Flies were flipped into fresh vials after 4 days for time-course experiments before the ovaries were isolated for analysis.

The Lines Used in Each Figure Are Listed Below

Figure 1B, 1C, 1D (flies were aged accordingly as shown and described in method details), and S1: w¹¹¹⁸ (WT), boi^e01708 (boi^e), UAS-stg.N16 (w;String). Figure 1E (flies were aged accordingly as shown and described in method details): P{hh^hs.PI} (hsHh). Figure 1F, 1G (MARCM labeling) (flies were aged accordingly as shown): Ub-RFP, Gal80 19AFRT Flp122/19AFRT; UAS-nls-GFP/109-30-Gal4 (control); Ub-RFP, Gal80 19AFRT Flp122/19AFRT; UAS-nls-GFP/109-30-Gal4; hs-hh/+

Figure 2A: w¹¹¹⁸ (WT), boi^e, boi^e; 109-30Gal4; RNAi of Atg1: P{ TRiP.JF02273}attP2, boi^e; 109-30Gal4; RNAi of Atg6: P{ TRiP.JF02897}attP2, boi^e; 109-30Gal4; RNAi of Atg8b: P{ TRiP.JF02706}attP2, boi^e; 109-30Gal4; RNAi of Atg9: P{ TRiP.JF02891}attP2. Figure 2C: w¹¹¹⁸ (WT); boi^e01708; boi^e; 109-30Gal4; RNAi of lkb1: P{TRiP. HMS01351}attP2, boi^e; 109-30Gal4; RNAi of PKA-C1 P{ TRiP. JF01218}attP2.

Figure 3A, 3A’, 3A”, 3B, 3B’, 3B”, 3C and S2: w¹¹¹⁸ (WT), boi^e. Figure 3D, 3E, and S2 : hsHh. Figure 3F, 3F’, 3F”, 3G, 3G’, 3G”, 3H, and S2: w; 109-30Gal4; UAS-Atg8a.GFP and boi^e; 109-30Gal4; UAS-Atg8a.GFP. Figure 3I and 3J (western blot for Ref(2)P) (1 week old flies were starved for 3 days or kept under fed conditions before protein preparation): w¹¹¹⁸ ; boi^e. Figure 3K: w¹¹¹⁸ ; boi^e, (transgene expressed in FSCs using w; 109-30Gal4): boi^RNAi; Atg2^RNAi P{ TRiP. JF02786}attP2), Atg8b^RNAi P{ TRiP. JF02706}attP2), Atg9^RNAi P{ TRiP.JF02891}attP2), UAS-hhN-GFP, boi^e; 109-30Gal4, boi^e; Atg1 ^RNAiP{ TRiP. JF02273}attP2), boi^e; Atg6^RNAi P{ TRiP. JF02897}attP2), boi^e; Atg9^RNAi P{ TRiP.JF02891}attP2), (transgene expressed in apical cells using w; babGal4): UAS-hhN-GFP and boi^RNAi. Figure 3L: w¹¹¹⁸ (WT), w;109-30Gal4; UAS-AMPKalpha.^T184D, w;109-30Gal4; UAS-unc-51/Atg1, boi^e, boi^e;109-30Gal4; RNAi of lkb1: P{ TRiP. HMS01351}attP2, boi^e;109-30Gal4; RNAi of SNF4Agamma/AMPK: P{ TRiP. GL00252}attP2, and boi^e;109-30Gal4; Pka-C1^B10. Figure 3M (MARCM labeling) (flies were aged accordingly as shown): (control) Ub-RFP, Gal80 19AFRT Flp122/19AFRT; UAS-nls-GFP/109-30-Gal4, (Hh-gof) Ub-RFP, Gal80 19AFRT Flp122/19AFRT; UAS-nls-GFP/109-30-Gal4; hs-hh/+, (AMPK^T184D) Ub-RFP, Gal80 19AFRT Flp122/19AFRT; UAS-AMPKalpha.^T184D/109-30-Gal4; UAS-nls-GFP/+, (Hh-gof; Atg5^RNAi) Ub-RFP, Gal80 19AFRT Flp122/19AFRT; UAS-nls-GFP/109-30-Gal4; Atg5^RNAi{ TRiP. HMS01244}/+.

Figure 4A, 4B, 4C, 4C’, 4I (flies were aged accordingly as shown), 4K, 4K’ (3-week-old female flies were used), S1 and S2: w¹¹¹⁸ (WT). Figure 4A, 4B’, 4F, 4F’ 4J (flies were aged accordingly as shown), 4L, and 4L’ (3-week-old female flies were used), S1 and S2: boi^e. Figure 4A, 4B”, 4E, 4E’, S1 and S2: w; 109-30Gal4; UAS-stg.N16. Figure 4A, 4H, 4H’, S1 and S2: boi^e, 109-30Gal4; UAS-stg.N16. Figure 4A, 4D, 4D’, S1 and S2: w; 109-30Gal4; UAS-smo.RNAi. Figure 4A, 4G, 4G’, S1 and S2: boi^e, 109-30Gal4; UAS-smo.RNAi.

Figure 5A and S2: w¹¹¹⁸ (WT), w; 109-30Gal4; UAS-smo.RNAi, w; 109-30Gal4/ptc^tuf-1; ptc^tuf-1, w; 109-30Gal4-ptc^tuf-1; 1/ptc^tuf-1; UAS-smo.RNAi, w;109-30tubGal80^ts/ptc^RNAi, w; 109-30babGal4/Hh-N-GFP; boi^e, boi^e, 109-30Gal4/UAS-smo. RNAi, boi^e; 109-30Gal4-ptc^tuf-1/ptc^{tuf- 1}, boi^e; 109-30Gal4-ptc^tuf-1/ptc^tuf-1; UAS-smo.RNAi, boi^e, 109-30Gal4/ptc.RNAi, boi^e; 109-30Gal4-ptc^tuf-1/ptc^tuf-1, boi^e, 109-53Gal4/hh.RNAi. Figure 5B: w;109-30tubGal80^ts. Figure 5B, 5C, 5C’, 5C”, 5D, 5D’, 5D” and S2 :w;109-30tubGal80^ts; UAS-smo^C320A, w;109-30tubGal80^ts/ptc^RNAi; UAS-smo^C320A, w; UAS-ptc/109-30tubGal80^ts, w;109-30tubGal80^ts/ UAS-ptc^Δloop2, w;109-30tubGal80^ts/ UAS-ptc^1130X. Figure 5E: w; UAS-ptc/109-30tubGal80^ts, w; UAS-ptc/109-30tubGal80^ts; PKA-C1^RNAi, w; UAS-ptc/109-30tubGal80^ts; LKB1^RNAi, w; UAS-ptc/109-30tubGal80^ts; AMPK^RNAi, w; UAS-ptc/109-30tubGal80^ts; Atg5^RNAi, w; UAS-ptc/109-30tubGal80^ts; UAS-tor. Figure 5F (western blot for Ref(2)P) (1 week old flies were kept under fed conditions before protein preparation): w; UAS-ptc/109-30tubGal80^ts, w; UAS-ptc/109-30tubGal80^ts; Atg5^RNAi, w; UAS-ptc/109-30tubGal80^ts; UAS-tor. Figure 5G (flies were aged accordingly as shown) and S2: w¹¹¹⁸ (WT), w; 109-30Gal4/ptc^tuf-1; ptc^tuf-1, w; 109-30tubGal80^ts/ptc^RNAi, w; 109-30tubGal80^ts/pka-C1^DN.

Figure 6A, 6C and S2: w¹¹¹⁸ (WT), w; 109-30Gal4/InR^RNAi, w;109-30tubGal80^ts/Akt1^RNAi, w;109-30tubGal80^ts/Tor^RNAi, w;109-30tubGal80^ts/Rheb^RNAi. Figure 6B, 6C and S2: boi^e, boi^e, 109-30Gal4; UAS-InR^A1, boi^e, 109-30Gal4; UAS-mys-AKT1, boi^e, 109-30Gal4; UAS-TOR, boi^e, 109-30tubGal80^ts/UAS-Rheb. Figure 6D, 6E and S2 (3-week-old female flies were used): w¹¹¹⁸ (WT), w; hs-dilp2, w;109-30tubGal80^ts/ UAS-InR^A1, w; 109-30tubGal80^ts/ UAS-mys-AKT1, w; 109-30tubGal80^ts/ UAS-TOR.

METHOD DETAILS

Immunofluorescence

Ovaries from 20-40 female flies were dissected in Grace’s insect cell culture medium (Gibco) using fine forceps for each experimental datapoint. Individual ovarioles were gently teased apart (Hartman et al. 2013). Ovaries were collected and fixed by gently rocking in 4% paraformaldehyde for 10 min followed by three 10min 1X PBST washes. Primary antibody mix was made in 5% BSA diluted with 1XPBST. Ovaries were incubated in primary antibody mix overnight at 4°C. Primary antibodies used were rabbit anti-VASA (1:2,000; Santa Cruz Biotechnology RRID: AB_793874), mouse anti-Fasciclin III (Fas III) (1:100; Developmental Studies Hybridoma Bank; Patel et al., 1987, RRID:AB_528238), chicken anti-GFP (1:1000; Invitrogen), or rabbit antiphospho-histone-H3 (PH3) (1:1,000; EMD Millipore: RRID:AB_10807285). Next day, ovaries were washed in 1X PBS, three times for 10mins, followed by secondary antibody incubation (made in 5% BSA diluted with 1XPBST) at RT for at least 2h in dark conditions. Secondary antibodies used were FITC, Cy3, and Cy5 conjugated to species-specific secondary antibodies (Jackson ImmunoResearch Laboratories, Inc. RRID:AB_2307443, RRID:AB_2315777, RRID:AB_2335588, RRID:AB_2337384, RRID:AB_2340607, RRID:AB_11180200). Ovaries were washed three times in 1X PBST and nuclei were stained with Draq5 (Cell Signaling Technology) (1:1000) for 10mins and washed in 1XPBS. Immunostained ovaries were mounted in Vectashield medium (Vector Laboratories) and coverslips were sealed using nail polish and left to dry before storing at −20°C.

Proliferation Assay

Ovaries were dissected/experiment (200-500 germaria) in Grace’s insect medium and stained with rabbit antiphospho-histone-H3 (PH3) (1:1,000; EMD Millipore: RRID:AB_10807285). After completing the immunofluorescence procedure described above, the number of dividing FSCs was quantified by scoring germaria for PH3-positive FSCs per germarium, divided by the total number of germaria (Hartman et al., 2010) (also see table S1).

LysoTracker Staining

10-50 flies were dissected/experiment (between 25-200 germaria) in Grace’s insect medium, incubated for five minutes in 100μM LysoTracker Red DND 99 (Invitrogen, Thermo Fisher Scientific), washed in PBS, and fixed in 4% paraformaldehyde for 10 min (modified from (Barth et al., 2011; DeVorkin and Gorski, 2014)). Lysotracker staining was quantified by scoring germaria containing Layer 2 FSCs positive for lysotracker staining divided by the total number of germaria (Nystul and Spradling, 2007) (also see table S2).

Image Analysis

After mounting, images were collected at room temperature (~22°C) using 40× (1.25 NA) or 20× (0.7 NA) oil immersion lenses (Leica) on an upright microscope (DM 5000; Leica) coupled to a confocal laser scanner (TCS SP5; Leica). LAS AF SP5 software (Leica) was used for data acquisition. Images representing individual channels of single confocal slices including FSCs in each germarium were exported as TIFF files, and images were converted to figures using Photoshop software (Adobe).

Nutrient Restriction Assays

Flies were subjected to nutrient restriction by rearing on fruit juice plates (50% grape juice, 3% bactoagar, 1% glacial acetic acid, and 1% methyl paraben) for 1-3 days in starvation chambers.

Egg Laying Assay

Virgin females were housed with males and aged for appropriate time-points in single vials (5 females per vial). For each independent experiment, the number of eggs per genotype was counted over a 24-hour period after aging for 1-6 weeks, and eggs laid/fly/week were quantified.

Fertility Screen

Five virgin females were placed in a vial with 3 males (day 0). Adult flies were transferred into a new vial every 3-4 day before eclosion of progeny. The number of days until female flies stopped producing viable eggs were counted from day 0 per genotype and recorded as days until oogenesis arrest.

Protein Isolation and Western Blot

Protein was isolated from whole ovaries from 10-40 dissected female flies, and ovaries were lysed using a dounce homogenizer in RIPA buffer (50 mM Tris, pH 8.0, 0.1% SDS, 1% Triton X-100, 150 mM NaCl, 1% deoxycholic acid, 2 mM PMSF, and protease inhibitors [Sigma-Aldrich]). Lysates were further passed through a 27^1/2-gauge needle, at least 5 times. Lysates were kept on ice during the isolation procedure. Protein lysates were run on a 10% SDS-PAGE gel, then transferred to nitrocellulose membrane (Amersham). Membranes were probed with anti-Ref(2)P (1:100) (Abcam) and β-actin (1:2000, Abcam) antibodies. Band intensities were quantified using ImageJ (Fiji). β-actin was used for normalization.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical Analysis

All the statistical details can be found in figures (error bars), figure legends and results (p values and comparisons), tables S1 and S2 provide actual numbers (n= the number of germaria scored/genotype), averages, fractions and which genotypes were used as controls as comparisons to experimental genotypes to yield a p-value, respectively. Significant differences between control and experimental genotypes were determined using Student’s unpaired t test for two samples using Prism (GraphPad), with significance determined at p ≤ 0.05. N is defined as the number of germaria scored for each genotype, averaged among multiple experimental repeats to yield values shown in Tables S1 and S2. Standard Error of the Mean is indicated by error bars in all figures except for Figure 1D, where error bars represent Standard Deviation. Supplementary Tables 1 and 2 include quantitation of error for data represented bar graphs.

Supplementary Material

Supplementary Table 1:

Quantification of PH3 staining in FSCs (region 2A/2B). Related to Figure 1 and Figure 4.

Supplementary Table 2:

Quantification of Lysotracker staining in FSCs (region 2A/2B). Related to Figures 3-6.

Highlights.

• Constitutively active Hh signaling induces autophagy in FSCs

• Age-associated autophagy is Smo-independent and Ptc-dependent

• Balance between Hh and IIS signaling maintains FSC lifespan