Tolerance to Hypoxia Is Promoted by FOXO Regulation of the Innate Immunity Transcription Factor NF-κB/Relish in Drosophila

Our cells and organs need oxygen from the air we breathe in order to survive. However, although air contains twenty percent oxygen, our cells are exposed to much lower levels....

Abstract

Exposure of tissues and organs to low oxygen (hypoxia) occurs in both physiological and pathological conditions in animals. Under these conditions, organisms have to adapt their physiology to ensure proper functioning and survival. Here, we define a role for the transcription factor Forkhead Box-O (FOXO) as a mediator of hypoxia tolerance in Drosophila. We find that upon hypoxia exposure, FOXO transcriptional activity is rapidly induced in both larvae and adults. Moreover, we see that foxo mutant animals show misregulated glucose metabolism in low oxygen and subsequently exhibit reduced hypoxia survival. We identify the innate immune transcription factor, NF-κB/Relish, as a key FOXO target in the control of hypoxia tolerance. We find that expression of Relish and its target genes is increased in a FOXO-dependent manner in hypoxia, and that relish mutant animals show reduced survival in hypoxia. Together, these data indicate that FOXO is a hypoxia-inducible factor that mediates tolerance to low oxygen by inducing immune-like responses.

OXYGEN is essential for the normal growth, development, and functioning of tissues and organs. However, while the air we breathe contains ∼20% oxygen, even under healthy physiological conditions, our cells and tissues receive considerably lower levels. These can be anywhere from 1 to 10% oxygen depending on the tissue (McKeown 2014). Hence, our tissues and organs need to function and maintain homeostasis at low levels of oxygen. This aspect of normal physiology is often neglected in tissue culture experiments where cells are routinely maintained in 20% oxygen. In addition, many pathologies such as heart disease, stroke, and chronic lung disease are characterized by severe oxygen deprivation (hypoxia) (Semenza 2011). This hypoxia has deleterious effects on tissue metabolism and function, and can lead to death. Understanding how cells, tissues, and organisms adapt to low oxygen is therefore an important question in biology.

One central hypoxic mechanism involves induction of the hypoxia-inducible factor (HIF)-1α transcription factor, which can control the expression of a diverse array of target genes that maintain cellular homeostasis in low oxygen (Semenza 2014). The importance of HIF-1α has been shown by loss-of-function genetic analyses in model organisms such as Caenorhabditis elegans, Drosophila, and mice. For example, in C. elegans and Drosophila, which are normally quite hypoxia-tolerant, HIF-1α mutants die when exposed to low oxygen (Jiang et al. 2001; Centanin et al. 2005; Li et al. 2013). Tissue-specific mouse knockouts have also shown how HIF-1α can control organ-level and whole-body adaptation to low oxygen in both physiological and pathological conditions (Schipani et al. 2001; Cramer et al. 2003; Tomita et al. 2003; Huang et al. 2004; Mason et al. 2004; Boutin et al. 2008). However, compared to our understanding of HIF-1α, less is known about other transcription factors that are important in mediating hypoxia adaptation in animals.

The conserved transcription factor Forkhead Box-O (FOXO) is an important mediator of adaptation to stress in animals (Webb and Brunet 2014). Studies in Drosophila have provided important insights into the role of FOXO as a regulator of organismal physiology. Here, different environmental stressors—such as starvation, oxidative stress, pathogens, and ionizing radiation—have been shown to induce FOXO transcriptional activity (Junger et al. 2003; Dionne et al. 2006; Karpac et al. 2009, 2011; Borch Jensen et al. 2017). Once induced, FOXO then directly controls the expression of an array of metabolic and regulatory genes that together function to maintain organismal homeostasis and survival (Gershman et al. 2007; Teleman et al. 2008; Alic et al. 2011; Birnbaum et al. 2019). Indeed, genetic upregulation of FOXO is sufficient to promote stress resistance in Drosophila, and it is one of the most effective ways to extend life span (Giannakou et al. 2004; Hwangbo et al. 2004; Kramer et al. 2008; Demontis and Perrimon 2010; Alic et al. 2014).

In this paper, we report our work using Drosophila melanogaster to explore hypoxia tolerance. In their natural ecology, Drosophila grow in rotting, fermenting food rich in microorganisms, an environment likely characterized by low ambient oxygen (Callier et al. 2015; Markow 2015; Harrison et al. 2018). Probably as a consequence of this environment, they have evolved mechanisms to tolerate hypoxia (Centanin et al. 2008; Li et al. 2013; Lee et al. 2019). Here, we explore whether induction of FOXO is one such mechanism. Previous work has shown that reduced insulin signaling and FOXO induction confers hypoxia tolerance in C. elegans and zebrafish (Scott et al. 2002; Mendenhall et al. 2006; Menuz et al. 2009; Liu et al. 2016). Moreover, the mammalian FOXO homolog FOXO3a can be induced in cell culture upon hypoxia exposure, where it regulates metabolic responses and cell death (Bakker et al. 2007; Jensen et al. 2011). In this paper, we show that FOXO is required for hypoxia tolerance in Drosophila, and that it functions by regulating the immune transcription factor NF-κB. Thus, the induction of FOXO is a conserved mechanism of hypoxia tolerance in animals.

Materials and Methods

Drosophila stocks

Flies were raised on medium containing 150 g agar, 1600 g cornmeal, 770 g Torula yeast, 675g sucrose, 2340 g D-glucose, and 240 ml acid mixture (propionic acid/phosphoric acid) per 34 liter water and maintained at 25°, unless otherwise indicated. The following fly stocks were used: w¹¹¹⁸, sima^KG07607/TM3,Ser,GFP [Bloomington Drosophila Stock Center (BDSC) #14640] (Centanin et al. 2008), Df (3R)X3F (BDSC #2352), foxo^Δ94/TM3,Ser, GFP (Slack et al. 2011), thor-LacZ (BDSC #9558) (Bernal and Kimbrell 2000), Relish^E38 (BDSC #9458) (Hedengren et al. 1999) Relish^E20 (BDSC #9457) (Hedengren et al. 1999) hsflp; UAS-dp110 (Britton et al. 2002), act > CD2 > Gal4,UAS-GFP (Britton et al. 2002), UAS-Fatiga RNAi (VDRC #103382), UAS-PTEN (Britton et al. 2002), and daughterless-GSG (Sun et al. 2014), r4-GAL4 (BDSC #33832).

Hypoxia exposure

For all hypoxia experiments, vials containing Drosophila were placed into an airtight glass chamber into which a premix of 5% oxygen/95% nitrogen, 1% oxygen/99% nitrogen, or 100% nitrogen continually flowed (Lee et al. 2019). Flow rate was controlled using an Aalborg model P gas flow meter. Alternatively, for some experiments, Drosophila vials were placed into a Coy Laboratory Products in vitro O₂ chamber that was maintained at fixed oxygen levels of 1% or 5% by injection of nitrogen gas.

Immunofluorescence staining

Larvae were inverted using fine forceps in 1× PBS. Inverted larvae were fixed in 8% paraformaldehyde for 30 min, washed in 1× PBS/0.1% Triton X-100 (PBST), and blocked for 2 hr at room temperature in 1× PBS/0.1% Tween 20/1% bovine serum albumin (PAT). Larvae were then incubated overnight with primary antibody diluted in PAT at 4°, washed three times with 1× PBS with 3% Triton X-100 (PBT) and 2% fetal bovine serum (FBS), and incubated with secondary antibody diluted 1:4000 in PBT with FBS for 2 hr at room temperature. Larvae were washed with PBT and stained with 1:10,000 Hoechst 33342 dye for 5 min, then washed three times more with PBT. Larval tissues were isolated using fine forceps and then mounted on glass slides with cover slips using Vectashield mounting media (Vector Laboratories, Burlingame, CA). The rabbit anti-FOXO antibody was used at 1:500 dilution (a gift from Marc Tatar). Goat anti-rabbit Alexa Fluor 568 (Invitrogen, Carlsbad, CA) was used as the secondary antibody. Hoechst 33,342 (Invitrogen) was used to stain nuclei. Quantification of FOXO nuclear staining was done by scoring cells with prominent nuclear FOXO staining, as observed in the representative images of hypoxia-treated (5 and 1% oxygen) fat-body cells in Figure 1A.

Figure 1.

Hypoxia induces FOXO activity. (A) Left, FOXO staining of 96-hr AEL w¹¹¹⁸ larval fat bodies following exposure to hypoxia for 2 h. Nuclei are stained with Hoechst (bottom panels). Bar, 25 μm. Right, quantification of FOXO nuclear localization in fat-body cells. n = total number of cells analyzed. (B–D) 4e-bp mRNA levels measured by qRT-PCR in control (w¹¹¹⁸) and foxo mutant (foxo^Δ94) following (B) 6 hr of 5% O₂ hypoxia in larvae, (C) 6 hr of 1% O₂ hypoxia in larvae, or (D) 16 hr of 1% O₂ hypoxia in adults. n > 6 cohorts of animals per condition. Data represent mean + SEM. *P < 0.05, two-way ANOVA followed by post hoc Student’s t-test. (E) LacZ staining in tissues of thor-LacZ larvae following 2-hr exposure to 5% O₂. Bar, 100 μm. AEL, after egg laying; FOXO, Forkhead Box-O; mRNA, messenger RNA; norm, normoxia; qRT-PCR, quantitative RT-PCR.

Quantitative PCR

Total RNA was extracted from either 96-hr after egg laying (AEL) larvae or from 1-week-old mated female adults, using TRIzol according to the manufacturer’s instructions (catalog number 15596–018; Invitrogen). RNA samples were then subjected to DNase treatment according to the manufacturer’s instructions (2238 G; Ambion) and reverse transcribed using Superscript II (catalog number 100004925; Invitrogen). The generated complementary DNA was used as a template to perform quantitative (q)RT-PCRs (ABI 7500 real time PCR system using SyBr Green PCR mix) using specific primer pairs. PCR data were normalized to β-tubulin levels. Each experiment was independently repeated a minimum of three times. The following primers were used:

• β-tubulin: forward 5′-ATCATCACACACGGACAGG-3′; reverse 5′ GAGCTGGATGATGGGGAGTA-3′.

• 4e-bp: forward 5′-GCTAAGATGTCCGCTTCACC-3′; reverse: 5′ CCTCCAGGAGTGGTGGAGTA-3′.

• relish: forward 5′-TCCTTAATGGAGTGCCAACC-3′; reverse 5′-TGCCATGTGGAGTGCATTAT-3′.

• dorsal: forward 5′-TGTTCAAATCGCGGGCGTCGA-3′; reverse 5′-TCGGACACCTTCGAGCTCCAGAA-3′.

• dif: forward 5′-CGGACGTGAAGCGCCGACTTG-3′; reverse 5′ CAGCCGCCTGTTTAGAGCGG-3′.

• attacin A: forward 5′-AGGAGGCCCATGCCAATTTA-3′; reverse 5′ CATTCCGCTGGAACTCGAAA-3′.

• cecropin A: forward 5′-TCTTCGTTTTCGTCGCTCTCA-3′; reverse 5′ ATTCCCAGTCCCTGGATTGTG-3′.

• ldh: forward: 5′-AGATCCTGACTCCCACCGAA-3′; reverse: 5′-GCCTGGACATCGGACATGAT-3′.

• fatiga: forward: 5′-ATTGAGCCCAAGTTTGATCG-3′; reverse: 5′-AGCTGCCAGATTGTTCGTCT-3′.

• Def: forward: 5′-TGAAGTTCTTCGTTCTCGTGG-3′; reverse: 5′-CACCAGGACATGATCCTCTG-3′.

• Mtk: forward: 5′-CGATTTTTCTGGCCCTGCT-3′; reverse: 5′-CCGGTCTTGGTTGGTTAGGAT-3′.

• Cec C: forward: 5′-TCATCCTGGCCATCAGCATT-3′; reverse: 5′-CGCAATTCCCAGTCCTTGAAT-3′.

• Dros: forward: 5′-TTTGTCCACCACTCCAAGCAC-3′; reverse: 5′-ATGGCAGCTTGAGTCAGGTGA-3′.

Lac Z staining

Larvae were inverted using fine forceps in 1× PBS. Inverted larvae were fixed in 8% paraformaldehyde for 30 min, washed in PBST, and then incubated in 500 μl of an X-Gal solution containing 10 mM sodium phosphate buffer, pH 7.2, 150 mM NaCl, 1mM MgCl₂, 10 mM K₄[Fe^II(CN)₆], 10 mM K₃[Fe^III(CN)₆], and 0.1% Triton X-100 with 12.5 μl of 8% X-Gal solution (in DMSO) added immediately prior to incubation. Samples were then incubated at 37° until the X-Gal staining was visible.

Measurement of hypoxia survival

Larvae:

Newly hatched larvae were placed in food vials (50 larvae per vial) and then maintained in either normoxia or hypoxia (5% oxygen). Larvae exposed to hypoxia were maintained in this environment until ∼80% of larvae had pupated. Then, vials were removed from hypoxia and the numbers of eclosing adults were counted.

Adults:

First, 4–5 days posteclosion, mated female adults were placed into hypoxia (1% oxygen) for 24 hr in cohorts of 20 flies per vial. Then, vials were removed from hypoxia and the flies were allowed to recover for 48 hr before the numbers of dead flies were counted.

Starvation:

At 4–5 days posteclosion, mated female adults were subjected to starvation by transferring them from food vials to vials containing 0.4% agar/PBS for 24 hr. The numbers of dead flies were then counted.

Glucose, glycogen, trehalose, and TAG assays

Adult female Drosophila were either exposed to hypoxia (1% oxygen) for 16 hr or maintained in normoxia, and then frozen on dry ice. Colorimetric assays for each of the metabolites were then conducted using the methods described in detail in Tennessen et al. (2014).

Preparation of protein extracts and western blotting

First, 96-h AEL Drosophila whole larvae were lysed with a buffer containing 20 mM Tris-HCl (pH 8.0), 137 mM NaCl, 1 mM EDTA, 25% glycerol, and 1% NP-40 with the following inhibitors: 50 mM NaF, 1 mM PMSF, 1 mM DTT, 5 mM sodium orthovanadate (Na₃VO₄), and Protease Inhibitor cocktail (Roche, catalog number 04693124001) and Phosphatase inhibitor (Roche, catalog number 04906845001), according to the manufacturer’s instructions. Protein concentrations were measured using the Bio-Rad (Hercules, CA) Dc Protein Assay kit II (5000112). Protein lysates (15–30 μg) were resolved by SDS-PAGE and electrotransferred to a nitrocellulose membrane, subjected to western blot analysis with specific antibodies, and visualized by chemiluminescence [enhanced ECL solution (Perkin-Elmer)]. Primary antibodies used in this study were: anti-Akt (1:500 dilution; Cell Signaling, catalog number #9272), anti-pAkt-T342 (1:1000 dilution; gift from Michelle Bland), and anti-pAkt-S505 (1:1000 dilution; Cell Signaling, catalog number 4504). Goat secondary antibodies were purchased from Santa Cruz Biotechnology (sc-2030, 2005, and 2020). For experiments looking at Akt phosphorylation, total Akt levels were used as a loading control because the level of this protein was unaffected by hypoxia.

Statistical analyses

Data were analyzed by Student’s t-test or two-way ANOVA. All statistical analysis and data plots were performed using Prism software. In all figures, statistically significant differences are presented as * and indicate P < 0.05.

Data availability

All reagents are available on request. The authors affirm that all data necessary for confirming the conclusions of the article are present within the article, main figures, and supplemental figures. Supplemental material available at figshare: https://doi.org/10.25386/genetics.12429398

Results

Hypoxia induces FOXO activity

The main way that FOXO is regulated is through nuclear–cytoplasmic shuttling. To determine if hypoxia exposure could induce FOXO, we transferred third-instar larvae growing on food to either moderate (5% oxygen) or severe hypoxic environments (1% oxygen), and then stained for FOXO localization using an anti-FOXO antibody (Figure 1A). We saw that exposure to hypoxia caused FOXO relocalization from the cytoplasm to the nuclei of fat-body cells (Figure 1A and Supplemental Material, Figure S1A). This effect was rapid; nuclear relocalization occurred within 15 min of exposing larvae to hypoxia (Figure S1, B and C). We next examined the effects of hypoxia on the expression of 4e-bp, a well-characterized FOXO target gene. We measured messenger RNA (mRNA) levels of 4e-bp using qRT-PCR in whole third-instar larvae exposed to either 5 or 1% oxygen. We saw that 4e-bp levels were strongly increased in control (w¹¹¹⁸) larvae exposed to both hypoxic conditions (Figure 1, B and C). As with the FOXO nuclear localization, this increase in 4e-bp was rapid and was seen within 15–30 min following hypoxia exposure (Figure S1D). However, the hypoxia-induced increase in 4e-bp mRNA levels was largely abolished in foxo^Δ94, a deletion line that is a null mutant for the foxo gene (Slack et al. 2011) (Figure 1, B and C). We also examined the effects of hypoxia in adults. We exposed adult females to 1% O₂ and found that, as in larvae, 4e-bp levels were increased in control (w¹¹¹⁸) animals and that this effect was blunted in foxo mutants (Figure 1D). Finally, we examined the tissue pattern of 4e-bp induction by examining LacZ staining in thor-LacZ flies, which is a LacZ-enhancer trap in the 4e-bp gene locus (Bernal and Kimbrell 2000). We found that larvae exposed to 2 hr of 5% O₂ showed increased LacZ staining in the majority of larval tissues including the fat body, the intestine, and the body wall muscle (Figure 1E), suggesting that the hypoxia induction of FOXO activity is not tissue-restricted. Together, these data indicate that exposure to hypoxia in both Drosophila larvae and adults results in rapid induction of FOXO transcriptional activity.

FOXO is required for hypoxia tolerance

Is FOXO activation required for Drosophila survival in low oxygen? To find out, we measured hypoxia survival in foxo^Δ94 animals. Under standard laboratory conditions (rich food and normoxia) foxo mutant animals are viable (Slack et al. 2011). Therefore, we examined how well these mutants tolerate low oxygen. We first examined hypoxia survival in larvae. Control (w¹¹¹⁸) and foxo mutant embryos were allowed to develop in normoxia, and then newly hatched larvae were transferred to hypoxia (5% oxygen) for the duration of their larval period, before being returned to normoxia. We then counted the number of animals that developed to viable adults. We found that the foxo mutant animals reared in hypoxia had a significant decrease in viability compared to control animals (Figure 2A). In contrast, survival of foxo animals to adulthood in normoxia was no different to control animals (Figure S2A). We next examined hypoxia survival in female adults. Control (w¹¹¹⁸) and foxo mutant animals were exposed to either severe hypoxia (1% oxygen) for 24 hr or anoxia (0% oxygen) for 6 hr. After these low oxygen exposures, flies were returned to normoxia and the number of surviving animals counted. As observed in larvae, we found that the adult foxo mutant animals showed significantly decreased survival in both the hypoxic and anoxic conditions (Figure 2, B and C), while survival in normoxia over the same time periods was unaffected (Figure S2B). During severe hypoxia and anoxia, adult flies become immobile. However, when foxo adults were exposed to starvation instead of hypoxia for 24 hr there was no effect on viability, indicating that the decrease in hypoxia survival in foxo mutants is not simply a consequence of reduced nutrient intake as a result of immobility (Figure S2C). Together, our data indicate that FOXO activation is required for organismal survival in low oxygen in both developing larvae and adults.

Figure 2.

FOXO is required for hypoxia tolerance. (A) Control (w¹¹¹⁸) and foxo mutant (foxo^Δ94) animals were exposed to hypoxia (5% O₂) throughout their larval period, before being returned to normoxia as pupae. The percentage of flies that eclosed as viable adults was then counted. (B and C) Adult control (w¹¹¹⁸) or foxo mutant (foxo^Δ94) flies were exposed to either (B) 24 hr of 1% O₂ or (C) 6 hr of 0% O₂, before being returned to normoxia. The percentage of viable flies was then counted. Data represent mean + SEM. *P < 0.05, Student’s t-test. n > 4 cohorts of animals per condition. FOXO, Forkhead Box-O.

Cells, tissues, and organisms adapt to low oxygen by altering their metabolism (Semenza 2011). In particular, a key adaptation is the upregulation of glycolysis. Therefore, we checked whether FOXO might be important for controlling glucose metabolism in hypoxic animals. We first measured total glucose levels in adult animals exposed to hypoxia. Control animals exhibited a decrease in glucose levels after 16 hr of hypoxia (Figure 3A). foxo mutant flies had lower levels of total glucose in normoxia and these levels were even further depleted upon exposure to hypoxia (Figure 3A, left panel); however, the relative decrease in total glucose levels in hypoxia was similar between control and foxo animals (Figure 3A, right panel). In contrast, we saw a different pattern when we measured levels of glycogen (the stored form of glucose) and trehalose (the circulating form of glucose in Drosophila). As with total glucose, both glycogen and trehalose levels in normoxia were lower in foxo mutant animals compared to control animals, and the levels of both forms of glucose were decreased in hypoxia (Figure 3, B and C, left panels). However, foxo mutants showed a significantly greater decrease in both glycogen and trehalose in hypoxia compared to control animals (Figure 3, B and C, right panel).

Figure 3.

foxo mutants have altered glucose homeostasis in hypoxia. (A–C) Levels of free glucose (A), glycogen (B), or trehalose (C), in adult control (w¹¹¹⁸) and foxo mutant (foxo^Δ94) flies exposed to normoxia or 1% O₂ hypoxia for 16 hr. n = 15. Left panels indicate relative metabolite levels in normoxia and hypoxia. Right panels indicate the % change in levels in hypoxia. Data represent mean + SEM. Left panels, *P < 0.05, Student’s t-test following significant two-way ANOVA; right panels, *P < 0.05, Student’s t-test. (D) Ldh mRNA levels measured by qRT-PCR in control (w¹¹¹⁸) and foxo mutants (foxo^Δ94) following 16 hr of 1% O₂ hypoxia in adults. Data represent mean + SEM. *P < 0.05, two-way ANOVA followed by post hoc Student’s t-test. n > 10 per condition. FOXO, Forkhead Box-O; mRNA, messenger RNA; ns, not significant; qRT-PCR, quantitative RT-PCR.

Finally, we investigated expression of lactate dehydrogenase (ldh)—a key glycolytic enzyme—in w¹¹¹⁸ and foxo^Δ94 adult females. We saw that control animals had increased ldh mRNA levels when exposed to hypoxia, as has been reported before (Lavista-Llanos et al. 2002; Li et al. 2013) and is consistent with an upregulation of glycolysis. In contrast, foxo mutant animals had increased ldh levels in normoxia, and this expression increased significantly further in hypoxia (Figure 3D). Taken together, these data indicate that foxo mutants show deregulated control over normal glucose metabolism in hypoxia; they show overproduction of ldh and they exhibit a larger depletion of both stored and circulating glucose in hypoxia compared to control animals.

Hypoxia induces FOXO by inhibiting phosphoinositide 3-kinase/Akt signaling

We next examined how hypoxia induces FOXO. The best-studied cellular response to hypoxia involves induction of the HIF-1α transcription factor (called sima in Drosophila). HIF-1α induces expression of metabolic and regulatory genes required for hypoxia adaptation, and HIF-1α is required for organismal tolerance to low oxygen in both Drosophila and C. elegans (Jiang et al. 2001; Centanin et al. 2005). However, we found that FOXO was still relocalized to the nucleus upon hypoxia in fat-body cells from both sima^KG7607 homozygote mutant larvae (Figure 4A and Figure S3A) and sima^KG7607/Df larvae (Figure S3, B and C). We also found that fat-body knockdown of fatiga, a prolyl hydroxylase that functions in normoxia to hydroxylate sima and target it for degradation, did not lead to increased FOXO nuclear localization (Figure S3D). Finally, we found that hypoxia induction of the sima target gene, fatiga, but not the FOXO target gene, 4e-bp, was suppressed in sima mutant larvae (Figure 4C). Together, these data suggest that induction of FOXO is independent of the classic HIF-1α response.

Figure 4.

Hypoxia induces FOXO independently of sima/HIF-1α. (A) FOXO staining in fat bodies of control (w¹¹¹⁸) and sima mutant (sima^KG07607) larvae exposed to either normoxia or 5% O₂ hypoxia for 2 hr. Bar, 25 μm. (B) Quantification of FOXO nuclear localization in fat-body cells of control (w¹¹¹⁸) and sima mutant (sima^KG07607) larvae exposed to either normoxia or 5% O₂ hypoxia for 2 hr at 96-hr AEL. n = total number of cells analyzed. (C and D) fatiga mRNA and 4E-BP mRNA levels measured by qRT-PCR in control (w¹¹¹⁸) and sima mutant (sima^KG07607) third instar larvae maintained in normoxia or exposed to hypoxia (5% O₂ hypoxia) for 6 hr. Data represent mean + SEM. *P < 0.05, two-way ANOVA followed by post hoc Student’s t-test. N > 4 per condition. AEL, after egg laying; FOXO, Forkhead Box-O; H, hypoxia; HIF, hypoxia-inducible factor; mRNA, messenger RNA; N, normoxia; qRT-PCR, quantitative RT-PCR.

One main way that FOXO can be regulated is via the conserved insulin/phosphoinositide 3-kinase (PI3K)/Akt pathway (Webb and Brunet 2014). This is best seen in response to nutrient availability in Drosophila. In rich nutrients, insulin signaling via PI3K to Akt kinase is high and Akt can phosphorylate FOXO, leading to its cytoplasmic retention. However, during starvation, insulin/PI3K/Akt signaling is low, thus reducing phosphorylation of FOXO and allowing it to relocalize to the nucleus to induce transcription. We investigated whether decreased Akt activation was involved in FOXO induction during hypoxia exposure. Akt is activated by phosphorylation at two sites: threonine 342 and serine 505. We measured the relative amounts of Akt phosphorylated at each site after exposure to hypoxia using phospho-specific antibodies. We saw that when third-instar larvae were exposed to hypoxia, there was a reduction in phosphorylation of Akt at both sites (Figure 5, A and B). To determine if suppression of Akt signaling was mediating the induction of FOXO, we used the flp-out technique to induce mosaic expression of the catalytic subunit of PI3K, dp110, to maintain Akt activity in fat-body cells. We found that during hypoxia, expression of dp110 was sufficient to prevent FOXO nuclear relocalization (Figure 5C). To further explore whether hypoxia induces FOXO by inhibiting Akt signaling, we compared the effects of hypoxia to other manipulations that suppress Akt. We first examined nutrient deprivation, which suppresses systemic insulin signaling and inhibits the Akt pathway. We found that nutrient starvation led to a similar increase in FOXO nuclear localization compared to hypoxia (Figure 5D). Moreover, we found that this localization was not stronger when we exposed larvae to simultaneous starvation and hypoxia exposure (Figure 5D and Figure S4A), suggesting that both starvation and hypoxia share a common mechanism to induce FOXO. We also examined the effects of genetic suppression of Akt signaling. To do this, we overexpressed PTEN, a phosphatase that reverses the effects of PI3K to suppress Akt. We found that PTEN expression in the fat body led to an increase in FOXO nuclear localization that was similar to that following hypoxia (Figure S4, B and C). Taken together, these data suggest that hypoxia induces FOXO by suppressing Akt signaling.

Figure 5.

Hypoxia induces FOXO by inhibiting PI3K/Akt. (A and B) Western blot analysis of phosphorylated T342 and S505 Akt, and total Akt in control (w¹¹¹⁸) larvae following 2 hr of normoxia or 5% O₂ hypoxia. Quantification of blots (relative phospho-Akt intensity/total Akt intensity) is shown in (B). n = 4 per condition. *P < 0.05, Student’s t-test. (C) FOXO staining in UAS-dp110-overexpressing fat-body clones (GFP-positive). Nuclei are stained with Hoechst dye (blue). Bar, 50 μm. (D) FOXO staining in 96-hr AEL larvae that were maintained in normoxia, starved (PBS only), exposed to hypoxia, or simultaneously starved and exposed to hypoxia. Top images show FOXO staining, while bottom images show corresponding nuclear staining (Hoechst dye). Bar, 50 μm. (E) Quantification of FOXO nuclear localization in fat body cells from 96-hr larvae that were maintained in normoxia, starved (PBS only), exposed to hypoxia, or simultaneously starved and exposed to hypoxia. n = total number of cells analyzed. AEL, after egg laying; cont, control; FOXO, Forkhead Box-O; H, hypoxia; N, normoxia; PI3K, phosphoinositide 3-kinase.

FOXO induces Relish-dependent hypoxia survival

In Drosophila, FOXO maintains tissue and organismal homeostasis in response to various stresses, including starvation, oxidative stress, irradiation, and infection. In each case, FOXO functions by regulating diverse and often distinct target genes. We surveyed potential FOXO targets that might be important for hypoxia tolerance and we identified a role for the NF-κB transcription factor relish.

In Drosophila there are three NF-κB transcription factors, Relish, Dorsal, and Dif. They have been best characterized as effectors of immune signaling downstream of the Immune Deficiency, IMD (Relish) and Toll (Dorsal and Dif) pathways, where they induce expression of antimicrobial peptides and promote innate immune responses (Buchon et al. 2014). We found that when exposed to hypoxia, adult Drosophila showed an increase in relish [(as reported previously by Liu et al. (2006) and Bandarra et al. (2014)], but not dorsal or dif, mRNA levels (Figure 6, A–C). Furthermore, we found that this hypoxia-induced increase in relish mRNA was blocked in both foxo mutant adults (Figure 6D) and larvae (Figure S5). Finally, we found that hypoxia could induce strong expression of Relish-regulated antimicrobial peptides in both adults (Figure 6, E and F) and larvae (Figure S6), and that this was also blocked in foxo mutants. These data suggest that in hypoxia, FOXO can induce an immune-like response via upregulation of Relish.

Figure 6.

FOXO induces Relish-dependent transcription in hypoxia. (A–C) Expression levels of relish (A), dif (B), and dorsal (C) mRNA in w¹¹¹⁸ adult females exposed to either normoxia or 16 hr of 1% O₂. Data represent mean + SEM, n = 10, *P < 0.05, Student’s t-test. (D–F) Expression levels of relish (D), attacin A (E), and cecropin A (F) mRNA in w¹¹¹⁸ and foxo^Δ94 adult females exposed to either normoxia or 16 hr of 1% O₂. Data represent mean + SEM, n = 10, *P < 0.05, two-way ANOVA followed by Student’s t-test. FOXO, Forkhead Box-O; mRNA, messenger RNA.

To test whether this immune-like response was important for hypoxia survival, we examined hypoxia survival in two independent relish null mutants, rel^E38 and rel^E20 (Hedengren et al. 1999). We found that both rel^E38 and rel^E20 adult flies showed a significant decrease in viability after hypoxia exposure (Figure 7, A and B), while survival in normoxia was unaffected (Figure S6). We also found that the reduction in hypoxia survival seen in relish, foxo double mutants was similar to that seen in either mutant alone (Figure 7C) suggesting that they function in the same genetic pathway to control hypoxia tolerance. Together, these data point to FOXO activation as a meditator of hypoxia tolerance via induction of an immune-like response through the NF-κB -like transcription factor Relish. We then tested whether induction of Relish-mediated transcription was sufficient to mediate the effects of FOXO on hypoxia tolerance. To do this we examined the effects of expression of a constitutively active version of Imd (ImdCA) (Petkau et al. 2017), the upstream activator of Relish. We used a ubiquitous gene-switch driver to express ImdCA in all tissues of adult flies. We found that this was sufficient to induce strong expression of relish target AMP genes (Figure S7). However, expression of ImdCA was not sufficient to reverse the decrease in hypoxia survival seen in foxo mutants (Figure 7D).

Figure 7.

Relish is required for hypoxia survival. (A and B) Survival of adult female w¹¹¹⁸, (A) relish^E38, or (B) relish^E20 flies after exposure to 24 hr of 1% O₂. Data represent mean + SEM, n = *P < 0.05, Student’s t-test. (C) Survival of adult female w¹¹¹⁸, foxo^Δ94, relish^E38, or foxo^Δ94, relish^E38 mutant flies after exposure to 24 hr of 1% O₂. Data represent mean + SEM, n = *P < 0.05, Student’ t-test, compared to w¹¹¹⁸ control group. (D) Hypoxia survival (24 hr at 1% O₂) of adult female w¹¹¹⁸ or foxo^Δ94 relish^E20 flies with (+) or without (−) expression of UAS-ImdCA with a ubiquitous da-GeneSwitch driver. Data represent mean + SEM, n = *P < 0.05, two-way ANOVA followed by Student’s t-test. ns, not significant.

Discussion

In this paper, we report that FOXO is a HIF required for organismal survival in low oxygen, and we show that one way that FOXO functions is through upregulation of Relish/NF-κB (Figure 8). We saw that the hypoxia induction of FOXO occurs via suppression of PI3K/Akt signaling. This response is most likely induced by hypoxia-mediated reduction of insulin release and signaling, the main activator of PI3K/Akt, as previously reported in Drosophila larvae (Wong et al. 2014; Texada et al. 2019). Together with previous studies in C. elegans and Zebrafish showing that reduced insulin signaling and FOXO induction confer hypoxia tolerance (Scott et al. 2002; Mendenhall et al. 2006; Menuz et al. 2009; Liu et al. 2016), our work suggests that FOXO is a conserved mediator of hypoxia responses.

Figure 8.

A model for FOXO- and Relish-dependent hypoxia survival. Upon hypoxia exposure, the PI3K/Akt pathway is inhibited and FOXO is able to relocalize to the nucleus. FOXO can then upregulate target genes, including the NF-κB factor Relish, to promote hypoxia survival. FOXO, Forkhead Box-O; PI3K, phosphoinositide 3-kinase; dILPs, Drosophila insulin-like peptides; InR, insulin receptor.

Interestingly we found that the induction of FOXO upon hypoxia occurs in sima mutants, suggesting that the FOXO hypoxic response occurs independently of the classically described HIF-1α response. Work in mammalian cell cultures has reported that upon hypoxia, HIF-1α can induce FOXO3a function (Jensen et al. 2011). However, in vivo genetic studies in model organisms suggest that the HIF-1α and FOXO transcription factors may act in parallel to mediate responses to hypoxia. For example, in Drosophila, a hypoxia-induced HIF-1α pathway that leads to target of rapamycin inhibition functions independently of FOXO (Reiling and Hafen 2004). In C. elegans, the extension of life span caused by hypoxia and increased HIF-1α protein levels occurs in the absence of FOXO nuclear localization and function (Mehta et al. 2009; Müller et al. 2009; Zhang et al. 2009; Leiser et al. 2011). Finally, both HIF-1α and FOXO have been shown to act in parallel in C. elegans to control iron homeostasis (Ackerman and Gems 2012). These genetic studies, and our work presented here, suggest that inductions of both FOXO and HIF-1α are two parallel responses to hypoxia in animals.

One key way that cells, tissues, and organisms adapt to low oxygen is by altering their glucose metabolism to maintain homeostasis (Nakazawa et al. 2016; Xie and Simon 2017). Our data suggest that one reason that foxo mutants may show reduced hypoxia tolerance is that they have deregulated control over glucose metabolism. Thus, we saw that foxo mutant animals had low levels of glucose in normoxia and that both stored and circulating forms of glucose were significantly decreased under hypoxia compared to controls. These results suggest that FOXO is needed for either gluconeogenesis during stress, as has been reported in C. elegans (Hibshman et al. 2017), or for proper control of glycolysis. Indeed, we saw that expression of ldh is markedly increased in foxo mutants. Ldh is a rate-limiting enzyme involved in the conversion of pyruvate to lactate, which is a key metabolic event that can drive increased glycolysis, and ldh levels have been shown to increase in larvae upon hypoxia exposure (Li et al. 2013). Thus, one possibility is that foxo mutant animals may engage in abnormally high levels of glycolysis in low oxygen, leading to depletion of glucose and reduced hypoxia tolerance. This is consistent with previous studies in Drosophila showing a major role for FOXO as a regulator of metabolic homeostasis in the context of other stress responses such as starvation and pathogenic infection (Dionne et al. 2006; Teleman et al. 2008). For example, FOXO often functions in a tissue-specific manner to control systemic sugar and lipid metabolism (Wang et al. 2011; Karpac et al. 2013; Borch Jensen et al. 2017; Zhao and Karpac 2017; Molaei et al. 2019). These effects have been shown to be important for FOXO to extend life span and to promote increased tolerance to stress.

A central finding of our work is that one way that FOXO provides protection in low oxygen is through induction of the immune transcription factor Relish. In Drosophila, there are two main immune effector pathways that respond to pathogen infection and that work through induction of NF-κB transcription factors: the IMD pathway, which targets the NF-κB homolog Relish, and the Toll pathway, which works via the Dorsal and Dif NF-κB transcription factors (Buchon et al. 2014). We found that hypoxia specifically induced Relish via FOXO, and that this response was required for hypoxia tolerance. These data, together with previous work showing hypoxia induction of Relish (Liu et al. 2006; Bandarra et al. 2014), suggest that induction of an immune-like transcriptional response may be a protective mechanism in low oxygen in Drosophila. In the context of animal immunity, there is increasing appreciation of the role for infection tolerance as a defense strategy against pathogens (Ayres and Schneider 2012; Medzhitov et al. 2012; Lissner and Schneider 2018). This tolerance is often mediated via alterations in systemic metabolism and physiology to limit infection-induced tissue damage (Wang et al. 2016; Weis et al. 2017; Ganeshan et al. 2019). Our findings suggest that tolerance to hypoxia may share some of these metabolic and physiological functions. In Drosophila, this interplay between hypoxia and innate immune responses may reflect the natural ecology of flies. In the wild, Drosophila grow on rotting, fermenting food, an environment rich in microorganisms, including pathogenic bacteria. In these anaerobic conditions, low ambient oxygen may “prime” animals to deal with subsequent pathogenic bacterial encounters. Hence, one speculative idea is that experimental exposure of Drosophila to hypoxia may induce Relish and provide protection against the detrimental effects of subsequent pathogenic infection. This concept of hypoxia preconditioning has been observed in C. elegans, where it is important in protecting against cell death and damage induced by pore-forming toxins (Dasgupta et al. 2007; Bellier et al. 2009).

It is possible that the effects of FOXO on metabolism in hypoxia could be mediated via Relish. For example, a recent report showed that Relish was required to control metabolic responses to nutrient deprivation in Drosophila (Molaei et al. 2019). Furthermore, constitutive activation of IMD signaling, which signals via Relish, was shown to lead to decreased circulating sugars in adult Drosophila (Davoodi et al. 2019). In mammals, NF-κB is activated in response to cytokines, and it functions as a central regulator of immune and inflammatory responses (Zhang et al. 2017). Several studies have shown that an important way that NF-κB works to mediate these effects is through the control of glycolysis and mitochondrial metabolic activity (Mauro et al. 2011; Tornatore et al. 2012). Indeed, links between immunity and metabolism are emerging as important components of infection tolerance in animals (Ayres and Schneider 2012). Our data suggest the possibility that organisms may also coopt some of these immune–metabolism interactions to tolerate low oxygen.

Relish has also been shown to influence systemic metabolism in response to different stresses by controlling endocrine signaling. For example, in response to radiation damage, Relish activity in the larval fat body can control systemic insulin signaling (Karpac et al. 2011). In addition, Relish can function in the adult muscle in response to mitochondrial stress to control expression of the TGFβ ligand activin, which in turn regulates fat-body lipid metabolism (Song et al. 2017). Hence, it is possible that these types of endocrine signaling effects may explain how Relish functions to control metabolism and survival in hypoxia.

Functional interactions between FOXO and Relish have been described in response to other stressors in Drosophila. For example, nutrient starvation induces Relish in larvae via FOXO and this is important for controlling systemic insulin signaling (Karpac et al. 2011). In addition, as adults age, FOXO is induced in the intestine and it, in turn, upregulates Relish to control intestinal homeostasis and life span (Karpac et al. 2013; Guo et al. 2014). Interestingly, Relish and FOXO have an antagonistic relationship in adult fat, and these interactions are important for metabolic adaptation and survival upon starvation (Molaei et al. 2019). Hence, the links between FOXO and Relish are likely to be tissue-specific, but they may have evolved to function as a general mediator of stress responses. Functional links between NF-κB and FOXO have also been reported in mammalian cells (Lin et al. 2004; Thompson et al. 2015), and together with the reported induction of NF-κB in hypoxia in mammalian cell culture (Rius et al. 2008; Fitzpatrick et al. 2011), they suggest that the hypoxia-FOXO-NF-κB regulation that we see in Drosophila may operate in mammalian cells too.