In almost all animals, low physiological levels of oxygen (hypoxia) reduce growth rate and adult body size. Despite the near ubiquity of this response, the systemic mechanisms that coordinate growth and development under hypoxia remain poorly understood. In Drosophila, hypoxia increases circulating levels of the steroid hormone ecdysone to inhibit insulin/IGF signaling and slow growth. At the same time, ecdysone biosynthesis is reduced to delay pupation and extend development. Traditionally, the secretion of lipid-soluble steroids is thought to be regulated at the level of biosynthesis. This therefore presents a paradox: how can a single environmental factor both increase ecdysone levels yet decrease ecdysone synthesis? Our data show that this paradox is resolved by the dual regulation of ecdysone at the levels of secretion and biosynthesis. We show that, in the short term, hypoxia increases basal levels of ecdysone through the Atet-dependent exocytosis of ecdysone-containing vesicles. In the long term, hypoxia decreases the expression of ecdysone synthesis genes to delay the peak in ecdysone that triggers pupation. We also present evidence that both ecdysone synthesis and secretion are regulated, in part, by NO-signaling. Collectively, our findings reveal regulated steroid secretion as a critical and environmentally-responsive component of endocrine control, expanding our understanding of how animals integrate growth and developmental timing in response to acute and chronic environmental change.

Drosophila stocks and maintenance

We used the following fly strains: the isogenic RNAi control (VDRC 60,000), UAS Atet RNAi^v42751 (VDRC 42751), phm-GAL4 (a gift from Michael O’Connor), P0206-Gal4 (a gift from Christen Mirth), UAS-Syt-GFP (BDSC 6925), UAS-Atet RNAi⁵⁰⁵⁶³ (BDSC 50563), and UAS-Bnl RNAi (BDSC 34572). All constructs were backcrossed into VDRC 60,000, for five generations prior to analysis. Flies were reared at low density (50-100 larvae per vial), 25°C, 21% O₂ in constant light on a sucrose-yeast diet containing 13 g of carrageenan, 100 g of yeast extract, and 50 g of sucrose in 1000 ml of water. To prevent bacterial and fungal growth, we added 3% Nipagin and 0.3% (v/v) propionic acid to the cooled mixtures.

Body size measurement

All larvae were reared under standard conditions (25^oC, 21% O₂) and staged into 2h cohorts at ecdysis to the third instar before being maintained at 25°C, 21% O₂ (normoxic flies) or moved to 25°C, 10% O₂ (hypoxic flies) for the remainder of their development. We used pupal case size as a measure of body size. We collected digital images of the pupal cases and measured the area of the pupal case when viewed from the dorsal aspect using ImageJ. Measurements were log transformed prior to analysis.

Growth rate measurements

P0206>+ and P0206>HIF-1α.RNAi larvae were reared under standard conditions (25^oC, 21% O₂) and massed at ecdysis to L3 (72h after oviposition) and 24h later (96h after oviposition). Specific growth rate (mg/d) was calculated as log M₉₆ – log M̅₇₂, where M_x is mass at age x.

Quantitative PCR (qPCR)

VDRC 60,000 larvae were reared under standard conditions (25^oC, 21% O₂) and staged into 2h cohorts at ecdysis to the third instar before being maintained at 25°C, 21% O₂ (normoxic flies) or moved to 25°C, 10% O₂ (hypoxic flies). Larvae were subsequently collected every four or six hours for 24 hours and preserved in groups of ten in RNAlater (Thermo Fischer Scientific). We extracted the RNA using Trizol (Invitrogen), treated it with DNase I (Thermo Fischer Scientific), quantified it with a NanoDrop One (Thermo Fischer Scientific) and reverse transcribed it with a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Quantitative RT–PCR was conducted using PowerUp SYBR Green Master Mix (Applied Biosystems) and measured on a QuantStudio 3 Real-Time PCR system (Applied Biosystems). We calculated mRNA abundance for three biological replicates of ten larvae, using a standard curve and normalizing against the expression of ribosomal protein 49 (RP49). Standard curves were generated using seven serial dilutions of total RNA extracted from the isogenic control line (60,000): 5x 1st instar larvae, 5x 2nd instar larvae, 3rd instar larvae (male), 5x pupae (male), 5x adult flies (male). Primer sequences used in the study are listed in Supplementary Table S1.

Statistical analyses

All statistical analyses were conducted in R (66). All the data and the R-scripts used to analyze them are provided as supplementary information. The effects of genotype on the response of body size and specific growth rate to hypoxia were tested using two-way ANOVAs (body size/specific growth rate = genotypeoxygen),* with post hoc Šidák-corrected comparisons. The effect of genotype on developmental time was tested using a binomial GLMM (pupation = agegenotype + vial),* where vial was treated as a random factor, with post hoc Tukey-corrected pairwise comparisons between genotypes. The effect of oxygen level on gene expression was tested using two-way ANOVAs (expression = ageoxygen*), where age is a categorical variable, with post hoc Šidák-corrected comparisons within specific ages.