The tracheole length grows as the optic lobe volume increases during larval development. Tracheolar trees of the optic lobe were reconstructed in brains stained with Calcofluor and anti-Discs large (not shown) at 12 h intervals from 24 to 96 h ALH. Dashed lines indicate the brain hemisphere and its corresponding optic lobe (identified by morphological landmarks visualised by anti-Discs large staining). Both the optic lobe (A–G) and its tracheoles (A–F,H) grow continuously along larval life. Analysis of the ratio of total tracheolar length in the optic lobe and optic lobe volume showed a decline in the proportion of optic lobe tracheolation (I). Scale bars: 20 µm for panels A–C, 50 µm for panels D–F. *P<0.05, **P<0.001, Mann–Whitney Wilcoxon test. ns., non-significant; ALH, after larval hatching; CB, central brain. Sample sizes for time points from 24 h to 96 h ALH, n=4, 6, 6, 6, 6, 5, 6, respectively.

The hypoxia response in central brain and optic lobe is differentially regulated throughout larval development. (A,C,E) Ratiometric images of single frontal confocal sections of a brain hemisphere of larvae expressing the green (ubi-GFP-ODD) and red (ubi-mRFP-nls) fluorescent proteins of the biosensor. The colour code (upper right) indicates average GFP-ODD/mRFP-nls ratios for each nucleus, which were segmented based on the mRFP-nls signal. (A′,C′,E′) Maximum intensity projection of the brain tracheal system (white). The dotted line denotes the border between central brain and optic lobe. (B,D,F) Histograms representing the frequency distribution of GFP-ODD/mRFP-nls ratios (normalised to whole-brain average) for the central brain and the optic lobe, showing a clear difference in the distribution of values between the two brain compartments. (B′,D′,F′) Box plots showing the non-normalised mean of the GFP-ODD/mRFP-nls ratios, showing maximum and minimum observation, upper and lower quartile, and median. Scale bars: 15 µm (A,A′), 20 µm (C,C′), 40 µm (E,E′). *P<0.05, **P<0.01, ***P<0.001 Student’s t-test or Mann–Whitney Wilcoxon test. Sample size n=6, 6, 8.

The hypoxia response correlates with the distance between a cell and its closest tracheole. (A) Ratiometric image of a larval hemisphere at 84 h ALH superimposed with a maximum projection of the tracheal system (Calcofluor). (A′) Magnified view of the brain area framed by the yellow square in (A), which points to cells (nuclei) within the optic lobe volume surrounding the optic lobe lateral tracheole (OLTl, white arrow in A). Inset showing a single nucleus (inside dotted circle) for which the corresponding ratio value and distance to tracheole are shown (bottom right). (B) ratio⁻¹ plotted against distance to OLTl tracheole. The black line shows the exponential fit to the data (n=6). (C) ratio⁻¹ plotted against minimum distance to tracheole for every nucleus in the brain (purple: central brain, orange: optic lobe; n=6). The black dotted line is an exponential fit to the data. (D) Exponential fits (grey lines) for values of eight different brains; the black dotted line shows the average fit for these brains. (E) The function resulting of the exponential fit to the data was used to calculate a predicted ratio value for each cell according to the distance from the cell to its closest tracheole. The values are depicted with a colour codescale from predicted high ratios (bright colours) to predicted low ratios (dark colours). Scale bars: 40 µm.

GFP-ODD degradation is driven by oxygen availability. Ratiometric images for larvae exposed to hyperoxia (A, n=7) and hypoxia (D, n=7) and the corresponding maximum projection of their tracheal system (Calcofluor; A′,D′). Larvae reared in ambient hyperoxia show a left-shift (higher oxygen) in the distribution of optic lobe ratio values (B). Brains of larvae exposed to hyperoxia show lower ratio values both for central brain and optic lobe (B′). (E) Larvae reared in hypoxia showed two distinct populations of values for central brain and optic lobe as observed in normoxia (see Fig. 3E). (E′) Hypoxia-reared larvae showed increased non-normalized mean ratio values (i.e. lower oxygen) as compared to normoxia. (C,F) ratio⁻¹ (oxygenation) plotted against minimum distance to tracheoles. Sky blue lines (C) and orange lines (F) show exponential fits for different brains from larvae reared in hyperoxia and hypoxia, respectively. Dotted lines show average of all fits for normoxia (black), hyperoxia (dark blue) and hypoxia (red). Scale bar is applicable to all panels and is 40 µm. *P<0.05, **P<0.01, ***P<0.001 Student’s t-test or Mann–Whitney Wilcoxon test.

The biosensor reveals cell-type specific hypoxia states in central brain and optic lobe. (A–E) Dorsal views (in relation to the neuraxis) of immunostained brains for cell-type specific markers (gray): (A) anti-Deadpan (neuroblasts, NB), (B) anti-Prospero (ganglion mother cells, GMC), (C) anti-Elav (neurons), (D) anti-Repo (glial cells). The cell-specific nuclear staining shown in (A–D) was utilized as segmentation signal to create a mask to obtain the corresponding ratiometric analysis of each cell type (A′–D′). The anti-Discs large staining (E) was used to outline neuroepithelial cells. The corresponding ratiometric analysis (E′) was based on manual segmentation of the IPC and OPC using TrakEM2. (F) Mean ratio⁻¹ (oxygenation) for each cell type in central brain and optic lobe, represented as a function of mean distance to tracheoles. The dotted line shows the average trend, obtained by averaging the fits of all normoxic brains. (G) Decaying exponential fits for all cell types. (H) Boxplot comparing ratiometric values of all cell types both in central brain and optic lobe (n=4). (I) Exposure to hyperoxia has a stronger effect in neuroblasts than in neuroepithelial cells (n=6). (J) Cell-type specific hypoxia response in the central brain and optic lobe based on biosensor data. Scale bar is applicable to all panels and is 40 µm. Error bars in (F) show s.e.m. *P<0.05, **P<0.01, ***P<0.001 Student’s t-test or Mann–Whitney Wilcoxon test.