Insulin-producing cells in the brain of adult Drosophila are regulated by the serotonin 5-HT1A receptor

Abstract

Insulin signaling regulates lifespan, reproduction, metabolic homeostasis, and resistance to stress in the adult organism. In Drosophila, there are seven insulin-like peptides (DILP1-7). Three of these (DILP2, 3 and 5) are produced in median neurosecretory cells of the brain, designated IPCs. Previous work has suggested that production or release of DILPs in IPCs can be regulated by a factor secreted from the fat body as well as by neuronal GABA or short neuropeptide F. There is also evidence that serotonergic neurons may regulate IPCs. Here, we investigated mechanisms by which serotonin may regulate the IPCs. We show that the IPCs in adult flies express the 5-HT(1A), but not the 5-HT(1B) or 5-HT(7) receptors, and that processes of serotonergic neurons impinge on the IPC branches. Knockdown of 5-HT(1A) in IPCs by targeted RNA interference (RNAi) leads to increased sensitivity to heat, prolonged recovery after cold knockdown and decreased resistance to starvation. Lipid metabolism is also affected, but no effect on growth was seen. Furthermore, we show that DILP2-immunolevels in IPCs increase after 5-HT(1A) knockdown; this is accentuated by starvation. Heterozygous 5-HT(1A) mutant flies display the same phenotype in all assays, as seen after targeted 5-HT(1A) RNAi, and flies fed the 5-HT(1A) antagonist WAY100635 display reduced lifespan at starvation. Our findings suggest that serotonin acts on brain IPCs via the 5-HT(1A) receptor, thereby affecting their activity and probably insulin signaling. Thus, we have identified a second inhibitory pathway regulating IPC activity in the Drosophila brain.

Fig. 1

The 5-HT_1ADro receptor locus and 5 kb region used for Gal4 construct. The genomic region of the 5-HT_1ADro locus (CG16720) is on the right arm of the second chromosome (light bars 5′ and 3′ untranslated regions, dark bars coding regions; arrows indicate mRNA transcription start sites). The 5-kb region of genomic DNA used to make the GAL4 construct is indicated immediately upstream of the mRNA start site. The adjacent CG15109 locus is 7 kb upstream of the first exon of the 5-HT_1ADro gene. The CG15109 transcript has been reported to be exclusively expressed in the male testis

Fig. 2

Expression of 5-HT1A receptor on IPCs in the Drosophila brain. All images of adult brains are frontal views (dorsal is up) and of larval brain horizontal views (anterior is up). a Anatomy of the IPCs seen with Dilp2-Gal4-driven GFP (projection of optic sections). A cluster of 16 cell bodies (cb) give rise to axons with branches in two regions of the pars intercerebralis: wide dorsal lateral branches (asterisks) and shorter median branches (arrow). b–d In the adult brain, the 5-HT_1A-Gal4 driver is expressed in a set of median neurosecretory cells (green), most of which also express DILP2- immunolabeling (magenta). In the merged image (d), it can be seen that all but one of the 5-HT_1A-expressing cells display DILP2 immunolabeling (whitish). e, f In the feeding third instar larval brain, there is no 5-HT_1A expression in the DILP2-immunolabeled IPCs (projection of several optic sections). The weakly labeled cell body in (e) (asterisk) does not express DILP2. g In the wandering (nonfeeding) third instar larva, 5-HT_1A expression starts. Here, one IPC (arrow) coexpresses the receptor and DILP2. h In another specimen of the same age, two IPCs on the left side coexpress receptor and DILP2, but no cells on the right (asterisk). This image is a projection of several optic sections. i–k A single optic section of the same specimen showing colocalization of markers in one cell body (arrow) to the left and none on the right (asterisk)

Fig. 3

Relationships between processes of IPCs and serotonin-immunoreactive neurons in the adult brain. We utilized the Dilp2-Gal4 to drive GFP in IPCs (green) and a mouse monoclonal antibody to serotonin (magenta) to visualize relationships between the two neuron types in the pars intercerebralis (shown in frontal view; dorsal up). a–c IPCs at the level with wide lateral branches dorsally. These IPC branches superimpose those of varicose serotonin immunoreactive ones (even the small set of branches more ventrally, at asterisk). d The framed area in (a) is shown at higher magnification. e–g The IPCs at the level of the median short branches (other specimen). Again, the branches of the two neuron types superimpose (e.g., at arrows). h The framed area in (e) is shown at higher magnification

Fig. 4

Knockdown of 5-HT_1A leads to an increase in DILP-immunolabeling in insulin-producing cells (IPCs). a Relative DILP immunofluorescence in IPCs in fed and starved flies with and without 5-HT_1A receptor knockdown (5-HT_1ARi) in IPCs (Dilp2-Gal4/UAS-5-HT_1A-RNAi). The DILP2 antiserum used is likely to cross react with DILP2, 3 and 5. Control flies (w¹¹¹⁸-5-HT_1A) display significantly lower levels of DILP-immunofluorescence than the flies with diminished 5-HT_1A (Dilp2-5-HT_1ARi), both in fed flies (***p < 0.001; One-way ANOVA with Tukey’s comparison) and after starvation (***p < 0.001). Also, the increases of fluorescence in controls (gray bars) and knockdown flies (red bars) when comparing fed and starved flies are significant (p < 0.001 for both). For each genotype and condition, IPCs of 5–7 brains were investigated. b Relative DILP immunofluorescence in IPCs in fed and starved mutant (5-HT_1A mut) and wild type (w¹¹¹⁸) flies. Wild-type flies display significantly lower levels of DILP-immunofluorescence than the 5-HT_1A mutant flies, both in fed flies (***p < 0.001) and starved ones (*p < 0.05). Again, the increases in immunolabeling in fed and starved controls and fed and starved mutants are significant (p < 0.001 for both). IPCs of 7–10 brains of each genotype and condition were investigated

Fig. 5

Knockdown of 5-HT_1A receptor in IPCs increases sensitivity to starvation. We performed GABA_B 5-HT_1A receptor knockdown in IPCs with two different RNAi constructs, one from VDRC [5-HT_1ARi(V)] (a) and another from Bloomington Stock Center [5-HT_1ARi (B)] (b). These flies were crossed to a Dilp2-Gal4 driver [40] used in all experiments, unless other specified. Male flies were kept on aqueous agarose (to induce starvation) and their survival monitored over time. All experiments were run in three replicates. a Using a Dilp2-Gal4 driver to knockdown the 5-HT_1A receptor in IPCs [Dilp2-5-HT_1ARi (V)], we obtained flies that display significantly reduced survival at starvation (p < 0.0001 to both parental controls, Log rank test, Mantel-Cox; n = 111–215 for each genotype). b Flies obtained from crossing Dilp2-Gal4 with the other strain 5-HT_1ARi (B) also displayed significantly reduced survival at starvation (p < 0.0001 and p = 0.0002 to the two controls; n = 159–209 for each genotype). In all subsequent graphs with 5-HT_1ARNAi, we used the (V) strain. We also used a different Dilp2-Gal4 driver (on the 3rd chromosome) [18] to drive the UAS-5-HT_1ARi(V) and obtained the same significantly reduced life span at starvation (see S. Fig. 3)

Fig. 6

Knockdown of 5-HT_1A receptor in IPCs or globally increases sensitivity to starvation. Using the same experimental conditions, we tested lifespan at starvation with another Gal4 driver that includes the IPCs and a 5-HT_1A mutant. a The enhancer trap Gal4 OK107 includes the IPCs in its expression pattern [23] and was used here to drive UAS-5-HT_1ARi. The lifespan of OK107-5-HT_1ARi flies is significantly reduced at starvation (p < 0.0001 to both controls, Log rank test, Mantel-Cox; n = 84–121 for each genotype, experiment run in two replicates). b Heterozygous 5-HT_1A mutant flies also display significantly reduced survival at starvation compared to wild-type (w¹¹¹⁸) and controls (p < 0.0001 to wild-type controls; n = 165 for each genotype, experiment run in three replicates)

Fig. 7

Targeted expression of 5-HT₇ in IPCs decreases lifespan at starvation. The 5-HT₇ receptor couples via Gs to stimulate adenylate cyclase. Therefore, we ectopically expressed this receptor by means of the Dilp2 Gal4 driver (Dilp2-UAS-5-HT₇) to test the effect on starvation resistance. Indeed, this ectopic expression produced flies with decreased the lifespan at starvation (p < 0.001 to both controls, Log rank test, Mantel-Cox; n = 93–122 for each genotype; experiment in two replicates). The decreased lifespan suggest that the ectopic 5-HT₇ couples to Gs and stimulates IPCs and insulin signaling

Fig. 8

Feeding flies a 5-HT_1A antagonist increases sensitivity to starvation. The 5-HT_1A antagonist WAY100635 was fed to the flies via aqueous agarose at a concentration of 0.18 mM (0.1 g/L) and the flies were kept on this agarose for the duration of the starvation experiment. a Survival of wild-type (w¹¹¹⁸) flies kept on agarose with or without the antagonist. Antagonist-fed flies displayed a significant reduction in lifespan (p = 0.001, Log rank test; n = 101 and 165 for the two test groups; run in two replicates). b Survival of heterozygous 5-HT_1A mutant (mut) flies fed agarose with or without antagonist. The antagonist further reduces lifespan in the mutant flies (p < 0.0001; n = 166 and 147 for the two test groups; run in two replicates). c Comparison of median survival (lifespan) of the four groups of flies tested in (a) and (b). It can be seen that the antagonist action is additive to the heterozygous mutation of the receptor. Thus, the shortest median life span is seen for mutant flies fed the antagonist, which is significantly shorter than both wild-type flies fed antagonist and mutants fed agarose alone (*p < 0.05, **p < 0.01; one-way Anova with Tukey comparison)

Fig. 9

Responses to temperature stress are influenced by knockdown of 5-HT_1A receptor in IPCs or globally. Flies with the 5-HT_1A receptor knocked down in IPCs by the transgene Dilp2-Gal4/UAS-5-HT_1A-RNAi or globally in the mutant were tested for responses to temperature stress. All experiments were run in two replicates. a, b Response to heat was tested by exposing flies to 39°C and monitoring time to knockdown (given as percent survival with heat). a The flies with the 5-HT_1A receptor diminished in IPCs (Dilp2-5- HT_1ARi) displayed a faster knockdown at 39°C (p = 0.0005 and p < 0.0001 to parental controls; Log rank test, Mantel-Cox; n = 48–55 for the three genotypes). b The 5-HT_1A mutant (mut) flies displayed a similar increased sensitivity to heat compared to wild-type flies (w¹¹¹⁸) (p < 0.0001; n = 50 and 54). c, d Recovery from cold knockdown (coma) was monitored in the same genotypes. Flies were kept at 0°C for 4 h and the time to recovery was monitored (given as percent in chill coma). c The flies with receptor knockdown in IPCs were slower in their recovery from cold (p = 0.0009 and p = 0.0017 to controls; n = 52–58 for the three genotypes). d The receptor mutant flies also display a longer recovery time (p < 0.0001 to wild-type control; n = 56 and 58)

Fig. 10

Knockdown of 5-HT_1A receptor in IPCs or globally affects lipid storage at starvation. Whole-body lipid was measured, as given in “Materials and methods”, in fed flies (0 h) and flies exposed to starvation for 12 and 24 h. a Flies with the 5-HT_1A receptor diminished in IPCs (Dilp2-5-HT_1ARi) displayed a significantly lower amount of lipid than control flies both in fed and starved flies (***p < 0.001; one-way Anova with Tukey’s comparison; n = 120 for each genotype; experiment run in three replicates). The decrease in lipid over time was also significant for both genotypes (p < 0.001; two-way ANOVA). b In the 5-HT_1A mutant flies, the lipid levels were lower than in wild-type flies at 12 and 24 h of starvation. (***p < 0.001, ns not significant; one-way Anova, n = 120 for each genotype; experiment run in three replicates). The decrease in lipid over time was also significant for both genotypes (p < 0.001; two-way ANOVA)