Ablation of Insulin-Producing Neurons in Flies: Growth and Diabetic Phenotypes

The Drosophilagenome contains five Drosophila insulinlike peptide genes (dilp1 through -5) with significant homology to mouse and human insulins and two others with far less similarity (dilp6 and -7). These genes are expressed in tissues ranging from early embryonic mesoderm to small clusters of larval brain neurons, ventral nerve cord neurons, salivary gland, and midgut (1). Using messenger RNA (mRNA) in situ hybridization, we established that the most prominent insulin gene expression during larval stages, a period of intensive feeding and rapid growth, is within two bilaterally symmetric clusters of neurosecretory cells in the pars intercerebralis region of the protocerebrum (fig. S1).

An 859–base pair promoter fragment, comprised of sequences immediately 5′ of dilp2, was sufficient to drive gene expression in the small clusters of larval brain neurons that express dilp1, -2, -3, and -5 (fig. S1). To assess the role of the brain IPCs as an insulin-producing endocrine system, we ablated the brain IPCs using the dilp2 promoter to express the cell death–promoting factor, Reaper (2). The IPC ablation results in deficiency of brain neuron-derived insulin only (fig. S2). IPC ablation caused undergrowth phenotypes, developmental delays, and lethality similar to Drosophila insulin receptor (DInR) mutants (1, 3). To rule out an underlying cause of these phenotypes other than insulin deficiency, such as non-IPC death from Reaper or loss of other essential brain IPC functions, we used a heat shock–inducible dilp2 transgene with ubiquitous expression to reverse the effect of the IPC ablation.

We quantified the defect in growth by comparing larval length after 120 hours of development, a time at which synchronized cultures of normal larvae will reach wandering third instar and puparium formation. [Larval phenotypes are shown in (fig. S3).] After IPC ablation, larvae attained a mean length only 58% of normal size (Fig. 1A, center). Larvae with IPCs ablated but expressing the inducible dilp2 transgene had their mean length rescued to 88% of normal (Fig. 1A, right). The developmental time to reach wandering third instar and puparium formation was approximately 5 days in normal larvae but lengthened to 12 days in larvae after IPC ablation, a developmental rate similar to that observed in animals homozygous for loss-of-function mutations in DInR (1). Larvae with ablated IPCs that expressed the inducibledilp2 transgene required approximately 6 days to reach puparium formation. Thus, systemic DILP2 expression was sufficient to compensate for IPC ablation. That brain IPC ablation is rescued bydilp2 alone suggests insulin made by brain IPCs may be partially redundant.

• Download high-res image
• Open in new tab
• Download Powerpoint

IPC ablation produced small-sized adults of normal proportion (fig. S3). Examination of adult wings revealed reductions in both cell size and number after IPC ablation. [Wing phenotypes are shown in (fig. S3).] Under the strongest condition of IPC ablation, mean wing size was reduced to 61% of normal, whereas wing cell number and size were reduced to 72% and 85% of normal, respectively (Fig. 1B). Under a less severe regimen of IPC ablation, mean wing size was reduced to 74% of normal, with reductions in cell number and size to 81% and 91% of normal, respectively (Fig. 1B). As in larval growth, the dilp2 transgene effectively reversed the effects of IPC ablation on wing growth and, in fact, caused a slight overgrowth effect (Fig. 1B), possibly due to the 20% lengthening of developmental time that allowed more growth. The overall reduction in cell size and number after IPC ablation is similar to that in DInR (1) and IRS1-4 mutants (4). This, together with our observation that brain IPC-derived insulins can activate the DInR in vitro (fig. S4), suggests that brain IPCs are a key source of insulin for this growth control pathway.

We also investigated the role of brain IPCs and insulin in the regulation of carbohydrate metabolism. Trehalose is a disaccharide composed of two glucose molecules and is the principal blood sugar in many insects (5). Using the same heat shock regimen, we compared the combined concentration of glucose and trehalose in the hemolymph of wandering third instar larvae, a brief and discrete developmental stage before puparium formation when feeding has ceased. The IPC-ablated larvae had an average combined glucose and trehalose level of 38% above normal, and these levels returned to normal when DILP2 was provided systemically by the transgene (Fig. 1C). Elevated hemolymph carbohydrate levels in larvae lacking IPCs indicate that insulin is an essential regulator of energy metabolism inDrosophila. This accumulation of carbohydrate in the blood is reminiscent of that seen in human diabetes mellitus, although it should be noted that we measured carbohydrate levels during development rather than in adults.

To investigate how central nervous system (CNS)–derived insulin regulates systemic functions, we examined theDrosophila IPC contacts outside the CNS. The morphology of the brain IPCs was examined with the use of the dilp2promoter to drive expression of mGFP, a membrane bound green fluorescent protein (GFP) (Fig. 2, A to C). The IPC clusters within the pars intercerebralis extend processes that terminate at the lateral protocerebrum and subesophageal ganglion. IPC processes also terminate on the heart and in the corpora cardiaca (CC) compartment of the ring gland, after crossing the midline and exiting the CNS (Fig. 2F). Labeling of the IPC processes with mGFP and DILP2 antibody (fig. S1) revealed localization of DILP2 peptide within the processes that contact the heart and ring gland (Fig. 2A). DILP2 peptide was concentrated in a graded distribution outside the cells of synthesis on the heart (Fig. 2C), and colabeling with myosin heavy chain antibody, which labels the columnar heart epithelium, showed that the IPC processes and DILP2 were localized outside the lumen of the heart (Fig. 2D). These results suggest the heart surface may be the site of insulin release to the openly circulating hemolymph. We propose that brain IPCs are essential for organismwide growth control and carbohydrate homeostasis through release of insulin peptides into circulating hemolymph. These cellular functions are notably similar to those of mammalian pancreatic β cells (6, 7).

• Download high-res image
• Open in new tab
• Download Powerpoint

In Drosophila (8) and other insects (9), a fraction of CC cells synthesize adipokinetic hormone (AKH). AKH resembles glucagon in its activation of glycogen phosphorylase through heteromeric GTP-binding protein (G protein) and adenosine 3′,5′-monophosphate (cAMP) signaling to elevate blood sugar (10), and the two proteins have some limited sequence similarity (fig. S5). Double labeling of AKH mRNA and DILP2 peptide showed that IPCs extend processes to the CC and that AKH-expressing cells contain DILP2 (Fig. 2E). We also found that CC cells accumulate DILP2 within membrane-bound particles of the perinuclear space, suggesting the possibility that DILP2 is taken up by AKH cells (fig. S6). We cannot formally rule out the possibility thatdilp2 is transiently or minutely transcribed by AKH cells, although we have not detected expression of either thedilp2 promoter or dilp1, -2, -3, and -5 mRNAs in the AKH cells. Thus, in addition to contacts between IPCs themselves, the primary sites of IPC contact outside the CNS are the heart and the CC. Though they lack strict morphological homology, these intercellular contacts are analogous to the association of pancreatic β cells with other β cells, with glucagon-expressing α cells, and with blood vessels in the islets of Langerhans (7) and may reflect underlying evolutionary conservation.

Thus, there is remarkable similarity of the organ systems underlying conserved insulin function in diptera and mammals. Moreover, the presence of IPCs in the nervous systems of other invertebrate and protochordate species (1,11–14) and in primary cell cultures from mammalian fetal brain (15) provides further evidence for a common ancestral insulin-producing organ of neural origin. Our results also raise the possibility that common mechanisms of cell specification regulate development of pancreatic β cells and Drosophilabrain IPCs.