Insulin, insulin-like growth factor and target of rapamycin (TOR) signaling pathways are critical for proper regulation of systemic organismal growth under variable nutritional and environmental conditions (1). In mammals, insulin receptor (INSR) and insulin-like growth factor receptors (IGF-1R and IGF-2R) are present ubiquitously pre- and postnatally. The Drosophila genome encodes only one insulin receptor, called Insulin-like receptor (InR), which is initially translated into a single polypeptide and subsequently processed to generate the alpha and beta subunits of InR. The structure of Drosophila CG3837, Secreted decoy of InR (Sdr), is distinct from other insulin receptor genes in that it only encodes an alpha subunit. Evidence of the presence of truncated insulin receptor is accumulating in many insect species and evolutionary conservation has been reported (2). Because some of these truncated genes lack the transmembrane and/or tyrosine kinase domains, they could act as decoy receptors. In mammalian species, secreted variant isoforms of receptor tyrosine kinases including truncated variants of the insulin receptor, can be formed by alternative splicing (3).

Insulin/Insulin-like growth factor signaling pathways have primarily been investigated in liver, fat and skeletal muscles as they were reported to be direct target tissues of the signal. Recently the role of insulin signaling in central nervous system (CNS) to integrate nutrients, metabolites and hormonal signal, and to maintain glucose and lipid homeostasis and neurogenesis is emerging (4-7). Nervous system development requires proper organization, specification and interaction between neural and non-neural cells. The blood-brain barrier (BBB) arbitrates communication between the brain and the body. The Drosophila and vertebrate species BBBs are similar in many structural and functional aspects (8, 9). CNS physiology is sensitive to energy balance and it is reported that IGF-1 plays a neuroprotective role (10, 11). Despite its significance, little is known about the mechanism that regulates traffic of growth factors across the BBB. In this paper, we show that Sdr functions in glia, which constitute BBB and BRB, and plays a critical, novel role in maintaining the integrity of these barriers in adult flies.

Lifespan and locomotor activity are reduced in Sdr mutants

Drosophila Sdr is 42-49% homologous to Drosophila InR, human insulin receptor (hIR), and human insulin-like growth factor 1 receptor (hIGF-1R). However, the homology is restricted to the N-terminal ectodomains and Drosophila Sdr lacks the transmembrane and cytoplasmic tyrosine kinase domains found in other members of the IR/InR protein family. Role of Sdr in development or growth had been investigated mostly in embryonic and larval stages, and Sdr is reported to act as an insulin receptor decoy. To further investigate the function of Sdr, we generated an Sdr null allele with a 546 bp deletion by stimulating imprecise excision of the P-element located in the 1st exon of Sdr (location 14918660 according to Refseq NT_033777). The resulting allele, Sdr^D190, lacks 217 nucleotides (nt) in the Sdr 5’UTR and 329 nt from the Sdr coding region, including the translation initiation site (Fig. 1A, B). Two mRNA transcripts were reported to be synthesized from the Sdr region, Sdr-RA and RB, that differ only in the 3’UTR and are predicted to generate the same polypeptide. However, because Sdr^D190 lacks a translation initiation site, flies carrying this allele are expected to lack SDR protein; whether a truncated Sdr transcript would be made was not clear. In fact, RT-PCR analysis using primers that hybridize to sequences in the 4th and 6th exons detected no mRNA from Sdr^D190 in adult flies (Fig. 2A, B). In addition, SDR protein was not detected by immunostaining Sdr^D190 adult tissues (Fig. 2F). These data confirm that Sdr^D190 is a null allele that generates no functional transcript or protein product.

Sdr^D190 flies did not show any obvious morphological defects during developmental stages from egg to adult. However, the lifespan of Sdr^D190 flies was shorter and locomotor activity was impaired relative to wild-type control flies. Although the ability of wild type flies to climb gradually declines with age, Sdr^D190 homozygous mutant flies lost the ability to climb at a much faster rate (e.g. time to 50% maximal climbing ability was 10 days or 24 days for Sdr^D190 or wild type, respectively) (Fig. 1C). The midpoint of the survival curve was 14 or 27 days for Sdr^D190 homozygous and wild type flies respectively, indicating that the lifespan of Sdr null flies was approximately 50% shorter than the lifespan of wild-type flies (Fig. 1D). Sdr^D190 heterozygous flies had similar defects in climbing ability and lifespan, but the effects were milder than in homozygous flies.

Sdr is expressed in the surface glia of the adult brain and eyes

We made polyclonal antibody against SDR protein. Because Sdr is homologous to InR, we tested whether the anti-SDR antibody cross-reacts with InR protein. Immunohistochemical studies were performed on IRP-Gal4>UAS-Sdr and IRP-Gal4>UAS-InR larvae, in which SDR and InR are overexpressed in larval neuroendocrine cells, respectively. Anti-SDR antibody detected ectopic expression of SDR in IRP-Gal4>UAS-Sdr larval brain but did not cross-react with overexpressed InR in IRP-Gal4> UAS-InR. This confirms the specificity of the anti-SDR antibody we generated (Fig. 2C). In wild type adult flies, SDR immunoreactivity was detected in the cortex area which covers central nervous system (i.e., central brain and medulla) and the laminal cortex of the eye. Co-staining with REPO or ELAV which are glial or neuronal markers, respectively, demonstrated glial-specific expression of SDR in wild type (Fig. 2D, E). On the contrary, SDR expression was no longer detected in Sdr^D190 flies (Fig. 2F).

Blood-brain-barrier and blood-retina-barrier defects in Sdr mutant

Because SDR is expressed in the glial cells in the CNS and the eyes of adult flies, we used Sdr^D190 mutant to investigate whether SDR is required for BBB and/or BRB function. The permeability of the BBB and BRB was examined by injecting tetramethylrhodamine-labeled dextran conjugates into the abdomen of adult flies. In healthy wild type flies, the labeled dextran conjugates do not cross the BBB or the BRB. We examined barrier leakage at the time when flies did not show any lethality but showed 50% decrease in climbing ability, the day 10 at 29°C. Significantly more dye penetrated the eyes of 10-day-old Sdr^D190 flies than wild-type flies (Fig. 3A). This phenotype was also observed at day 10 of Sdr^D190 flies raised at 25°C when mutants had normal survival and normal ability to climb. To confirm whether this leakage defect existed from the time when flies were born, we examined 2-day old young mutant flies. Most of the 2-day old young mutant flies did not show leakage of the dye into the eyes and only 30-40% of the flies showed small amount of leakage but at a greatly reduced level: indicating higher integrity of the BRB in 2-day-old Sdr^D190 flies. These results suggest that BRB integrity and function decrease with age in Sdr^D190 flies. This is consistent with the fact that the eyes of Sdr^D190 flies had abnormal patterns of rhodopsin staining (Fig. 3B). Rhabdomeres were significantly shortened and more disorganized than in control flies, and the severity of the defect increased with age.

To investigate the permeability and the integrity of the BBB in Sdr^D190 flies, tetramethylrhodamine-labeled dextran conjugates were injected into the abdomen of adult flies, brains were dissected, and the amount of dye that had crossed the BBB was measured with a spectrofluorometer. The results show approximately 10-fold higher fluorescence in Sdr^D190 brains than in wild type brains (Fig. 3C). These results indicate that the integrity of the BBB and BRB is significantly compromised in Sdr^D190 flies, leaving the brain and eye of these flies susceptible to damage and penetration.

Apoptosis is induced in aged Sdr mutants

To determine whether apoptosis is in progress in Sdr^D190 mutant, we examined expression and activation of markers of apoptosis including Decay (caspase3), Drice and p-JNK. Decay, Drice, and p-JNK were significantly induced in older mutant flies but not in wild type flies. Expression of Drice and p-JNK was higher in the laminal cortex of older Sdr^D190 flies (Fig. 4A, B) but caspase3 was not detected (data not shown). In contrast, caspase3 but neither Drice nor p-JNK was detected in the central brain cortex of Sdr^D190 flies (Fig. 4C). In addition, the nuclei of the caspase3-positive repo-positive Sdr^D190 glial cells appeared to be larger than the nuclei of repo-positive control glial cells, suggesting that nuclear swelling occurred in those caspase-positive repo-positive Sdr^D190 cells.

The human genome encodes 58 RTKs belonging to 20 RTK protein families. The Drosophila genome encodes 20 RTKs belonging to 11 of the 20 RTK protein families found in human (12). Mutations in genes encoding RTKs cause developmental disorders and cancer by altering various cell signaling and biological responses. As the availability of the complete genome sequences of various species has increased, the number of insect taxa with more than two InR copies are growing over the past decade (2). Among these InR paralogs, some encode truncated proteins that lack the tyrosine kinase domain. Both with and without transmembrane domain forms are also found. As they are expressed tissue- and sex-specifically (2), they would have distinct function but their function is mostly elusive and remain poorly characterized. In other animals including human, presence and expression of truncated RTK even in normal tissues has been reported (13, 14).

Recently, numerous effects of insulin signaling in brain are emerging. They include regulation of appetite, body temperature, reproductive function, systemic glucose metabolism (15, 16) and learning and memory (17, 18). Decreased expression of InR has been reported in the brains of animal models with Alzheimer’s disease or Parkinson’s disease, suggesting that defective or decreased insulin signaling could play a role in neurodegenerative disease (7).

In fact, we isolated Sdr mutant when we were searching for genes involved in neurodegeneration in adult flies (data not shown). Neuroprotective role and importance in brain homeostasis of IGF-1 has been reported (11, 19) and IGF can modulate blood vessel growth (20). To perform such functions, serum IGF-1 should cross the BBB in response to electrical, sensory, or behavioral stimuli (21). In general, the concentration of serum IGF-1 is stable; however, when serum IGF-1 increases, its concentration in cerebrospinal fluid is also reported to increase. This suggests the existence of a mechanism that regulates the transport of serum IGF-1 across the BBB.

Although Drosophila Sdr was described previously as a decoy of InR that regulates growth in fly larvae (22), it could also function in other processes and the amount of knowledge for its function is limited. When we examined dissociated primary cultures of embryonic cells, SDR was expressed in a punctate pattern in the cytoplasm of primary cultures of Drosophila embryonic cells (data not shown). Thus, translated SDR seem to exist in vesicular form until they are secreted. Trafficking of SDR-containing vesicles could be regulated to maintain homeostasis of the insulin signaling in the brain. The Drosophila genome encodes 8 insulin-like peptides (ilps). Seven of these peptides, ILP1 to ILP7, are reported to bind InR protein and repress growth inhibition and stress response pathways (23). In contrast, ILP8 binds to relaxin receptor homolog LGR3 to slow growth in response to injury (24, 25). Co-immunoprecipitation studies show that Drosophila SDR interacts with all Drosophila ILPs except ILP4 and ILP8 (22).

Brain activity depends on access to circulating oxygen and nutrients. Endothelial transport of oxygen and glucose and brain activity is facilitated by a neurovascular mechanism (26). Insulin signaling in the adult brain is required as insulin inhibits neuronal apoptosis and tau phosphorylation (19). SDR may continuously bind to and dissociate from ILPs (i.e., like IGFBP), releasing them when they are needed. Alternatively, secretion of SDR might be regulated, so that it is secreted from glial cells when it is needed. As high level of SDR is constantly synthesized in the brain, a high concentration of SDR may be required to rapidly bind to ILPs and block their function. And these processes may be required to maintain BBB and BRB in adult flies.

Longevity and climbing tests

Flies were collected as newborn adults over a 12-hour time window. Ten newborn flies were placed in a fresh vial and stored for 1 day at 25°C before being used in climbing and survival tests. During the experiment, flies were maintained at 29°C. After negative geotactic pressure was applied, the number of flies that climbed to the top of the vial within 20 seconds was counted and recorded.

Production and purification of rat anti-SDR polyclonal antibody

To construct an SDR overexpression plasmid, a 730 nucleotide DNA fragment was amplified by PCR using LD44769 (the EST clone of the SDR 5’ region) as template DNA and Pfu polymerase (Agilent/Strategene) and the following two primers: 5’ CGGGATCCGTGGCAGATGTATC3’ and 5’GGAATTCTGCGA TAGTCGTGATTG3’. BamHI and EcoRI digested PCR product was then cloned into the pRSET A vector. The resulting plasmid was used to express and purify 39 kDa recombinant SDR protein, which was used for primary and booster injections into rats to raise rat anti-SDR polyclonal antibodies. Recombinant SDR was coupled to CNBr-activated sepharose 4B (Amersham Biosciences, NJ, USA), and serum taken from rats after the booster injection was applied to this column. Bound antibodies were eluted to recover affinity-purified rat anti-SDR antibodies. Affinity-purified antibody was tested for specificity and cross-reactivity with Drosophila InR protein, as follows. InR and SDR were overexpressed in neuroendocrine cells of fly larvae using the IRP-Gal4 system (IRP-Gal4>UAS-InR or IRP-Gal4>UAS-Sdr). Ectopic expression of SDR was detected in SDR-overexpressing larvae but no immunoreactivity was detected in InR-overexpressing larvae (i.e., IRP-Gal4>UAS-InR) (Fig. 2). Therefore, the rat anti-SDR antibody is specific for Drosophila SDR and does not cross-react with InR in Drosophila larvae.

Construction of Sdr null mutant and revertant

GE22018 flies were purchased from Genexel (Taejun, Korea). In the genome of GE22018 flies, a P-element is inserted in the 1st exon of SDR at location 14918660 (RefSeq NT_033777). To mobilize the P-element in the GE22018 genome, GE22018 flies were crossed to flies harboring P[Δ2-3] transposase, and progeny of this cross were monitored for loss of eye pigmentation as a marker of P-element excision. To determine whether the P-element was excised precisely or imprecisely, we bred homozygous flies with putative P-element excisions, isolated their genomic DNA, and PCR amplified and analyzed the remaining Sdr gene sequences in each fly strain. The upstream primer used for this analysis started at −112 (relative to the transcriptional start site +1) and the downstream primer started at nt 1557 downstream of the transcriptional start site of Sdr. The DNA sequence of the products of PCR amplification was determined and the P-element excision sites were identified. The DNA sequence of one fly strain revealed imprecise P-element excision. This fly strain, Sdr^D190, had a deletion from nt 14918661 to nt 14919173 of the genomic DNA, according to the Refseq locus AE014297, and it retained 14 nt of P-element DNA (i.e., CATGATGAAATAAC). The DNA sequence of a second fly strain revealed precise excision of the P-element. This fly strain, Sdr^D193 was considered to be a phenotypically normal revertant of the parental GE22018 strain, and where indicated, the revertant strain (Rev) was used as a wild-type control.

Cryosectioning and Immunostaining fly brain

The heads of adult flies were fixed in 4% paraformaldehyde for 10 min, transferred to embedding media, and then cryosectioned at a thickness of 15 μm. Sections were placed on lysine-coated slides (Superior Marienfield, Lauda-Konigshfen, Germany) and immunostained as described in Kim et al. (27). Briefly, slides were fixed in 4% paraformaldehyde for 10 min at room temperature (RT), washed 3× for 5 min with 50 mM Tris-Cl pH7.5, 15 mM NaCl, 0.1% BSA, 0.1% Triton X-100 (1× TNBT), incubated with 5% goat serum for 30 min and then incubated with primary antibody overnight at 4oC. Finally, slides were washed with 1× TNBT for 30 min, and then incubated with secondary antibody for 2 h at RT. Slides were mounted with Vectashield and visualized under a Zeiss Microscope (Axioplan2, Carl-Zeiss, Oberkochen).

Permeability tests of the BBB and BRB

A solution containing tetramethylrhodamine-labeled dextran conjugates (10,000 MW, Molecular Probes) was loaded into a glass capillary prepared with a needle puller (PN-30, Narishige Scientific Instrument Lab, Tokyo), and injected into the abdomen of adult flies (150 nl/fly). After 2 h, injected flies were dissected, and the brains of 2 flies were placed into a single well containing 50 μl 0.1% SDS. The fluorescence from each well was measured using a SpectraMax Gemini EM microplate spectrofluorometer (Excitation = 555 nm, Emission = 580 nm). Each experiment was repeated at least 8 times. The amount of fluorescent dye in eyes of injected flies was estimated by examining eyes under a Zeiss Microscope (Axioplan2, Carl-Zeiss, Oberkochen). Experiments were repeated at least 10 times for each experimental group.

This work was supported by grant NRF-2017R1D1A1B06029547 funded by the Ministry of Education, Science, and Technology of Korea to J. Kim-Ha.

The authors have no conflicting interests.