Extracellular matrix induced by steroids and aging through a G-protein-coupled receptor in a Drosophila model of renal fibrosis

ABSTRACT Aldosterone is produced by the mammalian adrenal cortex to modulate blood pressure and fluid balance; however, excessive, prolonged aldosterone promotes fibrosis and kidney failure. How aldosterone triggers disease may involve actions independent of its canonical mineralocorticoid receptor. Here, we present a Drosophila model of renal pathology caused by excess extracellular matrix formation, stimulated by exogenous aldosterone and by insect ecdysone. Chronic administration of aldosterone or ecdysone induces expression and accumulation of collagen-like Pericardin in adult nephrocytes – podocyte-like cells that filter circulating hemolymph. Excess Pericardin deposition disrupts nephrocyte (glomerular) filtration and causes proteinuria in Drosophila, hallmarks of mammalian kidney failure. Steroid-induced Pericardin production arises from cardiomyocytes associated with nephrocytes, potentially reflecting an analogous role of mammalian myofibroblasts in fibrotic disease. Remarkably, the canonical ecdysteroid nuclear hormone receptor, Ecdysone receptor (EcR), is not required for aldosterone or ecdysone to stimulate Pericardin production or associated renal pathology. Instead, these hormones require a cardiomyocyte-associated G-protein-coupled receptor, Dopamine-EcR (DopEcR), a membrane-associated receptor previously characterized in the fly brain to affect behavior. DopEcR in the brain is known to affect behavior through interactions with the Drosophila Epidermal growth factor receptor (Egfr), referred to as dEGFR. Here, we find that the steroids ecdysone and aldosterone require dEGFR in cardiomyocytes to induce fibrosis of the cardiac-renal system. In addition, endogenous ecdysone that becomes elevated with age is found to foster age-associated fibrosis, and to require both cardiomyocyte DopEcR and dEGFR. This Drosophila renal disease model reveals a novel signaling pathway through which steroids may modulate mammalian fibrosis through potential orthologs of DopEcR.


Introduction
Aldosterone is a primary renal regulator of sodium and potassium homeostasis, but when chronically elevated as in diabetes and primary aldosteronism (1), aldosterone promotes kidney interstitial fibrosis and glomerulosclerosis (2)(3)(4). These events are preceded by elevated inflammation through monocytes and macrophage infiltration followed by proliferation of myofibroblasts that secrete fibrinogen, collagens and elastins. Aldosterone increases reactive oxygen species (ROS) to induce profibrotic factors such as Transforming Growth Factor-1 (TGF-1), Plasminogen Activator Inhibitor-1, and Enothelin-1 (4). TGF-1 contributes to fibrosis by activating myofibroblasts (5) as well as through suppressing matrix metalloproteinases, which can further promote excess extra-cellular matrix (6). Aldosterone affects these processes through its interaction with the mineralocorticoid nuclear hormone receptor (MR), as inferred from studies where blockade of MR activity prevents aldosteroneassociated inflammatory and fibrotic outcomes (7)(8)(9).
Many data also suggest that aldosterone contributes to fibrosis through rapid signaling independent of MR (4). Aldosterone enhances TGF-1 expression and fibrosis in part through stimulation of ERK1/2 (10)(11)(12), while aldosterone fosters hypertrophy in cardiomyocytes through action on ERK5 and PKC (13). As well, aldosterone effectively induces calcium influx in fibroblasts derived from MR-deficient mice (14). Angiotensin receptors crosstalk with MR to modulate NF-B in vascular smooth muscle cells (VSMC) stimulated with aldosterone (15), suggesting that aldosterone can in part act through G-protein-coupled receptors (GPCR). With considerable debate, GPER1 has been proposed as an alternative GPCR for aldosterone (16)(17)(18)(19). In VSMC, aldosterone was seen to activate PI3 kinase and ERK through both GPER1 and MR (20). Emerging evidence, however, shows that 17-estradiol is the steroid agonist of GPER1 (21)(22)(23), and no pharmacological evidence demonstrates GPER1 to interact with aldosterone. The problem remains: through which receptor aside from MR might aldosterone stimulate signaling, is this a GPCR, and how does this modulate fibrosis?
Here we develop a model of steroid-induced fibrosis based on Drosophila melanogaster. Genetic data reveal the Drosophila GPRC Dopamine-EcR (DopEcR) (reviewed in (24)) is expressed in cardiomyocytes, and is necessary for exogenous aldosterone and insect ecdysone to induce excess extracellular matrix at heart-associated nephrocytes, and to disrupt fly renal function. We likewise document elevated cardiac-renal fibrosis with age and find this pathology requires endogenous synthesis of ecdysone and cardiomyocyte DopEcR. Similar requirements are found for Drosophila EGFR (dEGFR) in terms of exogenous hormone treatments and endogenous aging. Based on our findings we propose that mammalian homologs of DopEcR may offer a novel entrée to understand fibrotic pathology in humans.

Steroid hormones induce renal dysfunction at the nephrocytes
The tubular heart of adult Drosophila is lined by pericardial cells, podocyte-like nephrocytes that conduct size-selective filtration of hemolymph (25,26) (Fig. 1A). The heart tube and the associated nephrocytes are enmeshed in an extracellular matrix composed of collagen-like proteins including Pericardin (collagen IV) (27,28). In a first step to develop a model of Drosophila renal fibrosis, we measured protein in adult excreta (frass) as an analog to proteinuria seen in humans with glomerular dysfunction (29). Frass is a by-product of both digestion and discharge from renal Malpighian tubules, gut-associated structures that maintain ionic and water balance (25,26,30,31). Previous work shows the appearance of frass can be modulated by diet, mating and internal metabolic state (32), and by the activity of heartassociated nephrocytes (33,34). We asked if frass protein content could be affected by nephrocyte function. We collected frass from adult males (to exclude eggs) in microcentrifuge tubes and measured total protein content, normalized to uric acid as a way to account for excretion volume. To manipulate nephrocyte function, we depleted nephrocyte slit diaphragm genes kirre and sticks-n-stones (sns), which encode homologs of mammalian nefrin. Previous reports show that reduced kirre and sns impairs nephrocyte filtration measured by uptake of fluoro-dextran beads (26,31). We replicated this result (Fig 1E, F) and likewise observed that reduced kirre and sns also elevated protein excretion ( Fig 1B). Thus, defects in nephrocyte function can induce proteinuria in Drosophila.
We next assessed how frass protein content was affected by nutrient and physiological conditions as occurs with human chronic kidney disease. Diets of high sugar or salt decreased protein excretion compared to normal diet (Fig 1C), perhaps by altering adult metabolic state.
To find a treatment that might increase proteinuria, we fed aldosterone to adult Drosophila.
Protein in frass was elevated in adults fed aldosterone for two weeks (Fig 1D) yet not when fed aldosterone for only 24 h (Fig S1). Drosophila do not synthesize aldosterone, a mammalian steroid hormone (Fig 1A) produced in the renal cortex. Rather, aldosterone likely acts in Drosophila as a mimic of insect steroids (Fig 1A) or by providing a precursor for the synthesis of insect steroids. 20-hydroxyecdyone (20E) is the primary active steroid in Drosophila. 20E is oxidized from the prohormone ecdysone by 20-hydroxylase (encoded by shade) at target cells. 20E activates the nuclear hormone Ecdysone Receptor (EcR) to modulate transcription.
Interestingly, feeding adults 20E for two weeks did not stimulate proteinuria, but proteinuria was elevated in adults chronically fed ecdysone ( Fig 1D). Likewise, chronic aldosterone and ecdysone, but not 20E, suppressed dextran filtration by nephrocytes ( Fig 1G). While only aldosterone and ecdysone affected nephrocyte function and associated proteinuria, all tested steroids (aldosterone, ecdysone and 20E) reduced survival of adults on high salt diet (Fig 1H), indicating that each exogenous hormone has some capacity to impart biological activity. We found no consistent association between exogenous steroids and adult survival on normal diet ( Fig 1I).

Elevated extracellular matrix drives renal dysfunction
Pericardial nephrocytes and the heart tube are surrounded by extracellular matrix made of collagen-like proteins including Pericardin (Fig 1A), col4a1 and Viking (27,28,35).
Adults fed aldosterone and ecdysone for 24 hours induced pericardin (prc) mRNA in their cardiac-nephrocyte tissue, but not when fed 20E (Fig 2A). Collagen encoding-transcripts col4a1 and Viking mRNA were not induced by any of these steroids (Fig 2B, C). Despite induction of prc mRNA, overnight steroid feeding itself did not elevate proteinuria ( Fig S1). In contrast, aldosterone and ecdysone fed to wildtype adults for two weeks had elevated extracellular matrix Pericardin protein (PRC) around the cardiac-nephrocyte complex ( Fig   2D,E). Depletion of pericardin mRNA from cardiomyocytes (tin4-gal4>prc(RNAi)) (efficiency in Fig S2) but not from nephrocytes (sns-gal4>prc(RNAi)) blocked the ability of aldosterone and ecdysone to induce excess PRC deposition (Fig 2D, E). We also determined that pericardin expression in cardiomyocytes was necessary for aldosterone and ecdysone to induce proteinuria and to repress nephrocyte filtration: depletion of prc mRNA from cardiomyocytes blocked the ability of aldosterone and ecdysone to induce pathology, while depletion of prc mRNA in nephrocytes did not (Fig 2F-K). In contrast, exogenous 20E continued to produce no effects on fibrosis or nephrocyte function, independent of prc knockdown ( Fig 2D, F-K). Thus, cardiomyocytes appear to be the source of Pericardin protein that accumulates in response to chronic exposure to aldosterone and ecdysone, and impairs nephrocyte function.

The GPCR dopEcR is required for steroids to drive fibrosis
It is striking that ecdysone but not 20E induces pericardin expression and associated renal pathology in Drosophila. This suggests that PRC protein in the ECM can be regulated independently of EcR, the canonical nuclear hormone ecdysone receptor of 20E. Indeed, depletion of EcR by RNAi in cardiomyocytes did not prevent the steroid-dependent induction of prc mRNA (Fig 3A), or associated ECM accumulation ( An alternative avenue for action involves Dopamine-EcR (DopEcR, CG18314), a membrane G-protein-coupled receptor (GPCR) of ecdysone that has been described in the fly brain (36,37) (24,38). We detected DopEcR mRNA in adult cardiac-nephrocyte tissue, and more so in adults fed aldosterone and ecdysone ( Fig 3G). Consistent with a model where DopEcR is required for aldosterone and ecdysone to stimulate renal pathology, cardiomyocyte-specific knockdown of DopEcR (via tin4-Gal4>DopEcR(RNAi)) blocked the ability of aldosterone and ecdysone to induce prc mRNA expression (Fig 3B), elevate proteinuria and inhibit nephrocyte filtration (Fig 3D, F). Likewise, DopEcR in cardiomyocytes is required for aldosterone and ecdysone to induce excess Pericardin protein (Fig 3H, J). In contrast, while elevated deposition of PRC was prevented by cardiac-specific KD of DopEcR, PRC was not inhibited in flies with nephrocyte-specific DopEcR or EcR knockdown ( Our work suggests that ecdysone acts as an agonist of DopEcR in the heart where receptor activation modulates organ fibrosis. Nevertheless, it is possible that treatment with exogenous ecdysone antagonizes production of endogenous steroids (E or 20E), and that this loss promotes fibrosis. If true, knockdown of endogenous ecdysone should itself promote fibrosis, and addition of exogenous ecdysone should not further increase fibrosis. To test this hypothesis, we employed a mutant of the nuclear zinc finger protein encoded by molting defective, DTS-3/mld (39), which inhibits transcription of enzymes required for endogenous ecdysone synthesis. DTS-3 flies grow and emerge normally at 18°C, while adults switched to 29°C produce little ecdysone. We grew cohorts of DTS-3 females and control-wildtype females following these temperature regimes. As expected, control-wildtype females showed little prc mRNA and PRC until treated with exogenous ecdysone (Fig 4F-H). Yet, contrary to the hypothesis, DTS-3 females, with anticipated low endogenous levels of ecdysone, also had low levels of prc mRNA and PRC until stimulated by exogenous ecdysone (Fig 4F-H).
Exogenous ecdysone appears to positively promote fibrosis rather than act by repressing production of endogenous steroids.

Renal fibrosis naturally occurs with age and is modulated by ecdysone and dopEcR
To this point we have induced fibrosis by treating flies with external steroids, begging the question, what is the physiological relevance in fibrosis of endogenous ecdysone acting through DopEcR and dEGFR? Endogenous ecdysone is normally elevated in aging Drosophila, where whole animal titers increase several fold between young and aged flies (40). As well, Vaughan at al. (41) found the collagen Viking accumulates in Drosophila cardiac ECM with age. We therefore measured Pericardin protein in the heart-nephrocyte ECM of untreated young and aged flies. Pericardin increased about 2-fold in 6-week old flies relative to young adults (Fig 4I, J). This age-dependent fibrosis was prevented by knockdown of endogenous ecdysone synthesis in adults, using the DTS-3 system as above (Fig 4I, J).
Likewise, knockdown of both DopEcR and dEGFR in cardiomyocytes prevented fibrosis in the aged flies (Fig 4I, J). These results suggest that PRC accumulation in the nephrocyte and cardiac-associated extracellular matrix is an intrinsic property of aging flies promoted by endogenous ecdysone acting through cardiac DopEcR and EGFR receptors.

Discussion
Mammalian aldosterone is synthesized from cholesterol in the adrenal cortex as a 21-carbon, C21-hydroxyl steroid to control plasma Na + and K + , water balance and blood pressure. Insect ecdysone is a 27-carbon steroid with hydroxyl groups at C21 and C27 (Fig 1A). Adult Drosophila produce ecdysone in ovaries and several somatic tissues including the Malpighian tubules (40,42). Circulating ecdysone is converted at target cells into 20-hydroxyecdysone (20E), which induces transcriptional programs by activating the nuclear hormone Ecdysone Receptor EcR. Our data show that exogenous aldosterone and ecdysone, but not 20E, stimulate deposition of PRC in adult heart-nephrocyte extracellular matrix acting through the G-protein coupled receptor DopEcR and not the canonical nuclear hormone receptor EcR.
How aldosterone mimics ecdysone in this context remains unknown. Work is needed to determine if aldosterone has affinity to DopEcR, or if aldosterone acts as precursor molecule that can be converted to ecdysone within Drosophila. We likewise do not understand why exogenous 20E does not stimulate fibrosis whereas ecdysone produces a strong response.
Previous work found that 20E as well as ecdysone has affinity for DopEcR in isolated Sf9 cell membranes (43), while exogenous 20E modulates DopEcR activity measured from fly brain cAMP levels, by brain nicotine-induced Ca 2+ -responses, and by adult behavior (36)(37)(38). It also remains to determine what roles ecdysone plays in the regulation of Pericardin at the heart during normal development; perhaps, we suggest, it facilitates cardiac remodeling during molt and pupation (44).
Ecdysone circulating in adult hemolymph may act at many sites aside from EcR in fat body and ovary (45), or from DopEcR in the fly brain (36,37). Our genetic results indicate that DopEcR message is required specifically in cardiomyocytes to modulate steroid-induced fibrosis. Using newly emerging tools well suited to study GPCR in Drosophila, we anticipate future work can directly identify which cells in this heart produce functional dopEcR proteins (46,47). Fibrosis in human hearts arises from myofibroblasts that secrete extracellular matrix proteins including fibronectins, elastins and collagens (48)(49)(50). Based on these parallels, we propose Drosophila cardiomyocytes and mammalian myofibroblasts have analogous functions to produce ECM.
We find that chronic induction of pericardin by steroid hormones stimulates excess PRC protein in the ECM surrounding the myocardial-nephrocyte cells, induces proteinuria and inhibits nephrocyte filtration. Excess heart-associated ECM was previously reported in aged Drosophila, measured by accumulation of Pericardin and the collagen subunit Viking (41).
Here we also find PRC increases in cardiac-nephrocyte ECM of old females. Remarkably, systemic knockdown of adult ecdysone synthesis, which otherwise increases with age (40), prevents elevated Pericardin in aged females, as does cardiomyocyte knockdown of DopEcR (and EGFR). From our observation that steroids elevate pericardin mRNA, we propose that DopEcR promotes fibrosis during aging by inducing pericardin mRNA, and subsequent translation and secretion of Pericardin, rather than by modulating ECM breakdown.
Activation by dopamine induces cAMP-mediated signal transduction. Ecdysone has greater affinity to DopEcR than does dopamine, and through unknown mechanisms will displace dopamine and induce alternative signal transduction mediated by MAP kinases (43) (38).
Reports are mixed on whether ecdysone also affects cAMP via DopEcR because dopamine alone can increase cAMP in Sf9 cells expressing DopEcR (37,43). In mammalian cells, cAMP can induce PKA to phosphorylate CREB, which then localizes to promoters. Human CREB targets include several collagen genes, and cAMP stimulation suppresses collagen-I expression in a CREB dependent manner (52)(53)(54). Accordingly, we hypothesize that dopamine-cAMP-associated transduction initiated from DopEcR may negatively regulate pericardin.
In contrast to the potential action of dopamine, DopEcR stimulated by ecdysone can signal through dEGFR to ERK1/2 as seen in transfected Sf9 cells and in a neuronal analysis of ethanol induced sedation (36,43). We now find myocardial dEGFR is also required for steroids to induce PRC and nephrocyte dysfunction, and for PRC to accumulate with age. In humans, EGFR signaling is a crucial regulator of fibrosis (55,56). The EGFR ligands TGF and epidermal growth factor are expressed in kidney cells where activated EGFR stimulates extracellular signal-regulated kinases1/2 (ERK1/2), Janus kinase/signal transducers and activators of transcription, and PI3-kinase/AKT. In renal interstitial fibrosis, EGFR regulates TGF-1 via ERK1/2 to activate myofibroblasts and promote expression of ECM collagens (57).
Notably, EGFR can be transactivated independent of its extracellular ligands, including by the activity of G-protein-coupled receptors (GPCR) such as the Angiotensin II receptor, and this action is mediated intracellularly by the sarcoma kinase Src. Furthermore, Src-mediated transactivation has been shown to accentuate renal fibrosis in mammals (58)(59)(60). Based on our current observations, we hypothesize that ecdysone-stimulated DopEcR might stimulate dSrc to facilitate ligand activation of dEGFR (Src42, (61)).
Studies in mammals suggest aldosterone may also signal via a membrane associated GPCR. GPER1 has been proposed to function as a non-genomic aldosterone receptor and as a potential homolog of DopEcR (21,22,62). GPER1-dependent induction by aldosterone is reported in renal cortical adenocarcinoma cells (17), and from mouse models with tissue specific mineralocorticoid receptor gene deletion (63). However, no data establish a mechanism of non-genomic action for aldosterone through GPER1 (23,64), and the current steroid candidate for GPER1 is 17-estradiol (65). Using the DIOPT Ortholog Prediction Tool, we identified several potential alternatives for the DopEcR homolog in the human genome including GRP52 (sequence similarity 46%) and UTS2R (sequence similarity 44%). GPR52 is an orphan G-protein coupled receptor described to modulate Huntingtin protein (HTT) through cAMP-dependent mechanisms (66). Knockdown of Gpr52 reduces HTT levels in a human tissue model, whereas neurodegeneration is suppressed by knockdown of DopEcR in Drosophila that express human Htt. The Urotensin II receptor (UTS2R) is a conserved GPCR implicated to function in renal fibrosis by trans-modulating EGFR and activating MAPK (67,68). The kidneys of diabetic rats express elevated Urotensin II, and UTS2R is required for exogenous Urotensin to induce TGF-1 and collagen in the renal ECM. If functional homology can be established between DopEcR and these mammalian candidates, Drosophila will provide a new model system to uncover mechanisms of fibrosis in humans.

Material and methods
Fly stocks. Unless noted, wildtype flies were yw (ywR). Tin4-Gal4 was a gift from the Manfred Frasch laboratory (69). sns-Gal4 was obtained from the Bloomington Stock Center old, then exposed to diets with appropriate hormone conditions for 24 hours. In all trails, renal traits and prc mRNA were assessed in adults at 3 weeks old. The same protocols were used to expose adults to high salt or high sugar, where instant media was moistened with water containing 1.5% NaCl. To vary dietary glucose, adults were aged to 3 weeks on otherwise standard lab diet where glucose was set at 5% (control, normal) or at 34% (high sugar diet).

Proteinuria.
For each biological replicate, frass of 15 males was collected for 2.5 hours in a 1.5ml centrifuge tube covered with a breathable foam plug, at 25 °C. Males were used in this assay to avoid complications of eggs also laid in the tubes by females. Deposited frass was fully dissolved with 20ul 1xPBS, providing 10ul to assess total protein and 10ul to measure uric acid, which serves as a proxy for the quantity of deposited frass. Total urine protein was determined by Pierce BCA Protein assay (Thermo Scientific #23227). Uric acid was measured by QuantiChrom Uric Acid Assay (Bioassay systems, DIUA-250).

Disease Models & Mechanisms • DMM • Accepted manuscript
Alexa555-phalloidin 1:100, ThermoFisher Scientific) diluted in PBTA overnight, washed 3x10 minutes with 1ml PBTA at room temperature, and mounted. Confocal images were obtained with a Zeiss 800 and quantified by imageJ software. The full length of the heart tube, pericardial cells and associated ECM network was imaged from all samples at 488 nm with the same laser intensity setting to produce a Z-stack comprised of 46 optical slices.     Dextran filtration as a measure of nephrocyte function was reduced in flies fed E or Aldo; this reduction was blocked by cardiac dEGFR knockdown. (D) prc mRNA, relative to Rp49, was induced by E or Aldo feeding (each genotype, n=3 biological replicates of 10 pooled tissues).