ELMO1 protects renal structure and ultrafiltration in kidney development and under diabetic conditions

Engulfment and cell motility 1 (ELMO1) functions as a guanine exchange factor for Rac1 and was recently found to protect endothelial cells from apoptosis. Genome wide association studies suggest that polymorphisms within human elmo1 act as a potential contributing factor for the development of diabetic nephropathy. Yet, the function of ELMO1 with respect to the glomerulus and how this protein contributes to renal pathology was unknown. Thus, this study aimed to identify the role played by ELMO1 in renal development in zebrafish, under hyperglycaemic conditions, and in diabetic nephropathy patients. In zebrafish, hyperglycaemia did not alter renal ELMO1 expression. However, hyperglycaemia leads to pathophysiological and functional alterations within the pronephros, which could be rescued via ELMO1 overexpression. Zebrafish ELMO1 crispants exhibited a renal pathophysiology due to increased apoptosis which could be rescued by the inhibition of apoptosis. In human samples, immunohistochemical staining of ELMO1 in nondiabetic, diabetic and polycystic kidneys localized ELMO1 in glomerular podocytes and in the tubules. However, ELMO1 was not specifically or distinctly regulated under either one of the disease conditions. Collectively, these results highlight ELMO1 as an important factor for glomerular protection and renal cell survival via decreasing apoptosis, especially under diabetic conditions.

podocytes in patients suffering from the microvascular complication 5 . Hyperglycaemia is also known to induce and subsequently increase apoptosis within renal glomerular cells, albeit via the activation of endothelial thrombomodulin-protein C system, which contributes to the pathogenesis of diabetic nephropathy 6 . Hence, this study evidentially demonstrates, that hyperglycaemia induces an increase in apoptosis within the human glomerular endothelial cells and podocytes. However, a better insight must be obtained with respect to the detrimental effects of increased apoptosis in the renal glomerular cells and the subsequent onset of diabetic nephropathy.
Genome wide association studies (GWAS) indicate that polymorphisms in a multitude of genes were associated with the heightened incidence of nephropathy in diabetic patients 7 . To develop new preventative and therapeutic strategies, these polymorphisms need to be studied carefully. Particularly, single nucleotide polymorphisms (SNPs) in human elmo1 are strongly suggested to be a potential contributing factor for the development of diabetic nephropathy, in various global populations [8][9][10][11] . The ELMO1/DOCK180 complex functions as a guanine nucleotide exchange factor for the small GTPase Rac1, thereby regulating cell migration 12 . ELMO1 is expressed in the vasculature in zebrafish and loss-of-function experiments identified an important role for ELMO1 in early developmental vascular processes 13 . Recently, ELMO1 was shown to protect endothelial cells from apoptosis via activation of the Rac1/PAK/AKT signalling cascade in vitro and in vivo 14 . ELMO1 has also been implicated in the development of ciliopathy-related phenotypes in zebrafish, by regulating Ezrin phosphorylation 15 . However, the functional role of ELMO1 in human and zebrafish kidneys, with respect to diabetic nephropathy, is currently unknown and is in the need to be established.
The zebrafish is a versatile model organism used to carry out research on various biomedical fields, including diabetes induced complications within the vasculature 16 . The renal system within a zebrafish embryo consists of a pronephros which, although simple in structure, functions to maintain homeostatic balance of ion and metabolite concentrations, and removal of toxic substances in the blood, similar to the human nephron 17 . Thus, this study aims to establish the functional role of ELMO1 in the zebrafish pronephric glomerulus, under hyperglycaemic conditions. We found that, in the zebrafish pronephros, ELMO1 protects the glomerulus from apoptosis and from hyperglycaemia induced damage. These findings are also supported by immunohistochemical analysis of human kidney sections, obtained from patients suffering from prolonged hyperglycaemia, due to diabetes mellitus type 2.

Results
Human diabetic patients and hyperglycaemic zebrafish illustrate no change in ELMO1 expression in the kidney. ELMO1 expression was analysed for potential changes within human kidney sections received from nondiabetic, diabetic (type 2) and polycystic kidney disease patients ( Fig. 1 and Supplementary Figure 1). Glomeruli in nondiabetic kidneys also showed ELMO1 positive reactivity in podocytes (Fig. 1D,E), but expression was almost absent in glomerular endothelial cells (Supplementary Figure 2D,E). It was also observed, that under nondiabetic conditions, ELMO1 was also expressed within the tubules. Remarkably, ELMO1 expression remained unchanged within the kidney of patients suffering from type 2 diabetes, although the renal expressional pattern of ELMO1 was similar to that of nondiabetic patients (  Figure 1C). Thus, histological results indicate that ELMO1 is equally expressed in different kidney diseases and not differentially regulated, due to a specific pathological condition.
To induce hyperglycaemia in zebrafish embryos, the morpholino technology was employed to knockdown PDX1 transiently 16 . Subsequently, we collected pronephric cells from the Tg(wt1b:EGFP) zebrafish line ( Fig. 2A) subjected to hyperglycaemia and performed a real-time quantitative PCR, to assess expressional changes of ELMO1 within the renal system of the zebrafish embryo under normo-and hyperglycaemic conditions. It is important to note that only podocytes are labelled with EGFP within the glomerulus in the Tg(wt1b:EGFP) zebrafish embryos 18 . No significant change in ELMO1 expression was observed (Fig. 2B,C) within the pronephros of hyperglycaemic embryos as compared to the control. These findings indicate that the data on the zebrafish embryos is relatable to patient data, where ELMO1 expression remains unaltered in diabetic patients (Fig. 1).
Hyperglycaemia and ELMO1 CRISPR injection lead to adverse changes in the zebrafish pronephros. We next aimed to assess the role of ELMO1 in renal structural and functional development, under hyperglycaemic conditions. Consequently, effects of PDX1 knockdown and knockout of ELMO1 were analysed with respect to the zebrafish pronephros (Fig. 3). As compared to the control (Fig. 3A), the PDX1 morphants, at 48 hpf, exhibited an enlarged glomerulus, where the width was significantly increased from 60 μ m in the controls to 75 μ m in PDX1 morphants (Fig. 3B,E). The neck of the PDX1 morphants was greatly shortened to 50 μ m in 4 ng injected PDX1 morphants and 37 μ m in 6 ng injected PDX1 morphants, as compared to a length of 63 μ m in the controls (Fig. 3B,F).
CRISPR/Cas9 mediated ELMO1 knockout (Supplementary Figure 3) was carried out and embryos aged 48 hpf were studied for renal phenotypes. Similar to hyperglycaemia induced changes, ELMO1 crispants exhibited alterations in the zebrafish pronephric structure. As compared to the control (Fig. 3C), ELMO1 crispants displayed a significant increase in width (from 60 μ m in controls to 76 μ m in the ELMO1 crispants) and strong reduction of the neck (from 80 μ m in controls to 40 μ m in the ELMO1 crispants) (Fig. 3D,G,H). To evaluate hyperglycaemia and ELMO1 knockout effects on renal functionality, we assessed pronephric ultrafiltration (Supplementary Figure 4). The assay to determine changes in the functionality of the zebrafish pronephros has already been established 19 . Upon heart injection of the 70 kDa Texas-Red ® dextran, a significantly increased loss of heart fluorescence over time, in PDX1 morphants and the ELMO1 crispants, was observed as compared to their controls (Fig. 3I). This indicates an adversity in the ultrafiltration, in both the hyperglycaemic fish and ELMO1 crispants. Thus, hyperglycaemic conditions and ELMO1 knockout, not only alter pronephric structure, but also adversely affect the function of the renal system, within zebrafish embryos. Hyperglycaemia induced pronephric structural and functional alterations within zebrafish embryos can be significantly rescued via ELMO1 overexpression. As hyperglycaemia and knockout of ELMO1 have similar detrimental effects on the zebrafish pronephric structure and function, we aimed to establish whether overexpression of ELMO1, could rescue hyperglycaemic renal alterations as shown in Fig. 3. Overexpression of ELMO1 13,14 (Supplementary Figure 6) could beneficially restore the compromised pronephric structure within PDX1 morphants ( Fig. 4A-D). The structural changes represented by glomerular enlargement and reduced pronephric neck length in PDX1 morphants (Fig. 4B,E,F) were almost entirely normalized upon the co-injection of ELMO1 mRNA. Similarly, increased clearance of the 70 kDa Texas-Red ® dextran within PDX1 morphants was decreased to levels corresponding to controls, upon ELMO1 overexpression (Fig. 4G). Thus, under hyperglycaemic conditions, overexpression of ELMO1 was found to rescue the altered renal structure and function within zebrafish embryos.
Podocyte foot processes are detrimentally affected in PDX1 morphants and ELMO1 crispants, altering the glomerular filtration barrier. Since the renal filtration barrier is drastically altered in a diabetic nephropathy setting 4 , we aimed to determine if at an electron microscopic level, PDX1 morphants and ELMO1 crispants were affected, especially with respect to podocyte structure and the formation of the ultrafiltration barrier. Consequently, we found that PDX1 morphants (48 hpf) display a highly dystrophic filtration barrier as compared to their controls (Fig. 5A,C), where the podocyte structure contributed to a highly under-developed filtration barrier. Foot processes between adjacent podocytes did not interdigitate appropriately and urinary space was absent. ELMO1 crispants (48 hpf) also exhibited inappropriate podocyte development and filtration barrier, as compared to the control (Fig. 5B,D). Specifically, in ELMO1 crispants foot processes did not develop and interdigitate to form the filteration barrier and podocytes appeared unable to filter the blood plasma (Fig. 5D). Thus, it is further established that, even at a subcellular level, hyperglycaemia and ELMO1 CRISPR injections lead to podocyte structural adversities.
Pancaspase inhibitor, zVAD-fmk, treatment can restore altered structural and functional changes in ELMO1 CRISPR injected embryos. Recently, ELMO1 has been established to protect endothelial cells from apoptosis during angiogenesis 14 . It was hypothesized that a similar function is being performed by the protein in the renal system. zVAD-fmk has been previously utilized as an apoptotic inhibitory compound on zebrafish embryos 20 . Anti-apoptotic treatment of zebrafish embryos was carried out at the age of  24 hpf for 24 hours, in order to observe any changes in the pronephros (Fig. 6). It was observed that upon treatment with zVAD-fmk, ELMO1 crispants showed a significant rescue of pronephric structure as compared to those embryos without treatment (Fig. 6A,B,D,E,G,H). The zVAD-fmk treatment reduced significantly the width of the enlarged glomerulus from 81 μ m to 60 μ m in length, and restored the length of the pronephric neck of the compromised ELMO1 embryos from 18 μ m to 75 μ m in length, which are comparable to the control embryos ( Fig. 6G,H). Ultrafiltration was also analysed in the ELMO1 crispants upon zVAD-fmk treatment. zVAD-fmk treatment of the ELMO1 CRISPR injected embryos also considerably rescued the fluorescence loss from the heart as compared to those ELMO1 crispants without treatment (Fig. 6I). Interestingly, it was found that the pronephros in PDX1 morphants displayed no signs of improvement in structural and functional integrity, upon pancaspase inhibitor treatment (Fig. 6C,F,G,H,J). Thus, the loss of ELMO1 causes structural and functional phenotype within the zebrafish embryos due to an increase of apoptosis, which can be rescued upon pancaspase inhibitor treatment. However, hyperglycaemia induced pronephric changes within the zebrafish embryos cannot be rescued with zVAD-fmk treatment alone, owing to the probable activation of other detrimental mechanisms 21 .
Finally, we aimed to establish if ELMO1 knockout indeed causes an increased incidence of apoptotic cells. The TUNEL assay was employed on ELMO1 crispant embryos, in order to visualize any significant increase in apoptosis universally and especially within the pronephros (Fig. 7). ELMO1 crispants displayed a significantly much higher incidence of apoptotic cells as compared to their controls, at 48 hpf (red dots, Fig. 7A,B). This increase in number of apoptotic cells is not only observed throughout the embryos, but also within the pronephros in the ELMO1 crispants as compared to their controls (Fig. 7E). This data has been further demonstrated with an activated Caspase3 assay (Supplementary Figure 7). Furthermore, zVADfmk treatment of ELMO1 crispants not only restored pronephric structural adversities but also drastically reduced the incidence of apoptotic cells within these crispants (Fig. 7D,E). Hyperglycaemia did not lead to an increase in apoptotic cells within the pronephros (Fig. 7C,E). Thus, it is evident that ELMO1 knockout in zebrafish embryos leads to an increase in apoptosis within the entire organism, and particularly within the renal system.

Discussion
In this study, we have established an important role of ELMO1 in renal function, under hyperglycaemic conditions, within the zebrafish and successfully translated our finding to the pathogenesis of human diabetic nephropathy (Fig. 8). ELMO1 not only shields endothelial cells from apoptosis 14 , but also functions to protect glomerular cells from the phenomenon, under high glucose conditions. Hyperglycaemic conditions in zebrafish embryos caused by PDX1 knockdown 16 , do not induce ELMO1 expression within the pronephros. This coincides with our data from type 2 diabetic patients, where no change in ELMO1 expression was observed within the kidney, compared to nondiabetic and polycystic kidneys, respectively. Thus, our study suggests that the work performed  Upon treatment with zVAD-fmk (300 μ M), pronephric structural defects in ELMO1 crispants at 48 hpf, are significantly restored to normalcy (A,B,D,E); n = 47 for each condition). The pancaspase inhibitor causes no adversity in the control embryos at 48 hpf (D). However, no significant restoration in pronephric structure is observed with zVAD-fmk treatment of PDX1 morphants, at the same age (C,F); n = 47 for both conditions). The white line in image (F) represents the scale bar corresponding to 200 μ m. The significant restoration of the glomerular width and neck via zVAD-fmk treatment in ELMO1 crispants as well as the lack of rescue in PDX1 morphants are quantified in graphs (G,H). Altered ultrafiltration in ELMO1 crispants is also improved significantly upon treatment with zVAD-fmk as seen in graph (I) but not in PDX1 morphants, as observed in (J) The n values for each sample set used for the functional assay are as follows: Co. CRISPR (22), Co. CRISPR with zVAD-fmk (23), ELMO1 CRISPR (22) on zebrafish embryos, can be translated to a clinical setting, where ELMO1 could protect the diseased kidney from increased apoptosis, irrespective of the underlying renal disease. Overexpression of ELMO1 within mammalian (rat) glomeruli has been previously linked to extracellular matrix protein accumulation under diabetic conditions, which is predicted to be a detrimental cause, leading to the pathogenesis of diabetic nephropathy 11 . However, several GWAS studies have indicated SNPs within the elmo1 gene relate to the susceptibility of patients suffering from diabetes to develop nephropathy, in various human populations [8][9][10] . These studies thus strongly indicate that mutations within the elmo1 gene may lead to pathogenesis of DN, thereby suggesting a protective function of this protein concerning the disorder. In accordance with the GWAS indication, we have established that overexpression of ELMO1 actually helps to protect renal structure from pathogenesis under hyperglycaemia. PDX1 knockdown causes similar detrimental alterations within the pronephric structure and function in zebrafish embryos, as observed within the pronephros of ELMO1 crispants. Furthermore, adversities within the renal structure of the zebrafish PDX1 morphants are significantly restored upon simultaneous overexpression of ELMO1. These findings lead to the conclusion that the increase in ELMO1 expression would be a protective response of the pronephros to try and compensate against the harmful effects of the potentially damaging hyperglycaemia. To additionally strengthen this inference, ELMO1 knockout was found to be damaging to the zebrafish pronephric structure and function.
It is known previously that podocyte-specific deletion of Rac1 in mice, a protein activated by ELMO1 and Dock180 interaction, causes foot process effacement of these cells; but only leads to loss of podocytes under hyperglycaemic conditions 22 . Since ELMO1, the upstream regulator of Rac1, is associated with endothelial cell protection against apoptosis and with pathogenesis of DN, we predicted that ELMO1 plays the same survival role in the renal system subjected to physiological and hyperglycaemic conditions. Further investigation of ELMO1 crispants displayed a significantly higher incidence of apoptotic cells within the whole embryo and especially within the pronephric structure, which we suggested to be the cause of structural and functional renal adversities. To confirm this hypothesis, treatment of ELMO1 crispants with the pancaspase inhibitor, zVAD-fmk, showed restoration of the renal disruptions, both structural and functional. Therefore, it is evident that ELMO1 protects renal cells from increased apoptosis within zebrafish embryos. Thus, we expect that an upregulation of ELMO1 in the diabetic patients would likely protect the nephron from increased apoptosis, ultimately delaying the onset of diabetic nephropathy.
To ensure the relevance of our research, it was essential to transfer our findings in the zebrafish embryos to human patient data. Hence, kidney sections from nondiabetic, diabetic and polycystic kidney disease patients were tested for ELMO1 expression. We found that these patients, that we analysed, displayed an unaltered expression of ELMO1 within the renal parenchyma. This is in accordance with our research conducted on zebrafish embryos, but is in contrast to a recent published study which suggested that low ELMO1 expression prevents diabetic nephropathy development in mice 23 . It is common knowledge that no animal model completely adheres to the pathogenic mechanisms of human diabetic nephropathy. For example, mice with different background strains have variable susceptibility to developing diabetic nephropathy. Additionally, all criteria for diabetic nephropathy such as renal histolopathology and abnormal glomerular filtration rate are not completely relatable to mice models of diabetic nephropathy 24,25 . Thus, keeping this discrepancy of translatable murine data to human diabetic nephropathy in mind, a need for a new animal model is evident. Our data reinstates the relevance of zebrafish as a versatile animal model in basic diabetes research 26 . The zebrafish is known for its ease to work with, low maintenance costs and high throughput. Our study further displays that the zebrafish is a valuable model not only to test an implicated protein in diabetic nephropathy but also to understand the underlying mechanisms of the protein within the renal system; establish possible treatments, and translate these finding to the human setting.
In conclusion, upregulation of ELMO1 is advantageous for renal cells in zebrafish, especially under unfavourable hyperglycaemic conditions, which has been clearly established in the present study. Future prospects should thus focus on the induction of ELMO1 in diabetic conditions, in order to compensate for unfavorable effects of hyperglycaemia on the renal structure, contributing to diabetic nephropathy. Our findings clearly show that ELMO1 protects the renal system under hyperglycaemic conditions, via the reduction of apoptosis, within zebrafish embryos. The research is further translated in human patients suffering from long term hyperglycaemia leading to development of nephropathy.
Injections of morpholinos, mRNA and gRNA. Morpholinos were diluted to 4 or 6 μ g/μ l in 0.1 M KCl. mRNA was diluted to 300 pg/nl; ELMO1 gRNA and Cas9 mRNA were diluted to 100 pg/nl and 150 pg/nl respectively, in 0.1 M KCl. One nanoliter of morpholino or mRNA was injected into the embryos at the 1-cell stage.
Scientific RepoRts | 6:37172 | DOI: 10.1038/srep37172 A Roche light cycler ® 480 was used for the qPCR. Comparative ratio for ELMO1 expression was calculated with the 2 -ddCP 33 . The real-time quantitative PCR was repeated twice with a total of 6 replicates for each condition. The number of embryos used for the hyperglycaemic condition was 1017 and 895 for normoglycaemic conditions. The comparative ratio (Pfaffl formula) indicates no significant difference between ELMO1 expression levels.
Ultrafiltration assay. 72 hours post-fertilisation, 5 nl of Texas-Red ® tagged 70 kDa dextran (2 ng/ml in PBS) was injected into the heart of tg(fli1:EGFP) embryos. Sequential images of the living fish were taken 1, 24 and 48 hours post injection (hpi) using an inverted microscope (Leica DMI 6000B) with a camera (Leica DFC420 C) and the Leica LAS application suite 3.8 software. Maximum fluorescence intensity in the heart area was measured using NIH's ImageJ application (Supplementary Figure 3). The fluorescence values were compared in relative units of brightness for each fish.
Microscopy and analysis of pronephric alteration. For in vivo imaging Tg(wt1b:EGFP) embryos were embedded in 1% low melting point agarose (Promega), dissolved in E3 medium. Images were taken with a Leica DFC420 C camera, attached to a Leica MZ10 F modular stereo microscope. Quantification of the pronephron was done using the Leica LAS V4.8 software. Imaging was carried out for TUNEL stained Tg(wt1b:EGFP) via a DMIRE2 microscope with Leica TCS SP2 True Confocal Scanner and a DM6000 B microscope with Leica TCS SP5 DS scanner. Images were taken with 600 Hz, 1024 × 512 pixels and z-stacks were 1 μ m thick.
TUNEL assay. TUNEL assay was performed on 48 hpf Tg(wt1b:EGFP) embryos, using the ApopTag ® Red In Situ Apoptosis Detection Kit (Millipore). The embryos were deyolked and fixed over-night at 4 °C, in freshly prepared 4% PFA/PBS. After washing in PBST (1x PBS/0.1% Tween-20), sequential dehydration with increasing concentrations of methanol, was done to store them in 100% methanol overnight at − 20 °C. Embryos were rehydrated and treated with Proteinase K (10 μ g/ml, Roche) for 21 min. TUNEL protocol was hence followed as previously described 14 .
Immunohistochemistry (IHC) and fluoro-IHC for ELMO1. Human kidney tissue samples were provided by the tissue bank of the National Center for Tumor Diseases (NCT, Heidelberg, Germany) in accordance with the regulations of the tissue bank and the approval of the ethics committee of Heidelberg University.
For IHC analysis 5 non-diabetic (2/3 male/female patients aged 39-71 years), diabetic (2/3 male/female patients aged 34-76 years) and polycystic kidneys (4/1 male/female patients aged 54-67 years) were analysed. All kidney organs were removed during organ recession between 2007-2015 for all patients. All diabetic patients suffered from long term type 2 diabetes and developed nephropathy where kidney function was reportedly disrupted. Two diabetic patients showed other signs of diabetic complications such as retinopathy. All diabetic patients are assured to be insulin dependent. Sections of formalin-fixed, paraffin-embedded (FFPE) kidneys, provided by the NCT Tissue Bank, were cut at a thickness of 1 μ m (HM 340E Electronic Rotary Microtome, Thermo Scientific), dryed and stored at room temperature. IHC analysis was performed by using BenchMark ULTRA (Ventana Medical Systems). Antigens were retrieved with Protease 1 (12 min, room temperature, Ventana Medical Systems) and ELMO1 was detected with a polyclonal rabbit anti-ELMO1 antibody at 36 °C for 24 minutes, using the OptiView DAB IHC Detection Kit (Ventana Medical Systems). Podocin and CD34 were detected with their respective antibodies at 36 °C for 24 minutes, using the ultraView Universal Alkaline Phosphatase Red Detection Kit (Roche). NanoZoomer slide scanner (Hamamatsu) visualised tumor-free kidney parenchym and imaged with the Aperio ImageScope viewing software (Leica Biosystems, version 11.0.2.725).
For immunofluorescence staining, citrate buffer antigen revival method was used after de-paraffination was performed on kidney sections. After which, sections were blocked for an hour and incubated overnight at 4 °C with primary and then secondary antibodies. Fluorescent images were taken using a Zeiss Axio Imager D1 with a Axiocam mRm camera.
Quantification and statistics. TUNEL assay confocal images were processed using NIH's ImageJ application. Apoptotic cell number/area was determined by the analyze particles tool of ImageJ, within the entire pronephric structure. Alterations in the pronephric structure were measured and quantified by measuring size in μ m of glomerular width and neck. Immunofluorescence staining analysis for ELMO1 expression in human kidney sections was carried out by measuring integrated density of the fluorescence value representing ELMO1 protein expression, via NIH's ImageJ application. Background fluorescence was duly substracted and corresponding averages/ group were compared using a student t-test. The number of patients analysed per group were 5 and 8 glomeruli and tubules taken from 4 different stained sections, was utilized for ELMO1 quantification. Statistical significance between different groups was analysed using Student's t-test. Results are indicated as the mean value + /− SD. P-values < 0.05 were considered as significant *< 0.05, **< 0.01, and ***< 0.001.