BMSCs and miR‐124a ameliorated diabetic nephropathy via inhibiting notch signalling pathway

Abstract BMSCs are important in replacement therapy of diabetic nephropathy (DN). MiR‐124a exerts effect on the differentiation capability of pancreatic progenitor cells. The objective of this study was to explore the molecular mechanisms, the functions of miR‐124a and bone marrow mesenchymal stem cells (BMSCs) in the treatment of DN. Characterizations of BMSCs were identified using the inverted microscope and flow cytometer. The differentiations of BMSCs were analysed by immunofluorescence assay and DTZ staining. The expression levels of islet cell‐specific transcription factors, apoptosis‐related genes, podocytes‐related genes and Notch signalling components were detected using quantitative real‐time reverse transcription PCR (qRT‐PCR) and Western blot assays. The production of insulin secretion was detected by adopting radioimmunoassay. Cell proliferation and apoptosis abilities were detected by CCK‐8, flow cytometry and TUNEL assays. We found that BMSCs was induced into islet‐like cells and that miR‐124a could promote the BMSCs to differentiate into islet‐like cells. BMSCs in combination with miR‐124a regulated islet cell‐specific transcription factors, apoptosis‐related genes, podocytes‐related genes as well as the activity of Notch signalling pathway. However, BMSCs in combination with miR‐124a relieved renal lesion caused by DN and decreased podocyte apoptosis caused by HG. The protective effect of BMSCs in combination with miR‐124a was closely related to the inactivation of Notch signalling pathway. MSCs in combination with miR‐124a protected kidney tissue from impairment and inhibited nephrocyte apoptosis in DN.

differentiation 1 (NeuroD1) 11 . Thus, we speculated that BMSCs could be used to migrate DN through reducing hyperglycaemia and glycosurias. In this study, we aimed to further directionally induce BMSCs to differentiate into islet-like cell and to explore the role of BMSCs in DN.
MicroRNAs (miRNAs) are a class of endogenous small non-coding RNA molecules that are composed of 18-24 nucleotides. By binding to complementary sequences of target messenger RNAs (mRNAs) or non-coding RNAs [12][13][14][15][16] , miRNAs are important in post-transcriptional regulation 17,18 . MicroRNA124a (miR-124a) plays significant roles in pancreatic development, insulin secretion, b cell differentiation, the regulation of blood glucose and lipid metabolism [19][20][21] In addition, miR-124a can accurately regulate the development of embryonic pancreas by controlling pancreatic precursor cells counts and regulating the expression levels of pancreatic development-related transcription factors 20,22 . Studies have shown that the down-regulation of miR-124a could reduce the differentiation of pancreatic progenitor cell into b cells 23,24 . Moreover, miRNAs exert important effect on the differentiation of BMSCs into osteoblast, adipocyte and neurocyte 25,26 . However, BMSCs could be detected after one month of transplantation in a DN model. Researchers indicated that the proliferation potency of BMSCs was reduced after the transplantation 27 .
Therefore, we speculated that miR-124a may play an important role in BMSCs differentiation into islet-like cell.
Notch signalling pathways are a critical signalling pathway in cell differentiation [28][29][30][31] . Studies have demonstrated that the activation of Notch signalling pathway can promote the proliferation of pancreatic precursor cell and the development of exocrine gland 32,33 . Notch signalling pathways also produce critical effects on pancreatic development-related cells 34,35 . Moreover, the expression levels of Ngn3mediated downstream target genes, for example, NeuroD, Hes6-1 and MyT1, could be down-regulated by activating Notch signalling pathways 36 . However, the inactivation of Notch signalling pathways can promote the differentiation of endocrine cell in pancreas including a cell and b cell, and reduce insulin resistance 37,38 . Studies have reported that a cross-talk existed between miRNAs and Notch signalling pathways 39,40 . In our previous study, we found that Notch signalling pathways took part in BMSCs differentiation in vitro.
Therefore, we speculated that miR-124a may play an important role in the differentiation of BMSCs, in which also involves Notch signalling pathways. The combined use of BMSCs and miR-124a may improve kidney injury in DN rats. The study provides new methods for the prevention and treatment of DN. CO2 at 37°C. The culture medium was renewed every two days.
The growth and morphology of cells were determined using an inverted microscope (Carl Zeiss Microscopy). The cells were sub-cultured, and the 2-3 generations of cells were used for FACS analysis.
The cells were harvested by incubating with trypsin (cat. no. 0458) and then centrifuged (1000 g for 10 minutes at 4°C) to obtain cell pellet. The harvested cells were then washed with ice-cold PBS once and centrifuged (1500 g for 5 minutes at 4°C). The cell pellet was then re-suspended with 100 lL ice-cold PBS. Finally, cells were observed under an ordinary optical microscope.
Next, cells were labelled with antibodies against CD29 (Cat: 562154, BD Biosciences, USA), CD44 (Cat: 550974, BD Biosciences, USA) and CD34 (cat# 345801). Cells were then detected using flow cytometer 41 . To be more specific, all antibodies were incubated for 4 minutes in phosphate buffer saline (PBS) with 0.1% of Triton X-100. For each determination, at least 10 000 cells were analysed using a FACSCalibur cytometer (Becton Dickinson). CellQuest software (BD Biosciences) was used for result analysis. Meanwhile, immunofluorescence assay was performed. In brief, cells were grown in 6-well plates and fixed in 4% paraformaldehyde for 30 minutes.
After incubating with 3% Triton X-100 for 30 minutes, the slides were blocked with normal serum for 20 minutes and were then incubated with primary antibodies against CD29 (eBioscience, cat.

| BMSCs transfection and groups
MiR-124a mimics, inhibitors and negative control were purchased from GenePharma (Shanghai, China). Then, the mimics, inhibitors or negative control was, respectively, transfected into the cells using HiPerFect Transfection Reagent (Qiagen, Hilden, Germany) following the manufacturer's protocols. As previously reported 27 , BMSCs were harvested 48 hours after the transfection. The final concentration of mimics, inhibitors and negative control was 50 nmol/L. The mimics, inhibitors and negative control were repeatedly transfected every 3 days for long-term detection. BMSCs were further allocated into five groups: control group (Con; BMSCs with no treatment), induction group (Ind; BMSCs differentiation into islet-like cells), miR-124a negative control group (M/Ind; BMSCs were induced for 28 days so as to differentiate into islet-like cells after transfecting with miR-124a negative control), miR-124a mimics group (m/Ind; BMSCs were induced for 28 days so as to differentiated into islet-like cells after transfecting with miR-124a mimics), miR-124a inhibitors group (i/ Ind; BMSCs were induced for 28 days so as to differentiated into islet-like cells after transfecting with miR-124a inhibitors).

| CCK-8 assay
Cell viability was tested by the CCK8 assay in BMSCs and podocyte.
In brief, cells were harvested after being cultured for 28 days, and then were seeded at a density of 4 9 10 3 cells/well in 96-well plates and incubated for 0, 12, 24, and 48 hours. Subsequently, 20 lL CCK-8 was added to each well for other 1-hour incubation.
The optical density (OD) values were read at 570 nm using a microplate reader (Thermo, USA). All experimental concentrations were assessed in triplicate.

| Radio immunoassay
The cells in Con, Ind, M/Ind, m/Ind and i/Ind groups were harvested and were used for detecting insulin secretion with radio immunoassay 11 . In brief, cells in all groups were incubated respectively in the DMEM medium containing 2.8 mmol/L glucose, the DMEM medium contains 20 mmol/L glucose and 50 lmol/L IBMX for 2 hours at 37°C. Stimulation index (SI) was calculated and used for testing the sensibility of pancreatic b-cells to glucose stimulation.

| Animal model and groups
Male Sprague-Dawley (SD) rats aged 6-8 weeks of age were kept, according to the policy of the local ethics committee of laboratory animal centre, in the Animal Center of the First Affiliated Hospital of Medical College, Xi'an Jiaotong University. Rats were fed with HFD (19.7 kJ/g, 45% fat, Research Diets, Cyagen) for 8 weeks. After fasting for 12 hours, the rats were injected intraperitoneally with lowdose STZ (100 mg/kg) for 2 weeks to establish DN model. Urine samples over 24 hours were collected. Then venous blood was collected by cutting off rats' tails. Blood glucose, triglyceride, free fatty acid, urinary albumin excretion and creatinine were detected after 8 hours fasting. Rats with a blood glucose level over 16.7 mmol/L were considered as having diabetic, and were therefore accepted in the study 43 . Six groups of rats were used: (i) normal control group (Con, n = 6; the rats were normally fed for 14 weeks), (ii) BMSCs transplanted group (B/Con, n = 6; 3 9 10 6 BMSCs were injected into rats by tail vein after the rats have been normally fed for

| The histological examination of kidney tissue
Samples were isolated and embedded in paraffin to prepare 4-lm tissue slices. The slices were then investigated using haematoxylin and eosin (HE) staining and Periodic Acid-Schiff (PAS) staining. As previously described 44 , mesangial expansion index (MEI) was assigned and scored within four levels from 0 to 3. The index scores were defined as the follows: 0, normal glomeruli; 1, matrix expansion occurred in up to 50% of a glomerulus; 2, matrix expansion occurred 50% to 75% of a glomerulus; 3, matrix expansion occurred 75% to 100% of a glomerulus. Scores were assigned for at least 30 glomeruli from kidney slices for each sample, and the means were calculated.
Each slide was scored by a pathologist who was unaware of the experimental details.

| Cell culture and glucose treatment
A conditionally immortalized murine podocyte cell line was maintained as previously described 45  Podocytes were seeded in a 6-well plate at a density of 3 9 10 5 cells/mL with complete culture medium. After 24 hours, the podocytes were exposed to media containing 40 mmol/L of glucose for 24 hours so as to simulate DN environment in vitro (called HG group). Cells were then harvested for further experimental study.
Podocyte was divided into five groups: (i) blank control group (Con; podocyte with no treatment), (ii) HG group, (iii) BMSCs transfected with miR-124a negative control group (M/HG group: BMSCs were transfected with miR-124a negative control and the BMSCs with miR-124a negative control were incubated with podocytes for 24 hours. Then podocytes were exposed to 40 mmol/L glucose for 24 hours) (iv) BMSCs transfected with miR-124a mimics group (m/ HG; podocyte pre-treated with BMSCs transfected with miR-124a mimics for 24 hours, and then were exposed to 40 mmol/L glucose for 24 hours), (v) BMSCs transfected with miR-124a inhibitors group (i/HG; podocyte pre-treated with BMSCs transfected with miR-124a inhibitors, and then were exposed to the 40 mmol/L glucose for 24 hours).

| Detection of podocyte death
The podocytes of Con, HG, M/HG, m/HG and i/HG groups were trypsinized and fixed with 50% ethanol. The cells were washed three times with cold phosphate-buffered saline, and then stained with propidium iodide (PI, Invitrogen, cat. no. P3566) for 10 minutes. Cells were finally analysed by a flow cytometry.

| Quantitative real-time reverse transcription PCR (qRT-PCR)
Total RNA was extracted using RNA Extraction Kit (TIANGEN biochemical science technologies co., Ltd, Beijing, China) following the manufacturer's protocols. In kidney tissues, total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Next, 2 lg of RNA was used for complementary (cDNA) synthesis by using a first strand cDNA kit (TransGen Biotech, Beijing, China), according to the manufacturer's protocol. PCR amplification was carried out in ABI 7300 Thermocycler (Applied Biosystems, Foster City, CA, USA) performed with a SYBR Green PCR kit (Thermo). The PCR cycles were set at 95°C for 10 minutes, followed by 40 cycles at 95°C for 15 seconds, annealing/extension at 60°C for 45 seconds. The primers were designed by Shanghai Sangon Company (Shanghai, China).
U6 and b-actin were used as the internal control. qRT-PCR data were analysed using 2 ÀMMCt calculation 46 . The primer sequences were shown in Table 1.

| Proteins isolation and Western blot
Protein in kidney tissues was extracted using RIPA (Solarbio, Beijing,

China) and protein in cells was extracted using EpiQuik Whole Cell
Extraction Kit (AmyJet Scientific, Wuhan, Hubei, China). Western blot analysis was performed following standard procedures. Protein lysates were separated in 10% sodium dodecyl sulphate-polyacrylamide gels and electroblotted onto to a polyvinylidene fluoride membrane (Roche Diagnostics, Mannheim, Germany). Next, The primer sequences in qRT-PCR assay

Genes
Primers sequences

| Terminal deoxynucleotidyl transferasemediated dUTP nick-end labelling (TUNEL) assay
Cell apoptosis was detected with In Situ Cell Death Detection Kit (R&D Systems, USA), according to the manufacturer's protocol. Cells (1 9 10 5 cells/well) were seeded in 24-well plates and were treated.
Next, cells were first fixed in 4% paraformaldehyde at 4°C for 30 minutes and permeabilized in 0.1% Triton X-100, and treated with fluorescein-12-dUTP. Finally, the fluorescence was detected using fluorescence microscopy (Zeiss Axiovert 100 M, Carl Zeiss, Germany).

| Statistical analysis
Data were expressed as mean AE standard deviation (SD). Differences between two groups were determined by Student's t test for unpaired variables. However, differences among three or four groups in vitro study were analysed by one-way ANOVA and subsequent Tukey's test. Probability (P) values <.05 was considered as statistically significant.

| Characterization of BMSCs
The rat BMSCs presented fibroblast-like shape. The size of cell increased, and vacuoles and granules appeared in the cells (Figure 1A). Results from immunofluorescence assay showed that the positive expressions of CD29 and CD44, and negative expression of CD34 in BMSCs ( Figure 1B). Furthermore, results from flow cytometry showed that the proportions of CD29, CD44 and CD34 were 98.05%, 95.16% and 1.94% in BMSCs, respectively ( Figure 1C).

| Differentiation of BMSCs
Our results showed that nestin was highly expressed in BMSCs after 24-hour induction. Insulin was positive in BMSCs after being induced for 7 days. The expression of glucagon and somatostatin were positive in BMSCs after being induced for 14 days (Figure 2A). In addition, the results of DTZ staining showed that BMSCs turned brown red after being induced for 14 days, suggesting that the cytoplasm of BMSCs contains zinc ion. The above results proved that BMSCs was induced into islet-like cells (Figure 2B). The expression levels of Pdx-1, Pax-6, Insulin-1, Ngn3 and GK in BMSCs were then determined by qRT-PCR and Western blot assays. The results indicated that accompanied with the differentiation of BMSCs, Pdx-1, Pax-6, Insulin-1 and GK expressions were gradually increased, and Ngn3 expression was gradually decreased (P < .05, P < .01, Figure 2C and D). Furthermore, the production of insulin in BMSCs was detected by performing radioimmunoassay.
The results indicated that the production of insulin was significantly increased with the differentiation of BMSCs progressed (P < .001, Figure 2E).

| MiR-124a promoted insulin secretion and the expression of islet cell-specific transcription factors in BMSCs
To investigate whether miR-124a play crucial roles in BMSCs or not, BMSCs were treated with PBS (Con) and 50 mg/L TSPG (Ind) for transfected with miR-124a mimics was higher than that in other groups ( Figure 3A). Moreover, the expression level of miR-124a went up significantly after BMSCs were transfected with miR-124a mimics (P < .05, P < .01, Figure 3B). CCK-8 results also showed that the cell viability increased significantly after BMSCs were transfected with miR-124a mimics, however, it decreased significantly after the BMSCs were transfected with miR-124a inhibitors for 12, 24 and 48 hours (P < .05, P < .01, Figure 3C). Furthermore, qRT-PCR and Western blot assays were used to detect the expression levels of more obvious in m/Ind group than that in M/Ind group. Moreover, Ngn3 expression was increased in i/Ind group than that in M/Ind group (P < .05, P < .01, Figure 3D and E).

| BMSCs in combination with miR-124a enhanced podocytes cell viability treated with HG
To further examine the biological significance of BMSCs in combination with miR-124a on podocytes cell proliferation treated with high-glucose (HG), immunofluorescence assay was first used to identify the characteristics of podocytes cells. The results indicated that synaptopodin protein highly expressed in podocytes ( Figure 4A).
Then, Podocytes were treated with PBS (Control) and glucose at Subsequently, podocytes were treated with PBS (Con) and 40 mmol/L HG (HG) for 24 hours and were then co-cultured with BMSCs, which were respectively transfected with mock (M/HG), miR-124a mimics (m/HG) and miR-124a inhibitors (i/HG) for 48 hours. CCK-8 results indicated that the cell viability decreased in podocytes treated with high-glucose. After co-culturing with BMSC that was transfected with miR-124a mimics, the cell viability increased significantly compared with M/HG groups. After co-culturing with BMSC that was transfected with miR-124a inhibitors, the cell viability decreased further, compared with M/HG groups (P < .05, P < .01, Figure 4C). In addition, we also found that HG significantly inhibited the expression levels of nephrin, podocin and CD2AP in podocytes, and that the pre-treatment of BMSC transfection with miR-124a mimics significantly increased nephrin, podocin and CD2AP expressions compared with M/HG groups. However, the pre-treatment of BMSC transfection with miR-124a inhibitors significantly decreased nephrin, podocin and CD2AP expressions in comparison with M/HG groups (P < .05, P < .01, Figure 4D and E).

| BMSCs in combination with miR-124a inhibited podocytes cell apoptosis treated with HG
Furthermore, we explored the effects of BMSCs in combination with miR-124a on the apoptosis of podocytes cells treated with HG. The results showed that HG significantly promoted apoptosis of podocytes cells, and that the pre-treatment of BMSC transfected with miR-124a mimics sharply inhibited apoptosis compared with M/HG groups. The pre-treatment of BMSC transfected with miR-124a inhibitors was observed to significantly enhance apoptosis in comparison with M/HG groups (P < .05, P < .01, Figure 5A and B). In addition, we also found that HG noticeably up-regulated caspase-3 and Bax expressions, and down-regulated Bcl-2 expression in podocytes, and that the pre-treatment of BMSC transfection with miR-124a mimics  Figure 5C and D).

| BMSCs combined with miR-124a regulated the expressions of Notch1, NICD, Hes1 and Delta
We further demonstrated the effect of BMSCs combined with miR-124a on Notch signalling pathway. Our results showed that HG sig-  Figure 6A and B).

| BMSCs combined with miR-124a reduced kidney injury in DN rats
We found that the bodyweights of DN rats were increased significantly in DN model group, compared to those of Con and B/Con groups. However, bodyweights of rats treated with BMSCs transfection with miR-124a mimics were decreased compared to DN group.
The 24-hours urinary albumin excretion (UAE) was increased significantly in DN groups, indicating that the renal function was recovered by the treatment of BMSCs in combination with miR-124a, although there was no significant difference in terms of kidney index (UMA/ Ucr) 47 between m/DN and the model group (P > .05, Figure 7A).
Compared to control and B/Con groups, blood glucose level higher than 16.7 mmol/L were observed in DN and M/DN groups.
Whereas blood glucose level was decreased in m/DN group compared to DN and M/DN groups (P > .05, Figure 7B). The expression of miR124a was shown in Figure 7C. According to our results, DN models were successfully established by HFD/STZ treatment, and the model could be used in further experimental study. Our data showed that the overexpression of miR-124a could decrease the weight of DN rats and blood glucose level.

| BMSCs in combination with miR-124a protect kidney tissue from damage
To confirm the effect of BMSCs transfection with miR-124a mimics/ inhibitors against DN rats, SD rats were divided into control group  Figure 8A and B). After treating with BMSCs in DN rats, the extent of pathological injury of kidney injury was markedly alleviated. The effect of BMSCs transfection with miR-124a mimics was more productive than that with miR-124a inhibitors/negative control. In addition, we found that the expression levels of nephrin, podocin and CD2AP were lower in DN group than those in Con and B/Con groups, and that nephrin, podocin and CD2AP expressions increased significantly in m/DN group in comparison with DN group (P < .05, P < .01, Figure 8C and D). Furthermore, we also detected the expression of fibrotic related proteins 48 .
The result showed that compared to DN group, the expression of fibronectin (FN), TGF-b1, Col-I, and Col-III was depressed by the treatment of BMSCs in combination with miR-124a. By contrast, the co-treatment of BMSCs and miR-124a inhibitors enhanced the expression of FN, TGF-b1, Col-I, and Col-III compared to DN group ( Figure 9A and B).

| BMSCs in combination with miR-124a inhibited nephrocyte apoptosis
The method of TUNEL was used to detect apoptosis in SD rats that were divided into Con, B/Con, DN, M/DN, m/DN and i/DN groups.
We found that the apoptosis rate was higher in DN group than that in Con and B/Con groups ( Figure 10A). Cell apoptosis was increased significantly in DN group. But the apoptosis was decreased in m/DN group compared to DN group. The apoptosis of i/DN group was increased compared to that of the DN group ( Figure 10A). In addition, apoptosis-related genes including caspase-3, Bax and Bcl-2 were detected by qRT-PCR and Western blot assays with an aim to further explore the mechanism. Our results showed that the expression of pro-apoptosis factors (caspase-3 and Bax) was significantly increased, whereas the expression of anti-apoptosis factor (Bcl-2) was decreased noticeably in DN group. However, the presence of BMSCs transfected with miR-124a mimics decreased the expression the effects that opposite to BMSCs transfected with miR-124a mimics (P < .05, P < .01, Figure 10B and C).

| DISCUSSION
Diabetes mellitus (DM) has a high morbidity rate and is usually accompanied with complications that seriously threats human health 49,50 . The main treatments for DM are oral hypoglycaemic medications and/or insulin. However, some severe medication side effects emerge from the treatment 51 . In recent years, many studies have proved that the application of hematopoietic stem cell or islet transplantation was effective method in treating type 1 diabetes 52,53 .
Although this treatment could produce certain effects, the long-term use of immunosuppressive agents could produce side effects for the patients of islet transplant 54,55 . Thus, it is necessary to seek for a method that is safe and long-acting for treating diabetes and its complications. In recipient mice, bone marrow transplantation can generate islet cells, which can produce genetic markers and insulin of b-cells 63 74 . This pathway consists of four receptors, which are Notch1-Notch4 and five ligands: Jagged1, Jagged2, Delta-like (Dll) 1, Dll3 and Dll4 75 . The binding of ligand and Notch receptor induces a conformational change, which allows the c-secretase-mediated protease to release the Notch intracellular domain (NICD) 76 . NICD travels into the nucleus in which it activates the transcription of downstream genes such as Hes1 and Hey1 genes, and affects cellular differentiation, proliferation and apoptosis 77 . In our study, we demonstrated that HG significantly up-regulated the expressions of Notch1, NICD, Hes1 and Delta in podocytes, and that BMSCs in combination with miR-124a inhibited the expressions of Notch1, NICD, Hes1 and Delta. Thus, those data suggested that BMSCs and miR-124a inhibited Notch signalling pathway.
Nephrin is a transmembrane protein in the podocyte that belongs to a member of the immunoglobulin superfamily 78  However, this mechanism would be disrupted under HG conditions.
Our results showed that the BMSCs in combination with miR-124a inhibited the expressions of FN, TGF-b1, Col-I and Col-III, suggesting that BMSCs in combination with miR-124a depressed the progression of renal fibrosis. With the fibrosis of the kidney, renal failure is also related to the progressive deletion of renal cells 82 . We found that BMSCs in combination with miR-124a inhibited nephrocyte apoptosis by regulating the expressions of Bcl-2, caspase-3 and Bax.
Collectively, miR124a helped BMSCs differentiate into islet-like cells and the insulin secretion of BMSCs. Therefore, the depressed renal injury in DN may be explained by that the insulin conferred by BMSCs in the presence of miR124a. In addition, as a safe and effective delivery vehicle in exosomal trafficking, MSCs has been used to deliver miRNA to targeted cells to treat diabetes 83 . Thus, it was also possible that the presence of MSC elevated the transfection efficiency of miR124a and then enhanced the effect of miR124a.
However, one limitation of this study was that there was discrepancy among changes in renal function and in molecular biology.
This may due to the fact that the final organ function may affect by not only the molecules we detected, but also some other unknown signals or the cell context. In addition, the precise regulatory mechanisms via which MSCs combined with miR-124a exerted was not clear in this study. Thus, it will be intriguing to illustrate the regulatory signal pathways involved in this study model, and it will enrich our knowledge on the molecular progression of DN.
In summary, our studies have successfully identified and isolated BMSCs. We proved that miR-124a promoted insulin secretion, proliferation and affected the expressions of Pdx-1, Pax-6, Insulin-1, Ngn3 and GK in BMSCs. Our results also indicated that MSCs in combination with miR-124a enhanced proliferation and inhibited apoptosis of podocytes cells mediated by HG. In addition, we observed that BMSCs in combination with miR-124a inhibited Notch signalling pathway in podocytes cells. We have demonstrated that BMSCs in combination with miR-124a protected kidney tissue from impairment and inhibited nephrocyte apoptosis in vivo.

CONF LICT OF I NTEREST
The authors declare no conflict of interest.