FTO‐mediated m6A modification of serum amyloid A2 mRNA promotes podocyte injury and inflammation by activating the NF‐κB signaling pathway

Diabetic kidney disease (DKD) is one of the severe complications of diabetes mellitus, yet there is no effective treatment. Exploring the development of DKD is essential to treatment. Podocyte injury and inflammation are closely related to the development of DKD. However, the mechanism of podocyte injury and progression in DKD remains largely unclear. Here, we observed that FTO expression was significantly upregulated in high glucose‐induced podocytes and that overexpression of FTO promoted podocyte injury and inflammation. By performing RNA‐seq and MeRIP‐seq with control podocytes and high glucose‐induced podocytes with or without FTO knockdown, we revealed that serum amyloid A2 (SAA2) is a target of FTO‐mediated m6A modification. Knockdown of FTO markedly increased SAA2 mRNA m6A modification and decreased SAA2 mRNA expression. Mechanistically, we demonstrated that SAA2 might participate in podocyte injury and inflammation through activation of the NF‐κB signaling pathway. Furthermore, by generating podocyte‐specific adeno‐associated virus 9 (AAV9) to knockdown SAA2 in mice, we discovered that the depletion of SAA2 significantly restored podocyte injury and inflammation. Together, our results suggested that upregulation of SAA2 promoted podocyte injury through m6A‐dependent regulation, thus suggesting that SAA2 may be a therapeutic target for diabetic kidney disease.


| INTRODUCTION
Diabetic kidney disease (DKD), one of the severe complications of diabetes mellitus, is the leading cause of end-stage renal disease (ESRD). 1 Although new drugs, such as glucagon-like peptide-2 antagonists and sodiumdependent glucose transporter 2 inhibition, have emerged in recent years for the treatment of DKD, 2 they are still insufficient.Hence, the search for new methods to treat DKD is still a challenge we need to solve.
Podocytes are terminally differentiated epithelial cells, which together with endothelial cells and the basement membrane, form the glomerular filtration barrier. 3,4A key event in the pathogenesis of DKD is podocyte injury, leading to the development of proteinuria and eventually glomerulosclerosis. 5,6Accumulating studies have observed that podocyte injury is manifested by altered expression levels of related markers, such as a reduction in ZO-1 and overexpression of Desmin. 7,8Therefore, investigating the mechanism of podocyte injury will provide more reference for the clinical treatment of DKD.
N6-methyladenosine (m6A) RNA methylation, defined as RNA methylation on the sixth N atom of adenylate, is the most prevailing modification in eukaryotic mRNA. 96A modifications on mRNA are dynamic and reversible.It is mainly mediated by m6A methyltransferases, which are referred to as "writers," including METTL14 and METTL3, and can be removed by m6A demethylases, which are known as "erasers," consisting of ALKBH5 and fat-mass and obesity-associated protein (FTO) 10,11 Furthermore, specific RNA-binding proteins, which are referred to as "readers," such as YTHDF1 and YTHDF2, can recognize the methyl groups on mRNAs to influence RNA function. 10m6A modifications are mainly involved in diverse biological regulatory processes, including mRNA stability, splicing, transport, positioning, translation, and protein-RNA interactions. 12As a demethylating enzyme, FTO usually requires demethylation of specific mRNAs and regulates mRNA metabolism.It has been reported that FTO-mediated m6A modification of mRNA regulates the progression of diabetic kidney diseases. 13Additionally, a previous study by our group found that FTO was overexpressed in high glucose-induced podocytes. 8However, its function and detailed molecular mechanisms in DKD are largely undefined.
In our previous MeRIP-seq data, we observed that SAA2 was significantly upregulated in high glucoseinduced podocytes.In addition, the m6A level of SAA2 was decreased in high glucose-induced podocytes, while silencing FTO reversed this phenomenon, suggesting that FTO-mediated m6A modification of SAA2 might play a role in high glucose-induced podocyte injury.However, the exact mechanism is not yet known.SAA2, one of the subtypes of SAA, is a potent proinflammatory protein in humans. 14Studies have reported that SAA is increased in human DKD and corresponding mouse models. 15oreover, the effect of SAA exposure could enhance the inflammatory response in podocytes. 14However, no studies have yet reported the link between FTO and SAA2 and the mechanisms by which they play a role in podocyte injury and inflammation.
In the present study, we found that FTO caused podocyte injury in diabetic kidney disease by regulating the expression level of SAA2.We revealed a mechanism of diabetic kidney disease pathogenesis, which provides a new target for future prevention and treatment of diabetic kidney disease.

| Cell culture
Mouse primary podocytes were extracted as described above. 16The immortalized human podocytes (HPC) gifted by Peter Mundel, Albert Einstein College of Medicine, Bronx, New York, were grown in a complete medium composed of RPMI 1640 medium (F211130, Gibco, USA), 10% fetal bovine serum (A5669701, Gibco, USA), and 1% penicillin streptomycin.HPCs were incubated in a humid environment with 5% CO 2 at 37°C.After cell attachment, pedunculated cells stimulated with low glucose (LG group) formulated with 5.5 mM d-glucose (G8150, Solarbio, China), high glucose (HG) formulated with 30 mM d-glucose, and hyperosmolality (HO) formulated with 5.5 mM d-glucose plus 24.5 mM mannitol (SM8120, Solarbio, China) were used for other experiments in this study.

| Cell transfection and infection
Transient transfection of plasmids and siRNAs was performed using Lipofectamine 3000 reagent (L3000150, Invitrogen, USA).Knockdown and overexpression of FTO with lentiviral particles were performed at a multiplicity of infection (MOI) of 50.Polybrene (10 mg/mL) was added to stabilize the infection, and puromycin was added to screen for stable clones.Design and construction of lentiviruses for FTO from Genomeditech (Shanghai, China), including: overexpression vectors, empty vectors, shRNA, and interfering shRNA.Overexpression plasmids, empty vectors, siRNAs, and negative controls for SAA2 were also provided by Genomeditech (Shanghai, China).The corresponding shRNA and siRNA sequences are as follows: FTO shRNA: GGTGG CAG TGT ACA GTT ATAG Scramble shRNA: TTCTC CGA ACG TGT CACGT SAA2 siRNA: GCCGA UCA GGC UGC CAA UAAA NC siRNA: UUCUC CGA ACG UGU CACGU

| Quantitative real-time PCR (qRT-PCR) analysis
TRIzol reagent (Invitrogen, CA, USA) was used for the extraction of total RNA from cells and kidney tissue.The concentration and purity of the RNA were measured by a NanoDrop 2000 spectrophotometer (Thermo Fisher Science, USA).cDNA was synthesized using the PrimeScrip™ RT kit with gDNA Eraser (AG11711, Kogen, Japan) with equal amounts of RNA.The above cDNAs were analyzed for RNA expression levels using SYBR® PreMix Ex Taq (RR820A, Takara, China) on a LightCycler® 480II (Roche, USA).The relative expression levels of the target genes were analyzed by the 2 −ΔΔCT method using βactin as an internal reference gene.The primer sequences of the target genes, which were designed and synthesized by Biosune (Shanghai, China) are shown in Table 1.

| Immunofluorescence assay
Podocytes plated on six-well plates were fixed with 4% paraformaldehyde after different stimuli.Cells were permeabilized with 0.1% Triton X-100 before being blocked with 5% bovine serum albumin for 30 min at room temperature.The cells were then incubated overnight at 4°C with primary antibodies against synaptopodin, Desmin, and ZO-1.The following day, the cells were incubated with goat anti-rabbit IgG H&L (Alexa Fluor® 488) (1:200, ab150077, Abcam, USA) and goat anti-rabbit IgG H&L (Alexa Fluor® 594) (1:200, ab150080, Abcam, USA) at 37°C for 1-2 h away from light.After incubation with DAPI for 10 min, the cells were observed and recorded with a fluorescence microscope (Leica, Germany).Tissue immunofluorescence: human kidney tissues were obtained from the Department of Nephrology, Shandong Provincial Hospital.Mouse kidneys were fixed with 10% formalin, embedded in paraffin and cut into 4-μm sagittal sections.Paraffin sections were dehydrated, dewaxed, and then repaired by Tris under high pressure.This was followed by the same steps as for cellular immunofluorescence.

| Dual-luciferase reporter gene assay
Podocytes were evenly seeded into 6-well plates and incubated overnight.The cells were then transfected using Lipofectamine 3000 (Invitrogen) with 1 μg pNF-κBluc plasmids (Beyotime) and 0.02 μg pRL-TK-vector plasmids (Promega) as an internal control.At the same time, four of the wells were individually transfected with overexpression plasmids, empty vectors, siRNAs, and negative

| mRNA stability assay
Podocytes were seeded in two 6-well plates, one of which served as the control, and the other was transfected with lentivirus to knock down FTO as described above.After 48 h, podocytes were treated with actinomycin D (10 mg/ mL) for 24, 9, 6, 3, and 0 h.Then, total RNA was isolated using TRIzol.Real-time PCR was performed to detect the expression level of SAA2 after reverse transcription.

| Histology and immunohistochemistry
Human and mouse tissue sections were prepared as described above.Periodic acid Schiff (PAS) staining (G1281, Solarbio, Beijing, China) and Masson trichrome staining (G1340, Solarbio, Beijing, China) were performed according to the instructions for the reagents.
Immunohistochemistry was also performed according to standard procedures.Similarly, paraffin sections were first dewaxed and hydrated in xylene and gradient ethanol before antigen repair.The sections were then inactivated with endogenous peroxidase and biotin, blocked with serum and incubated overnight with primary antibody.DAB chromatography was performed after incubation with the secondary antibody the following day.The sections were then stained with hematoxylin and divided into ethanol with hydrochloric acid.The slices were then sealed after ethanol dehydration.The above-stained sections were observed under a fluorescence microscope (Leica, Germany) and photographed.Primary antibodies included anti-Wilms Tumor (ab89901, Abcam, USA) and anti-F4/80 (28463-1-AP, Proteintech, China).

| Dot blot analysis of N6-methyladenosine (m6A)-RNA modification
Total RNA was isolated as described above.mRNA samples were uniformly diluted to 500 ng/μL.Then, the samples were denatured at 95°C within 3 min.One microgram of RNA from each group was loaded onto nylon membranes (Beyotime, China).The membrane was UV crosslinked for 45 min and washed with TBST.After being blocked with 5% nonfat milk, the membrane was incubated with specific m6A antibody (dilution 1:3000, 68055-1-Ig, Proteintech) overnight at 4°C.Dot blots were incubated with HRP-conjugated anti-mouse immunoglobulin G (IgG) for 1 h before visualization by an imaging system (GE, USA).Then, the membrane was stained with 0.02% methylene blue (G1300, Solarbio, China), followed by the imaging to indicate the total content of input RNA.

| Methylated RNA immunoprecipitation (MeRIP) assay
Total RNA was extracted with a TRIzol reagent (Takara, Japan).The Magna MERIP™ m6A kit (17-10499, Millipore, USA) was used to immunoprecipitate chemically fragmented RNA (~100 nucleotides).Subsequently, the concentration and purity of the immunoprecipitated RNA were measured by a NanoDrop 2000 spectrophotometer (Thermo Fisher Science, USA).Then, the immunoprecipitated RNA was reverse transcribed and applied for real-time PCR as described above.

| Human kidney specimens
Tissues, including kidney biopsy tissues from DKD patients and normal kidney tissues from nondiabetic renal cancer patients undergoing surgical resection, were obtained from the Department of Nephrology, Shandong Provincial Hospital affiliated with Shandong First Medical University.Our studies were carried out after the review and approval of the Experimental Animal Ethics Committee of Shandong Provincial Hospital affiliated with Shandong First Medical University, in agreement with the guidelines set forth by the Declaration of Helsinki.All participants signed informed consent forms prior to using the tissues for scientific research.

| Animals
C57BL/6 mice were provided by the Experimental Animal Center of Shandong First Medical University.All animal studies were carried out after the review and approval of the Experimental Animal Ethics Committee of Shandong Provincial Hospital affiliated to Shandong First Medical University.Based on the podocyte specificity of NPHS1, we constructed a recombinant adeno-associated virus (AAV9-SAA2-shRNA) and a recombinant adenoassociated virus-negative control (AAV9-NC) targeting podocytes by genetic engineering techniques.The vector structure was constructed as pAV-NPHS1-GFR-miR30-shRNA. The construction of AAV9-GFP has been widely used. 17The shRNA target sequences were as follows: SAA2-shRNA: A AGA GA G C TT TCA GG A A TT CTT TA G T GA AGC CA C A GA TGT AA A G AA TTC CTGAAAGCTCTCTC.
We then performed PCR for genotyping with the following primers: Mus-SAA2-71F CTGAC ATG AAG GAA GCTGGC Mus-SAA2-71R GGCAG CAT CAT AGT TCCCCC 4-week-old C57BL/6 mice were randomly divided into 4 groups (WT-CTL, n = 10; WT-STZ, n = 15; SAA2 NC-Pod -STZ, n = 6; SAA2 KO-Pod -STZ, n = 6).Mice in the SAA2 KO-Pod -STZ and SAA2 NC-Pod -STZ groups were injected with AAV9-SAA2-shRNA and AAV9-NC through the tail vein at 4 weeks of age with an injection volume of 1*10 12 vg/ mouse, while mice in the WT-CTL and WT-STZ groups were injected with PBS.After 4 weeks, the WT-STZ, SAA2 NC-Pod -STZ and SAA2 KO-Pod -STZ groups were given streptozotocin (STZ, Sigma-Aldrich, St Louis, MO, USA) dissolved in citrate buffer (50 mg/kg) by intraperitoneal injection for 5 consecutive days and a high-fat diet, while the WT-CTL group was given citrate buffer alone.Mice were considered diabetic when the blood glucose level was higher than 300 mg/dL A total of 16 days after the last STZ injection.Urine protein, urea nitrogen, and creatinine were measured in mice as previously described. 18All mice were euthanized at 20 weeks of age.Kidney tissue from mice was used in the follow-up study.

| Statistical analysis
GraphPad Prism 8.0.1 software was used for all statistical analyses.Analysis of variance was performed by t-tests and one-way ANOVA.Spearman correlation analysis was performed to assess the correlation between SAA2 expression and clinical indicators in the mouse kidney.p < .05 was considered statistically significant.The above experiments were repeated three times.

| High glucose-induced podocyte injury and inflammation
We used podocytes treated with high glucose to mimic the human DKD microenvironment.Western blot and realtime PCR showed that the expression level of Desmin, a marker of podocyte damage, was increased in high glucose-induced podocytes (Figure 1A,E), while ZO-1, which maintains the integrity of podocytes, was decreased (Figure 1B,E), consistent with the results of immunofluorescence staining (Figure 1F,G).Considering that exposure of podocytes to a high-glucose environment induced inflammation, we further detected the expression of inflammatory factors in high-glucose-induced podocytes.We observed that TNF-α, IL-6, and MCP-1 were highly expressed under high glucose conditions (Figure 1C-E).In conclusion, these results suggest that high glucose levels cause podocyte injury and inflammation.

| FTO is upregulated in high glucose-treated podocytes and associated with podocyte injury
We used real-time PCR to evaluate the level of FTO in podocytes treated with high glucose, and the results showed that the expression of FTO was upregulated (Figure 2A).Additionally, Western blot results found increased protein expression levels of FTO in high glucose-induced podocytes (Figures 2B and S1A).This is consistent with our previous study. 14Next, to explore the relationship between FTO expression and podocyte injury, we constructed a lentiviral vector for stable high expression and knockdown of FTO in podocytes, and the transfection efficiency is shown in Figure S1B-E.Western blot and real-time PCR showed that overexpression of FTO could increase the expression of Desmin and reduce the level of ZO-1 (Figure 2C,E).The immunofluorescence staining was consistent with the above results (Figure 2I,J).Moreover, silencing FTO reversed the upregulation of Desmin and downregulation of ZO-1 induced by high glucose (Figure 2F,H-J); thus, knocking down FTO could alleviate high glucose-induced podocyte injury.Interestingly, we observed that the inflammatory factors TNF-α, IL-6, and MCP-1 were also upregulated in low glucose-induced and FTO overexpression cotreated podocytes (Figures 2D,E and S1G).In contrast, the expression level of inflammatory factors decreased when FTO was knocked down (Figures 2G,H and S1H).These results suggest that FTO is involved in podocyte injury and inflammation, but the exact mechanism is still unclear and needs to be further explored.

| FTO regulated SAA2 in a m6A-dependent manner in podocytes
To explore the downstream targets regulated by FTO through the m6A mechanism, we previously performed RNA-seq and MeRIP-seq with control podocytes and high glucose-induced podocytes with or without FTO knockdown.The results showed that the m6A level of SAA2 was reduced in high glucose-induced podocytes and was significantly increased after silencing FTO expression (Figure 3A).RNA-seq showed that SAA2 was upregulated in the HG-input group compared with the LG-input group (Figure 3B).This suggests that there may be a potential link between FTO and SAA2.To further determine the effect of FTO on the upregulation of SAA2, we again used previous podocytes that used lentivirus to overexpress or knockdown FTO.Compared to the LG group, real-time PCR and Western blot showed that the expression of SAA2 was significantly reduced after silencing FTO expression (Figure 3C,D), while overexpression of FTO reversed the above results (Figure 3E,H).To evaluate whether FTO has an effect on SAA2 mRNA stability, we treated FTO stable knockdown and control podocytes with actinomycin D (ActD, 5 μg/mL), which is widely used for the experimental blockage of the transcriptional process, and we found that silencing FTO caused increased degradation of SAA2 mRNA (Figure 3I).Subsequently, we performed dot blot analysis and showed a significant decrease in m6A-RNA methylation in the podocytes treated with high glucose compared to the LG group (Figures 3F and S1F).The results of MeRIP-PCR revealed that the m6A modification level of SAA2 was reduced in high glucose-induced podocytes (Figure 3G), consistent with our sequencing results.Meanwhile, the m6A abundance of SAA2 was downregulated upon FTO overexpression, and was increased after FTO knockdown (Figure 3G).Taken together, FTO-mediated m6A demethylation might be responsible for the upregulation of SAA2 by influencing SAA2 mRNA stability, and SAA2 might be a potential target of FTO.

| SAA2 is involved in podocyte damage
To investigate whether SAA2 is involved in high glucoseinduced podocyte injury, we examined the mRNA and protein levels of SAA2 in high glucose-treated podocytes and showed that high glucose stimulation significantly promoted the expression of SAA2 (Figure 4A,B).We constructed SAA2 expression vectors and siRNAs to enhance and downregulate the expression of SAA2 in podocytes for further studies on the regulatory role of SAA2 (Figure S2A-D).First, qRT-PCR showed that overexpression of SAA2 significantly upregulated Desmin and downregulated ZO-1 (Figure 4D), while silencing SAA2 reversed the upregulation of Desmin and downregulation of ZO-1 induced by high glucose stimulation (Figure 4C).Western blot assays showed changes in protein levels consistent with the levels of mRNAs (Figure 4E,F).Additionally, immunofluorescence staining showed that SAA2 could inhibit the expression of Desmin and promote the expression of ZO-1 (Figure 4G).Last, immunofluorescence co-staining of SAA2 and synaptopodin in kidney tissue sections from patients with diabetic kidney disease showed that SAA2 was localized in podocytes and was highly expressed in diabetic kidney disease (Figure 4H).These results suggest that SAA2 participates in podocyte injury and is thus responsible for the progression of diabetic kidney disease.

| SAA2 contributes to FTO-mediated podocyte injury by regulating the NF-κB pathway
SAA2 is a proinflammatory protein and is involved in the inflammatory response.Meanwhile, many studies have shown that the NF-κB signaling pathway is closely associated with the development of inflammation. 15Therefore, we hypothesize that SAA2 acts through this pathway to promote podocyte injury and inflammation.We explored the impact of SAA2 overexpression and knockdown on the NF-κB pathway in podocytes.The results showed that overexpression of SAA2 promoted the phosphorylation of P65 (Figure 5A) while knockdown of SAA2 inhibited the phosphorylation of P65 (Figure 5B).In addition, we found that overexpression of SAA2 enhanced NF-κB activity, whereas knockdown of SAA2 decreased its activity, as detected by dual-luciferase reporter assays (Figure 5C,D).The effect of SAA2 on NF-κBdependent gene expression was further examined by Western blot and qRT-PCR.
The increase in the inflammatory factors TNF-α, IL-6, and MCP-1 caused by high glucose stimulation was reversed by knockdown of SAA2 (Figures S2H and 5E,G), while overexpression of SAA2 promoted the expression of inflammatory factors (Figures S2I and 5F,G).Next, we further confirmed whether FTO mediated podocyte injury through SAA2.Through functional reversion experiments, we found that both the upregulation of pP65 and Desmin and the downregulation of ZO-1 caused by the overexpression of FTO were reversed by the knockdown of SAA2 (Figure 5H-L).Meanwhile, silencing SAA2 was found to significantly inhibit FTO-mediated inflammatory damage in podocytes (Figure 5I-M).Taken together, SAA2, as the downstream target of FTO, participates in podocyte injury by activating the NF-κB pathway.

| SAA2 depletion alleviates kidney damage in STZ mice
To uncover the roles of SAA2 in vivo in diabetic kidney disease, we generated mice with podocyte-specific ablation of SAA2 (SAA2 KO-Pod mice) by intravenously injecting adenoviruses expressing shRNA targeting SAA2(Ad-shSAA2) into the mouse tail vein.qRT-PCR demonstrated effective deletion of SAA2 in primary podocytes isolated from SAA2 KO-Pod mice (Figure S2E).In addition, we validated the podocyte-specific knockdown of SAA2 again by quantifying SAA2 expression in endothelial and renal tubular cells (Figure S2F,G).Consistent with the above study, we found that SAA2 expression was significantly higher in the kidneys of STZ mice than in wild-type mice (Figure 6A).Next, we conducted a series of studies on the basis of STZ mice.First, clinical indicators related to renal function in mice were detected, and the results showed that the levels of albuminuria and urea nitrogen in STZ mice were significantly elevated, while the levels in STZ mice with SAA2KO-Pod were significantly lower than those in STZ mice (Figure 6E,F).Serum creatinine also showed the above trend (Figure 6G).In addition, we assessed the correlation between clinical indicators of renal function and SAA2 expression in the kidneys of STZ mice.Spearman correlation analysis showed that BUN (r = .558,p = .031),serum creatinine (r = .525,p = .044),and ACR (r = .754p = .001)were correlated with the expression of SAA2 (Figure 6B-D).Second, histopathological examination of the glomeruli revealed similar results.Periodic acid-Schiff (PAS) staining showed significant endothelial cell dilation, mesangial proliferation, matrix augmentation, and capillary stenosis in STZ mice, which was inhibited by SAA2 silencing (Figure 6H,I).At the same time, Masson staining suggested that SAA2 knockdown could significantly improve the renal collagen fibrosis in diabetic mice (Figure 6H,I).Consistent with this result, the decrease in glomerular podocyte marker WT-1 expression was reversed by SAA2 depletion (Figure 6K).Quantitative analysis showed that the number of podocytes increased with the depletion of SAA2 (Figure 6J).In addition, immunofluorescence detection of podocyte damage markers revealed an increase in Desmin fluorescence intensity and a decrease in ZO-1 in diabetic mice compared to control mice, while knockdown of SAA2 significantly improved these trends (Figure 6K).Histochemical staining for the macrophage marker F4/80 also suggested that knockdown of SAA2 significantly inhibited the infiltration of inflammatory cells in the glomeruli of diabetic mice (Figure 6H).These results confirm that glomerular damage in mice with diabetic kidney disease can be ameliorated by depletion of SAA2.

| DISCUSSION
As a result of the global diabetes pandemic, the incidence of diabetic kidney disease, one of the complications of diabetes, is also increasing.Approximately 40% of individuals with diabetes suffer from diabetic kidney disease. 18odocytes are an important component of the glomerular filtration barrier and contribute to the maintenance of renal physiological function.A study suggested that damage to podocytes is the initial feature of diabetic kidney disease. 19Inflammation caused by sustained high glucose stimulation is a key feature of podocyte damage and plays a key role in the pathogenesis of diabetic kidney disease. 20he NF-κB pathway is a major signaling pathway that contributes to the inflammatory response in glomerular disease and has been found to be involved in podocyte injury. 21A modifications are the most abundant internal chemical modifications in mRNA. 22The discovery of the first demethylase, FTO, facilitated worldwide research on the dynamic m6A modification of RNA. 224][25] The m6A modification, as a dynamically reversible modification, can be installed by methyltransferases and removed by demethylases. 26In recent years, it has been shown that m6A-modified regulatory proteins play important roles in diabetic kidney disease.METTL14 can promote diabetic kidney disease progression through m6A modification of α-Klotho leading to glomerular endothelial cell injury. 27NSD2 can alleviate renal damage and interstitial fibrosis in mice with DKD.METTL3 exerts reno-protective effects by upregulating the expression of NSD2 with m6A modification. 28n the present study, we verified that m6A levels are reduced by m6A spot blotting, whereas the expression of FTO was increased in high glucose-stimulated podocytes.Further, qRT-PCR and Western blot analysis showed that upregulation of FTO was associated with podocyte damage.The podocyte, as a terminally differentiated cell, is not regenerated in the adult body.The damage suffered by podocytes is easily caused by high glucose stimulation.Thus, podocytes are the weakest link in diabetic kidney disease. 29Additionally, FTO-mediated m6A modification of SOCS1 mRNA has been shown to contribute to the pathogenesis of diabetic kidney disease. 30Taken together, it is clear that FTO makes a significant contribution to the progression of diabetic kidney disease.Further exploration of the mechanisms of FTO-mediated podocyte damage is necessary.
FTO, as a demethylating enzyme, usually requires the demethylation of specific mRNAs.For example, FTO promotes the progression of HIF-2α-deficient (HIF-2α low/− ) ccRCC by demethylating BRD9. 31Therefore, how does FTO work for podocyte damage?The development and application of high-throughput sequencing technology have paved the way for further research.To further identify the downstream targets of FTO-mediated podocyte injury, our group performed RNA sequencing (RNA-seq) and methylated RNA immunoprecipitation sequencing (MeRIP-seq) on three cases each of control podocytes, high glucose-cultured podocytes, and high glucose-cultured podocytes with silenced FTO in the previous period.The results showed that the knockdown of FTO resulted in a strong downregulation of SAA2 expression in podocytes and a significant increase in the m6A modification of SAA2.This suggests that FTO may mediate high glucose-induced podocyte injury through the regulation of SAA2.We then measured the RNA and protein levels of SAA2 after silencing and overexpressing FTO in podocytes.MeRIP-qRCR was used to detect the m6A levels of SAA2 mRNA after knockdown and overexpression of FTO.All results are consistent with the sequencing results, suggesting that SAA2 is the downstream target of FTO in the regulation of podocyte damage.Studies have shown that m6A can affect SAA2 expression by regulating mRNA stability, and we conducted stability-related assays to further verify whether FTO also regulates the expression of SAA2 by affecting its mRNA stability.Actinomycin D assays suggested that FTO regulated the expression of SAA2 by acting on the stability of the mRNA.
Serum amyloid A, a class of small molecule proteins, is an important component of the acute phase response (APR) proteins induced by various stimuli. 32SAA is classified into four subtypes: SAA1, SAA2, SAA3, and SAA4, of which SAA1-3 are the acute phase subtypes and SAA4, which is not induced in the acute phase, is the main predominant SAA subtype in the circulation under physiological conditions. 14SAA is mainly present in human tissues in the form of SAA1 and SAA2.In contrast, SAA3 expression was predominant in the mouse kidney. 14,15It was demonstrated that SAA3 is involved in the inflammatory damage of podocytes with the regulation of JAK2 in a mouse model of diabetic kidney disease. 33SAA1 and SAA2 in human tissues are similar in function and distribution to SAA3 in mouse tissues. 34,35RJ Anderberg, RL Meek, KL Hudkins, SK Cooney, CE Alpers, RC Leboeuf, and KR Tuttle 15 found enhanced NF-κB activity in podocytes in response to stimulation with exogenous SAA.Therefore, we suspect that SAA2 is also involved in FTOmediated podocyte injury through activation of the NF-κB signaling pathway.Western blot assays revealed that overexpression of SAA2 promoted the phosphorylation of NF-κB P65 while knockdown of SAA2 did the opposite.SAA2 promotes the expression of TNF-α, IL-6, MCP-1, and Desmin and inhibits the expression of ZO-1 in podocytes.Dual luciferase reporter assays similarly suggested that NF-κB activity was positively correlated with SAA2 expression.Immunofluorescence of podocyte markers and immunohistochemical staining for the macrophage marker F4/80 in SAA2 knockdown STZ mice further suggest that knockdown of SAA2 reduces inflammatory damage of podocytes in diabetic kidney disease.In addition, a cohort study showed that the higher the serum SAA level was, the higher the risk of death and end-stage renal disease in patients with diabetic kidney disease. 36These studies suggest that SAA2 may be a potential target for anti-inflammatory therapy in diabetic kidney disease in the future, and could effectively predict the progression and prognosis of diabetic kidney disease.As the upstream gene of SAA2, FTO was shown to be positively correlated with Desmin, TNF-α, and IL-6, and negatively correlated with ZO-1.The replication experiments also further demonstrated that FTO alleviated inflammatory damage in podocytes by modulating SAA2 expression.However, it has been shown that FTO is negatively correlated with the expression of the inflammatory factors TNF-α and IL-6 in cardiomyocytes.Decreased FTO levels lead to increased m6A methylation and inflammation. 34This suggests that FTO works differently in different cell types and diseases.To further confirm the role of SAA2 in diabetic kidney disease, we performed the in vivo study.NPHS1 is a podocyte-specific promoter. 37We therefore constructed an adenoviral vector-mediated mouse SAA2 shRNA targeting podocytes with specificity for NPHS1.Detection of clinical indicators such as plasma creatinine and BUN suggested that knockdown of SAA2 in podocytes significantly alleviated renal functional impairment in STZ mice.Renal collagen fibrosis is alleviated and reduction in the number of podocytes is mitigated in STZ mice with SAA2 knockdown.
In summary, we found that SAA2 regulated by FTOmediated m6A modification is involved in podocyte inflammatory injury in diabetic kidney disease through regulation of the NF-κB pathway (Figure 7).This study further demonstrates the significance of epigenetic modifications in diabetic kidney disease and offers potential targets for its future control.

F I G U R E 1
High glucose-induced podocyte injury and inflammation.(A-D) Real-time PCR analyses show the mRNA levels of Desmin, ZO-1, TNF-α, and MCP-1 in podocytes treated with LG (5.5 mM d-glucose), HG (30 mM d-glucose), and HO (5.5 mM d-glucose + 24.5 mM mannitol) for 48 h.(E) Western blot analyses show the protein levels of Desmin, ZO-1, IL-6, and TNF-α in podocytes treated with LG, HG, and HO for 48 h.(F, G) Representative immunofluorescence images showing Desmin and ZO-1 in podocytes treated with LG, HG, and HO.Scale bar = 50 μm.Data are shown as mean ± SD. *p < .05,**p < .01,and ***p < .001.The experiment was performed in triplicate.F I G U R E 2 FTO is upregulated in high glucose-treated podocyte and associated with podocyte injury.(A, B) Real-time PCR and Western blot showing the expression of FTO in podocytes treated with LG, HG, and HO.(C, D) Real-time PCR analyses show the mRNA level of Desmin, ZO-1, TNF-α, and MCP-1 in podocytes treated with FTO overexpression lentiviral (OE-FTO) or control (OE-NC) vectors in the presence of LG. (E) Western blot analyses show the protein levels of Desmin, ZO-1, and TNF-α in different groups of podocytes.(F, G) Realtime PCR analyses show the mRNA levels of Desmin, ZO-1, TNF-α, and MCP-1 in podocytes treated with FTO shRNA (sh-FTO) or scramble shRNA (sh-NC) in the presence of HG. (H) Western blot analyses show the protein levels of Desmin, ZO-1, and TNF-α in different groups of podocytes.(I, J) Representative immunofluorescence images of Desmin and ZO-1 in different groups of podocytes.Scale bar = 50 μm.Data are shown as mean ± SD. *p < .05,**p < .01,and ***p < .001.The experiment was performed in triplicate.

F I G U R E 3
FTO regulated SAA2 in a m6A-dependent manner in podocytes.(A) IGV map from MeRIP-seq shows that SAA2 m6A methylation levels in podocytes treated with LG, HG, and FTO shRNA (sh-FTO) in the presence of HG. (B) Hierarchical clustering analysis of DEmRNAs between LG and HG groups.The horizontal axis represents the names of the samples.The vertical axis represents the names of DEmRNAs.Each column indicates a sample and each row indicates a gene.(C) Western blot analyses show the protein level of SAA2 in podocytes treated with FTO overexpression lentiviral (OE-FTO) or control (OE-NC) vectors in the presence of LG. (D) Real-time PCR analyses show the mRNA levels of SAA2 in different groups of podocytes.(E) Real-time PCR analyses show the mRNA levels of SAA2 in podocytes treated with FTO shRNA (sh-FTO) or Scramble shRNA (sh-NC) in the presence of HG. (F) Global m6A level of RNA extracted from podocytes treated with LG, HG, and HO was measured via m6A dot blot assays.Methylene blue staining was used to detect input RNA.(G) MeRIP PCR was performed to detect the m6A modification of SAA2 in different groups of podocytes.(H) Western blot analyses show the protein levels of SAA2 in different groups of podocytes.(I) Podocytes transfected with or without FTO-shRNA (sh-FTO) were treated with ActD (5 μg/mL) for 0, 3, 6, 9, and 24 h.SAA2 mRNA abundance, relative to βactin, was quantified by real-time PCR.Data are shown as mean ± SD. **p < .01 and ***p < .001.The experiment was performed in triplicate.

F I G U R E 4
SAA2 is involved in podocyte damage.(A) Real-time PCR analyses show the mRNA level of SAA2 in different groups of podocytes.(B) Western blot analyses show the protein levels of SAA2 in different groups of podocytes.(C) Real-time PCR analyses show the mRNA level of Desmin and ZO-1 in podocytes treated with SAA2 siRNA (si-SAA2) or negative control (si-NC) in the presence of HG. (D) Real-time PCR analyses show the mRNA level of Desmin and ZO-1 in podocytes treated with SAA2 overexpression plasmid (OE-SAA2) or control (OE-NC) vectors in the presence of LG. (E) Western blot analyses show the protein levels of Demin and ZO-1 in podocytes treated with SAA2 siRNA (si-SAA2) or negative control (si-NC) in the presence of HG. (F) Western blot analyses show the protein levels of Desmin and ZO-1 in podocytes treated with SAA2 overexpression plasmid (OE-SAA2) or control (OE-NC) vectors in the presence of LG. (G) Representative immunofluorescence images of Desmin and ZO-1 in podocytes treated with SAA2 siRNA (si-SAA2) or negative control (si-NC) in the presence of HG and SAA2 overexpression plasmid (OE-SAA2) or control (OE-NC) vectors in the presence of LG.Scale bar = 50 μm.(H) Representative immunofluorescence images of SAA2 and synaptopodin in the kidney tissues of control and DKD patients.Scale bar = 50 μm.Data are shown as mean ± SD. *p < .05,**p < .01,and ***p < .001.The experiment was performed in triplicate.| 13 of 17 LANG et al.

F I G U R E 5
SAA2 contributes to FTO-mediated podocyte injury by regulating the NF-κB pathway.(A) Western blot analyses show the protein levels of P65 and pP65 in podocytes treated with SAA2 overexpression plasmid (OE-SAA2) or control (OE-NC) vectors in the presence of LG. (B) Western blot analyses show the protein levels of P65 and pP65 in podocytes treated with SAA2 siRNA (si-SAA2) or negative control (si-NC) in the presence of HG. (C) Dual luciferase reporter analysis shows the NF-κB activity in podocytes treated with SAA2 overexpression plasmid (OE-SAA2) or control (OE-NC) vectors in the presence of LG. (D) Dual luciferase reporter analysis shows the NF-κB activity in podocytes treated with SAA2 siRNA (si-SAA2) or negative control (si-NC) in the presence of HG. (E) Realtime PCR analyses show the mRNA level of TNF-α and MCP-1 in podocytes treated with SAA2 siRNA (si-SAA2) or negative control (si-NC) in the presence of HG. (F) Real-time PCR analyses show the mRNA level of TNF-α and MCP-1 in podocytes treated with SAA2 overexpression plasmid (OE-SAA2) or control (OE-NC) vectors in the presence of LG. (G) Western blot analyses show the protein levels of TNF-α in podocytes treated with SAA2 siRNA (si-SAA2) or negative control (si-NC) in the presence of HG and in podocytes treated with SAA2 overexpression plasmid (OE-SAA2) or control (OE-NC) vectors in the presence of LG. (H) Real-time PCR analyses show the mRNA levels of Desmin and ZO-1 in different groups of podocytes.(I) Real-time PCR analyses show the mRNA levels of TNF-α and MCP-1 in different groups of podocytes.(J) Western blot analyses show the protein levels of Desmin, ZO-1, and TNF-α in different groups of podocytes.(K-M) Western blot analyses show the protein levels of IL-6, Pp65, and P65 in different groups of podocytes.(N) Representative immunofluorescence images of Desmin and ZO-1 in different groups of podocytes.Scale bar = 50 μm.Data are shown as mean ± SD. *p < .05,**p < .01,and ***p < .001.The experiment was performed in triplicate.| 15 of 17 LANG et al.

F I G U R E 6
SAA2 depletion alleviates kidney damage in STZ mice.(A) Real-time PCR detects the expression of SAA2 in the kidney from WT-CTL (n = 10) and WT-SZT (n = 15) mice.(B-D) The correlation between SAA2 levels in STZ mice kidney and BUN, serum creatinine, and ACR.(E-G) Abumin-to-creatinine ratio (ACR), blood urea nitrogen (BUN; mg/dL) levels, and plasma creatinine (Scr; mg/dL) levels in mice.(H) PAS staining and Masson's trichrome staining of mouse tissues and immunofluorescence staining for macrophage markers F4/80.(I) Fractional mesangial area (FMA%) and Masson staining area were qualified.(J) The number of podocytes was assessed by immunofluorescence staining of WT-1.(K) Representative immunofluorescence images of Desmin, ZO-1, synaptopodin, and WT-1 in mice kidney.Scale bar = 50 μm.Data are shown as mean ± SD. *p < .05,**p < .01,and ***p < .001.The experiment was performed in triplicate.F I G U R E 7 Model depicting the effect of increased expression of FTO in diabetic kidney disease (DKD): FTO regulated SAA2 in a m6Adependent manner and leading to podocyte injury and inflammation by regulating the NF-κB pathway.