Abstract

   Decreased plasma spermine levels are associated with kidney
   dysfunction. However, the role of spermine in kidney disease remains
   largely unknown. Herein, it is demonstrated that spermine oxidase
   (SMOX), a key enzyme governing polyamine metabolism, is predominantly
   induced in tubular epithelium of human and mouse fibrotic kidneys,
   alongside a reduction in renal spermine content in mice. Moreover,
   renal SMOX expression is positively correlated with kidney fibrosis and
   function decline in patients with chronic kidney disease. Importantly,
   supplementation with exogenous spermine or genetically deficient SMOX
   markedly improves autophagy, reduces senescence, and attenuates
   fibrosis in mouse kidneys. Further, downregulation of ATG5, a critical
   component of autophagy, in tubular epithelial cells enhances SMOX
   expression and reduces spermine in TGF‐β1‐induced fibrogenesis in vitro
   and kidney fibrosis in vivo. Mechanically, ATG5 readily interacts with
   SMOX under physiological conditions and in TGF‐β1‐induced fibrogenic
   responses to preserve cellular spermine levels. Collectively, the
   findings suggest SMOX/spermine axis is a potential novel therapy to
   antagonize renal fibrosis, possibly by coordinating autophagy and
   suppressing senescence.

   Keywords: autophagy, cell senescence, renal fibrosis, SMOX, spermine
     __________________________________________________________________

   The relationship between polyamine metabolism and kidney fibrogenesis
   is elucidated through the spermine oxidase (SMOX)/spermine axis.
   Supplementation with spermine or genetic deficiency in SMOX improves
   autophagy, suppresses cell senescence, and alleviates renal fibrosis.
   Moreover, the interplay between autophagy protein 5 (ATG5) and SMOX
   plays a crucial role in regulating the progression of renal fibrosis by
   preserving intracellular spermine levels. Created with BioRender.com.

   graphic file with name ADVS-11-2306912-g004.jpg

1. Introduction

   The prevalence of chronic kidney disease (CKD) is gradually increasing
   worldwide, and CKD may become the fifth leading cause of death in the
   world by 2040.^[ [50]^1 , [51]^2 ^] Despite the different etiologies,
   renal fibrosis, characterized by tubular atrophy, myofibroblast
   activation, and extracellular matrix (ECM) deposition, is the final
   manifestation of CKD.^[ [52]^3 ^] The decline in renal function
   accompanied by fibrotic pathological changes could be induced by many
   factors, among which metabolic disorders have emerged as one of the
   main drivers in recent years.^[ [53]^4 , [54]^5 , [55]^6 ^] Renal
   tubular epithelial cells (TECs) are the major targets of various
   insults, and tubular injury is involved in the initiation and evolution
   of renal fibrosis.^[ [56]^7 ^] Therefore, modulating critical cellular
   metabolic pathways may provide a novel therapeutic strategy to delay
   the progression of renal fibrosis.

   Unlike metabolites that serve as building blocks or energy sources,
   polyamine has been suggested to modulate many fundamental cellular
   processes.^[ [57]^8 ^] Spermine, an end‐product of polyamine
   metabolism, is ubiquitous and its levels are tightly regulated through
   de novo biosynthesis, catabolism, and transportation across the plasma
   membrane. Spermine oxidase (SMOX) is a key and highly inducible enzyme
   in polyamine catabolic pathway. It specifically oxidizes spermine to
   generate spermidine with toxic byproducts, including hydrogen peroxide
   and 3‐aminopropanal. Under cell stress conditions, SMOX is upregulated
   and links to increased reactive oxygen species that can cause cell
   injury and death.^[ [58]^9 , [59]^10 , [60]^11 , [61]^12 ^]
   Alternatively, spermine is degraded via the spermidine/spermine
   N1‐acetyltransferase (SAT1), a rate‐limiting enzyme for polyamine
   reverse transformation.^[ [62]^13 ^] Spermine is an organic cation
   containing multiple amino groups that has a variety of biological
   functions, including immune regulation, maintaining mitochondrial
   homeostasis, enhancing autophagy, and promoting cell proliferation and
   differentiation.^[ [63]^14 ^] It has been demonstrated that abnormal
   spermine metabolism is closely related to the occurrence and
   development of various human diseases.^[ [64]^15 ^] Clinical
   observations have indicated that the plasma spermine level is
   significantly decreased in CKD patients and negatively correlated with
   serum creatinine and urea nitrogen.^[ [65]^16 , [66]^17 ^]
   Consistently, SMOX activity in the blood is markedly increased in CKD
   patients.^[ [67]^18 ^] Spermine supplementation or targeting its
   metabolic enzymes could be potential therapeutic strategies for
   diseases.^[ [68]^19 , [69]^20 , [70]^21 ^] However, spermine metabolism
   in the kidney, especially in TECs, which may be involved in the
   pathogenesis of CKD, has yet to be fully investigated.

   In this study, we demonstrated that SMOX expression was upregulated in
   tubular cells and spermine contents were decreased in fibrotic kidneys
   of CKD patients and mouse models. Supplementation with exogenous
   spermine or genetically deficient SMOX improved autophagy, inhibited
   tubular cell senescence, and ultimately alleviated renal fibrosis.
   Further, ATG5 negatively regulated SMOX protein expression, and the
   ATG5/SMOX interaction preserved cellular spermine content.
   Collectively, our study provides insights into targeting spermine
   metabolism as a novel promising strategy for retarding the progression
   of CKD.

2. Results

2.1. SMOX Is Induced in Injured Kidneys and Is Associated with Fibrogenesis

   Despite reports have shown the tight connection of serum spermine with
   CKD patients, the SMOX expression and content of spermine in kidneys
   and their relationship to renal fibrosis have never been directly
   tested. We employed two renal fibrosis mouse models, the unilateral
   ureteral obstruction (UUO) and the unilateral renal
   ischemia‐reperfusion (UIRI), to determine spermine metabolism in the
   development of kidney fibrosis. The results showed that the mRNA levels
   of Smox were upregulated at day 7 and continued to elevate at day 10
   after UUO (Figure [71] 1A). Western blotting analysis confirmed that
   the protein levels of SMOX were also enhanced in fibrotic kidneys
   (Figure [72]1B,C). Immunohistochemical analysis revealed that during
   fibrosis progression, upregulated SMOX was predominantly concentrated
   in TECs (Figure [73]1D). This distribution was further validated in
   co‐immunostaining for SMOX and the proximal tubular marker LTL (Figure
   [74]S1A, Supporting Information). Likewise, the expression of SMOX in
   the kidneys of UIRI mice also gradually increased with progressive
   fibrosis, showing a similar pattern (Figure [75]1E–H and Figure
   [76]S1B, Supporting Information). Next, mass spectrometry and ELISA
   were performed to accurately determine the changes of spermine content
   in kidney tissues. The mass spectrometry results showed that the
   spermine content in fibrotic kidneys of UIRI and UUO mice was decreased
   compared with that in the normal control (Figure [77]1I). Moreover,
   ELISA data revealed a gradual and significant decrease of spermine
   content in kidneys with prolongated time after UUO and UIRI
   (Figure [78]1J,K). These results together demonstrate a disturbance of
   spermine oxidase/spermine axis in fibrotic kidneys.

Figure 1.

   Figure 1
   [79]Open in a new tab

   SMOX is induced in renal tubular epithelium in CKD mouse models. A)
   qRT‐PCR showed the Smox mRNA levels in the injured kidney of UUO mice.
   B,C) Renal SMOX expression was determined by Western blotting. GAPDH
   was used as a control (n = 6). Dot plots represent quantitative
   densitometric data from Western blotting. D) Representative images of
   Masson trichrome staining (upper panel) and immunostained for SMOX
   (lower panel) in kidneys after UUO. Scale bar, 100 µm. E) qRT‐PCR
   showed the Smox mRNA level in the injured kidney of UIRI mice. F,G)
   Renal SMOX expression was induced at indicated time points after UIRI.
   GAPDH was used as a control (n = 6). Dot plots represent quantitative
   densitometric data from western blotting. H) Representative images of
   Masson trichrome staining (upper panel) and immunostained for SMOX
   (lower panel) in kidneys after UIRI. Scale bar, 100 µm. I) The relative
   fold change in spermine in the kidneys of sham surgery, UUO at day 10,
   or UIRI at day 14 by mass spectrometric detection (n = 6). J,K) The
   relative fold change in spermine in the kidneys at indicated time
   points after UUO and UIRI by ELISA (n = 6). Data are expressed as means
   ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001 versus sham kidneys. UUO,
   unilateral ureteral obstruction; UIRI, unilateral renal
   ischemia‐reperfusion.

   To explore the relevance of the primary polyamine catabolic enzymes in
   human CKD, we performed immunohistochemical staining on renal biopsies
   from 32 CKD patients and nephrectomized normal tissues. As shown in
   Figure [80] 2A, SMOX was weakly expressed in normal human kidneys,
   while it displayed a pronounced expression in renal tubular epithelium
   of CKD patients, and relatively faint staining was detectable in
   interstitial cells. Importantly, SMOX expression was significantly
   increased in kidneys with the aggravation of renal fibrosis
   (Figure [81]2B,C). Correlation analysis revealed that SMOX expression
   was positively correlated with renal fibrosis, and negatively
   correlated with estimated glomerular filtration rate (eGFR)
   (Figure [82]2D,E). In contrast, the protein expression of SAT1, an
   enzyme that positively regulates spermine catabolism, also appeared to
   slightly increased but lacked correlation with renal fibrosis and eGFR
   (Figure [83]S2A,B, Supporting Information). In addition, by mining the
   published database, renal mRNA levels of SMOX and SAT1 in CKD patients
   were upregulated (Figure [84]S2C, Supporting Information), suggesting
   that these enzymes‐mediated abnormal spermine metabolism may be
   involved in the progression of human CKD.

Figure 2.

   Figure 2
   [85]Open in a new tab

   SMOX is elevated in the kidneys of CKD patients. A) Representative
   images of Masson trichrome staining (upper panel) and immunostained for
   SMOX (lower panel) in kidney tissues from CKD patients. Non‐tumor
   kidney tissues from patients with renal carcinoma were used as controls
   (Ctrl). Scale bar, 100 µm. B,C) Graphical representation showing the
   relative abundance of SMOX expression in the kidneys of Ctrl and CKD
   patients with varying degrees of kidney fibrosis. D,E) The correlations
   between the positive area for D) SMOX staining and renal fibrosis and
   E) eGFR in CKD patients. The P value, number of patients (n), and
   Spearman correlation coefficient (R) are indicated on the graph. Data
   are expressed as means ± SEM, ***P < 0.001 versus the Ctrl group. CKD,
   chronic kidney disease; eGFR, estimated glomerular filtration rate;
   Ctrl, control; Mod, moderate; Ser, severe.

2.2. Exogenous Spermine or Knockdown of SMOX Alleviates Renal Fibrosis in
Mice

   To investigate the effect of spermine supplementation on renal
   fibrosis, UUO mice were intraperitoneally injected with vehicle or
   spermine (1 mg kg^−1) starting 2 d before surgery, and continuously
   injected daily for 10 d (Figure [86]S3A, Supporting Information).
   Masson trichrome and Picrosirius red staining revealed substantial
   collagen deposition in the tubulointerstitium 10 days after UUO
   compared with sham‐operated mice, which was attenuated by
   administration of spermine (Figure [87] 3A). Consistently, western
   blotting analysis showed that UUO‐induced renal expression of
   fibronectin (FN), collagen I (Col I), and α‐smooth muscle actin (α‐SMA)
   were dramatically mitigated by spermine (Figure [88]3B). Similar
   results were also obtained in UIRI mouse model (Figure [89]3C,D and
   Figure [90]S3B, Supporting Information). Further, to assess the
   therapeutic effect of spermine on renal fibrosis, mice were treated
   with spermine starting on day 3 after UUO surgery (Figure [91]S3C,
   Supporting Information). Administration of spermine reduced the extent
   of fibrosis, as determined by Masson trichrome and Picrosirius red
   staining, as well as suppressed expression of fibrosis‐related proteins
   (FN and vimentin) (Figure [92]S3D,E, Supporting Information).

Figure 3.

   Figure 3
   [93]Open in a new tab

   Exogenous spermine alleviates renal fibrosis in CKD mouse models. A)
   Masson trichrome and picrosirius red staining images in various groups
   as indicated were shown. Scale bar, 100 µm. Quantitative determination
   of collagen deposition area based on Masson trichrome and picrosirius
   red staining in different groups (right panel) (n = 6). B)
   Representative western blotting and quantitative data for
   fibrosis‐related protein levels of FN, Col I, and α‐SMA in the kidneys
   of various groups as indicated were shown. C) Masson trichrome and
   Picrosirius red staining images in various groups as indicated were
   shown. Scale bar, 100 µm. Quantitative determination of collagen
   deposition area based on masson trichrome and picrosirius red staining
   in different groups. D) Representative western blotting and
   quantitative data for fibrosis‐related protein levels of FN, Col I, and
   α‐SMA in the kidneys of different groups as indicated are shown. Data
   are expressed as means ± SEM, **P < 0.01, ***P < 0.001 versus sham
   controls; #P < 0.05, ##P < 0.01 versus UUO/UIRI with vehicle (n = 6).
   FN, fibronectin; Col I, collagen I; α‐SMA, α‐smooth muscle actin.

   Given that increased SMOX expression and decreased spermine level in
   fibrotic kidneys, as well as the anti‐fibrotic impact of spermine in
   mouse models of renal fibrosis, we hypothesized that elevated
   endogenous renal spermine content exerts favorable effects on
   fibrogenesis. To this end, Smox heterozygote (Smox^+/− ) mice were
   constructed by the CRISPR/Cas9 technique (Figure [94] 4A). Western
   blotting confirmed a significantly decreased SMOX expression in kidneys
   of Smox^+/− mice, as compared with their wild‐type (Smox^+/+ )
   littermates (Figure [95]4B). As expected, a substantial increase in
   spermine concentration measured by ELISA was observed in kidneys from
   Smox^+/− mice (Figure [96]4C). HE staining, Masson trichrome, and
   Picrosirius red staining revealed that there is no obvious difference
   in kidney morphology between sham‐operated Smox^+/+ and Smox^+/− mice
   (Figure [97]4D and Figure [98]S4A, Supporting Information). However,
   SMOX knockdown markedly attenuated UUO‐induced collagen deposition
   (Figure [99]4D). Western blotting further verified that expression of
   profibrotic markers, including FN, Col I, and α‐SMA, were also
   significantly reduced in the UUO kidneys of Smox^+/− mice
   (Figure [100]4E). Similar results were obtained in UIRI mouse model, in
   which interstitial fibrosis were significantly reduced in the injured
   kidneys of Smox^+/− UIRI mice in contrast to Smox^+/+ mice
   (Figure [101]4F,G). These in vivo findings suggest that SMOX/spermine
   axis is critical for inhibiting renal fibrosis.

Figure 4.

   Figure 4
   [102]Open in a new tab

   Knockdown of SMOX ameliorates renal fibrosis in UUO and UIRI mice. A)
   Schematic diagram of the Smox^+/− mouse construction strategy. B)
   Representative western blotting and quantitative data showing the
   expression of SMOX in the kidneys of Smox^+/+ and Smox^+/− mice. *P <
   0.05 versus Smox^+/+ (n = 3). C) Graphical representations showing the
   relative fold‐change of spermine in the kidney of Smox^+/+ and Smox^+/−
   mice. D) Representative images of kidney tissues from indicated groups
   with either Masson trichrome (upper panel) or Picrosirius red staining
   (lower panel). Scale bar, 100 µm. Quantitative assessment of the
   collagen deposition area based on masson trichrome and picrosirius red
   staining in different groups. E) Representative western blotting and
   quantitative data showing fibrosis‐related protein levels of FN, Col I,
   and α‐SMA in the kidneys of different groups as indicated are shown. F)
   Representative images of kidney tissues from indicated groups with
   either masson trichrome staining (upper panel) or picrosirius red
   staining (lower panel). Scale bar, 100 µm. Quantitative assessment of
   the collagen deposition area based on masson trichrome and Picrosirius
   red staining in different groups. G) Representative western blotting
   and quantitative data for fibrosis‐related protein levels of FN, Col I,
   and α‐SMA in the kidneys of different groups as indicated are shown.
   Data are expressed as means ± SEM, **P < 0.01, ***P < 0.001 versus
   Smox^+/+ sham; #P < 0.05, ##P < 0.01 versus Smox^+/+ mice with either
   UUO or UIRI (n = 6).

2.3. Spermine Supplementation or SMOX Depletion Abrogates TGF‐β1‐Induced ECM
Production in TECs

   TGF‐β1, as a key factor driving fibrosis, actively alters metabolism in
   diverse cell types.^[ [103]^22 , [104]^23 ^] To ascertain the direct
   role of spermine in ECM production in kidney cells, we applied TGF‐β1
   stimulation of the mouse tubular epithelial cells (mTECs) line as an in
   vitro model. The results showed that the steady‐state mRNA and protein
   levels of SMOX were increased at both 24 and 48 hours after TGF‐β1
   stimulation (Figure [105] 5A,B), accompanied with decreased cellular
   spermine content (Figure [106]5C). Furthermore, pretreatment with a
   relatively high dose (10 × 10^−6 m) of spermine caused a reduction in
   TGF‐β1‐induced FN expression (Figure [107]5D). Concordantly,
   immunofluorescence staining showed that 10  × 10^−6 m spermine
   incubation significantly inhibited vimentin and α‐SMA expression under
   TGF‐β1 in mTECs and restored loss of E‐cadherin at cell junctions
   (Figure [108]5E and Figure [109]S4B, Supporting Information). To
   identify perturbed biologic networks and understand the biological
   processes and signaling pathways involved in spermine‐mediated
   response, mTECs were treated with TGF‐β1 in the absence or presence of
   spermine (10  × 10^−6 m) for 24 h, followed by transcriptome sequencing
   analyses. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway
   analysis showed that the downregulated differential genes were mainly
   enriched in ECM receptor and focal adhesion signaling pathways
   (Figure [110]5F). Notably, the heat map display of differentially
   expressed genes also demonstrated that spermine significantly reversed
   the expression of fibrosis‐related genes induced by TGF‐β1
   (Figure [111]5G). To verify these changes indeed existed, we performed
   qRT‐PCR on specific genes in the pathways identified by sequencing data
   analyses and found that pretreatment with spermine led to reduced mRNA
   levels of ECM‐associated factors such as FN, Col I, α‐SMA, connective
   tissue growth factor (CTGF), and periostin (POSTN) induced by TGF‐β1
   (Figure [112]5H).

Figure 5.

   Figure 5
   [113]Open in a new tab

   Spermine inhibits TGF‐β1‐induced ECM production in renal tubular cells
   (mTECs). A,B) Cells were treated with 10 ng/mL of TGF‐β1 for the
   indicated time periods. The Smox mRNA and SMOX protein levels were
   analyzed by A) qRT‐PCR and B) Western blotting, respectively. *P <
   0.05, ***P < 0.001 versus Ctrl (n = 3). C) Graphical representation
   showing the relative level of spermine in mTECs with or without TGF‐β1
   stimulation, measured by ELISA. *P < 0.05 versus Ctrl (n = 5). D) Cells
   were preincubated with 5 × 10^−6 and 10 × 10^−6 m of spermine (Spm) for
   12 h and then treated with 10 ng mL^−1 of TGF‐β1 for 24 h. Western
   blotting and quantitative data of FN expression are presented. ***P <
   0.001 versus the Ctrl group; ##P < 0.01 versus TGF‐β1‐stimulated cells
   without Spm treatment (n = 3). E) Representative micrographs with
   immunofluorescence staining of vimentin, α‐SMA, and E‐cadherin are
   shown in different groups. Scale bar, 10 µm. F) Kyoto Encyclopedia of
   Genes and Genomes (KEGG) enrichment analysis showed that differential
   signaling pathways were identified by the downregulated differential
   genes between the TGF‐β1+vehicle and TGF‐β1+Spm groups. G) Heatmap of
   RNA‐seq data showing the expression of fibrosis‐related genes in
   different groups (n = 3). H) qRT‐PCR showing the relative mRNA levels
   of fibrosis‐related genes in different groups. *P < 0.05, **P < 0.01,
   ***P < 0.001 versus control cells treated without Spm; #P < 0.05, ##P <
   0.01, ### P < 0.001 versus cells‐treated with TGF‐β1 (n = 3). I) Cells
   were transfected with either of control siRNA (siNC) or siRNA specific
   for SMOX (siSMOX) followed by treatment with 10 ng mL^−1 TGF‐β1 for 24
   h. Western blotting and quantitative data of FN expression are
   presented. ***P < 0.001 versus TGF‐β1‐untreated cells; #P < 0.05 versus
   TGF‐β1‐treated cells without siSMOX (n = 3). J) Cells were transfected
   with either control vector or SMOX‐overexpression plasmid followed by
   treatment with 10 ng mL^−1 TGF‐β1 for 24 h. Western blotting and
   quantitative data of FN expression are presented. **P < 0.01 versus
   TGF‐β1‐untreated cells; #P < 0.05 versus TGF‐β1‐treated cells without
   SMOX overexpression (n = 3).

   Since dysregulation of SMOX directly leads to changes in cellular
   spermine levels and our in vivo study showed that Smox^+/− mice had
   less renal fibrosis compared with wild‐type controls (Figure [114]4),
   we therefore sought to determine the effect of SMOX on TGF‐β1‐induced
   fibrogenesis. The results showed that depletion of SMOX by siRNA
   significantly reduced TGF‐β1‐induced FN expression compared to scramble
   siRNA (siNC) (Figure [115]5I). Conversely, overexpression of SMOX
   markedly promoted FN upon TGF‐β1 stimulation (Figure [116]5J). These in
   vitro data suggest that spermine metabolism may play an antagonistic
   role in ECM production.

2.4. Exogenous Spermine or Knockdown of SMOX Improves Autophagy and Inhibits
Cellular Senescence in Renal Fibrosis

   To clarify the mechanisms by which spermine metabolism regulates renal
   fibrosis, a proteomic study was performed on the kidney cortices from
   the UUO mouse model. As shown in Figure [117] 6A, a total of 70
   proteins were markedly down‐regulated and 168 proteins were
   up‐regulated in Smox^+/− mice compared to Smox^+/+ mice. Strikingly,
   gene ontology analysis confirmed that the differential proteins were
   significantly enriched in autophagy‐related signaling pathway
   (Figure [118]6B), demonstrating that the SMOX/spermine metabolism is
   closely associated with autophagy in renal fibrosis. Furthermore, Smox
   ^+/‐ markedly abrogated increased autophagy receptor and selective
   substrate p62 in UUO mice compared to Smox ^+/+, suggesting that SMOX
   restrained autophagy (Figure [119]6C). As both spermine and autophagy
   play a role in aging,^[ [120]^24 , [121]^25 ^] we examined whether the
   SMOX/spermine axis affects cellular senescence in the injured kidney.
   Significant reductions in p16‐positive senescent tubules
   (Figure [122]6D) and blue granule‐positive areas stained with
   Senescence‐associated β‐gal (SA‐β‐gal) (Figure [123]6E) were found in
   Smox ^+/‐ UUO kidneys compared to controls. These alterations were
   validated in UIRI mice, where SMOX knockdown significantly improved
   autophagy and delayed cellular senescence in fibrotic kidneys (Figure
   [124]S5A–C, Supporting Information).

Figure 6.

   Figure 6
   [125]Open in a new tab

   Downregulation of SMOX represses autophagy impairment and cellular
   senescence in renal fibrosis. A) Volcanic map on proteomic analysis
   shows significantly differential proteins in the kidneys of Smox^+/+
   and Smox^+/− mice with UUO (n = 3). B) GO enrichment analysis shows
   several important pathways. Data are expressed as means ± SEM. C,D)
   Representative micrographs and quantitative data with
   immunohistochemistry staining of p62 and p16 in the kidneys of Smox^+/+
   and Smox^+/− mice. Scale bar, 25 µm. ***P < 0.001 versus the Smox^+/+
   sham; #P < 0.05 versus Smox^+/+ UUO 10d (n = 3–6). E) Representative
   micrographs and quantitative data with SA‐β‐gal activity staining in
   the kidneys of Smox^+/+ and Smox^+/− mice. Scale bar, 50 µm. ***P <
   0.001 versus the Smox^+/+ sham; #P < 0.05 versus Smox^+/+ UUO 10d (n =
   6).

   We next examined renal autophagy and cellular senescence in UUO and
   UIRI mice supplemented with spermine. As shown in Figure [126] 7A,
   compared with vehicle‐treated UUO mice, spermine administration
   significantly increased LC3‐II levels, ATG5, and concomitant
   degradation of p62 in obstructed kidneys, indicating the activation of
   autophagy. Meanwhile, spermine reduced UUO‐induced expression of
   senescence markers p53 and p21 in the kidneys. The p62 and p16
   staining, as well as the SA‐β‐gal positive area, which were markedly
   induced in the UUO kidney, were also abrased by spermine
   (Figure [127]7B–D). Spermine supplementation in UIRI mice resulted in
   similar outcomes (Figure [128]S5D–G, Supporting Information).
   Collectively, these data demonstrate that impaired autophagy and
   accelerated cellular senescence occur in fibrotic kidneys, which is
   alleviated by SMOX knockdown or spermine supplementation.

Figure 7.

   Figure 7
   [129]Open in a new tab

   Spermine regulates the activation of autophagy and inhibits cellular
   senescence in renal fibrosis. A) Representative western blotting and
   quantitative data for indicated proteins in kidneys in different groups
   as indicated are shown. **P < 0.01, *P < 0.05 versus sham; #P < 0.05
   versus UUO mice treated with vehicle (n = 6). B,C) Representative
   micrographs and quantitative data with immunohistochemistry staining of
   p62 and p16 in different groups. Scale bar, 25 µm. ***P < 0.001 versus
   the sham; #P < 0.05 versus UUO mice treated with vehicle (n = 6). D)
   Representative micrographs and quantitative data with SA‐β‐gal activity
   staining in different groups. Scale bar, 50 µm. ***P < 0.001 versus the
   sham; ##P < 0.01 versus UUO mice treated with vehicle (n = 6).

2.5. The Interaction between ATG5 and SMOX Affects the Spermine Levels in
Kidney

   An interaction loop has been reported between polyamine metabolism and
   autophagy,^[ [130]^26 ^] we therefore sought to unravel the importance
   of this crosstalk in kidney fibrogenesis. AlphaFold 2.3.0 multimers
   were used to predict protein–protein interactions, yielding reliable
   DockQ values, which showed the highest DockQ for SMOX‐ATG5 interaction
   (Figure [131] 8A). Coimmunoprecipitation in mTECs confirmed that SMOX
   did have a strong interaction with ATG5 but a weak interaction with
   ATG3 and ATG7 (Figure [132]8B). Meanwhile, the crystal structure of the
   protein was analyzed by molecular docking, and it was found that ATG5
   and SMOX proteins derived from human and mouse had multiple potential
   binding sites (Figure [133]8C and Figure [134]S7A, Supporting
   Information). Further, we utilized coimmunoprecipitation in combination
   with mass spectrometry to detect what ATG5‐binding partners are in
   mTECs. Consistently, we were able to find that these ATG5‐interacting
   proteins included 16 amino acid metabolism‐related proteins, in which
   SMOX was present (Figure [135]8D). The ATG5‐SMOX interaction was
   further verified by co‐immunoprecipitation experiments in mTECs and
   HEK‐293T cells following overexpression of both proteins under basal
   conditions (Figure [136]8E,F) or exposure to TGF‐β1 (Figure [137]8G).
   Since the most important non‐covalent interaction in biomolecules is
   hydrogen bonding,^[ [138]^27 ^] an ATG5 multisite‐mutated (MMut)
   plasmid based on the potential binding sites of hydrogen bonds was
   constructed and transfected into mTECs, with wild‐type ATG5 as a
   control. Unlike wild‐type ATG5, the ATG5 MMut protein failed to
   interact with SMOX, indicating that ATG5 and SMOX interact through
   these predicted binding sites (Figure [139]S7B, Supporting
   Information). Additionally, based on the prediction that ATG5 GLU244
   binds to 3 loci of SMOX via hydrogen bonds, we examined the possibility
   of GLU244 as a potential key binding site. However, the results showed
   that the E244A single‐site mutation (SMut) of ATG5 had no significant
   effect on the interaction between ATG5 and SMOX (Figure [140]S7C,
   Supporting Information), suggesting that the interaction between the
   two does not mainly depend on this site, but on multiple sites.
   Regardless, colocalization of ATG5 and SMOX was observed by
   immunofluorescent staining in 293T cells (Figure [141]8H) and
   TGF‐β1‐stimulated mTECs (Figure [142]8I), as well as in the kidneys of
   CKD patients (Figure [143]S7D, Supporting Information).

Figure 8.

   Figure 8
   [144]Open in a new tab

   ATG5 interacts with and suppresses the expression of SMOX. A) Images
   represent the 3D structure and the potential binding of SMOX (green)
   proteins and ATG5 (purple), ATG7 (yellow), and ATG3 (blue)
   respectively. DockQ represents the confidence of the potential binding.
   B) Co‐immunoprecipitation of SMOX and autophagy‐related protein ATG5,
   ATG7, ATG3. C) Images represent the 3D structure of ATG5 (green) and
   SMOX (blue) proteins, and the dotted box indicates the binding site of
   ATG5 and SMOX protein. D) mTECs were transfected with ATG5
   overexpression plasmid, and cell lysate was immunoprecipitated (IP)
   with anti‐ATG5 antibody followed by examination the affected proteins
   by mass spectrometry. E–G) mTECs or 293T cells were treated with or
   without TGF‐β1 as indicated. Coimmunoprecipitation of ATG5 and SMOX. H)
   Double immunofluorescence staining of ATG5 and SMOX in 293T cells.
   Scale bar, 5 µm. I) Double immunofluorescence staining of ATG5 and SMOX
   in TGF‐β1‐treated mTECs. Scale bar, 10 µm. J) Representative western
   blotting and quantitative data for protein levels of SMOX and ATG5 in
   mTECs transfected with either control siRNA (siNC) or siRNA specific
   for ATG5 (siATG5). *P < 0.05 versus siNC (n = 3). K) Representative
   western blotting and quantitative data for SMOX and ATG5 protein levels
   in mTECs transfected with ATG5 empty vector or overexpression plasmid.
   *P < 0.05 versus empty vector (n = 3). L) Western blotting showing SMOX
   protein expression in the kidneys of Atg5^+/+ and Atg5^−/− mice with or
   without UUO. M,N) mTECs were transfected with either negative control
   siRNA (siNC) or siRNA specific for ATG5 (siATG5) followed by treatment
   with 10 ng mL^−1 of TGF‐β1 for 24 h. M) The heatmap shows the different
   metabolites in cells by non‐target metabolomics. N) The graphical
   representation shows the relative level of intracellular spermine. *P <
   0.05 versus siNC (n = 5). O) Graphical representation showing the
   relative spermine levels by ELISA in the kidneys of Atg5^+/+ and
   Atg5^−/− mice with or without UUO. **P < 0.01, ***P < 0.001 versus
   Atg5^+/+ sham controls; #P < 0.05 versus Atg5^+/+ UUO mice (n = 8).
   Data are expressed as means ± SEM.

   The clear evidence of the combination of the ATG5 and SMOX drove us to
   further observe the cross‐talk between their expression and function.
   Strikingly, downregulating ATG5 increased the protein expression of
   SMOX in mTECs (Figure [145]8J), while, overexpressed ATG5 decreased
   SMOX expression (Figure [146]8K). Further, in UUO model, it was also
   found that SMOX expression in the kidneys of Atg5^−/− mice was
   significantly increased compared with Atg5^+/+ mice (Figure [147]8L).
   Collectively, both in vitro and in vivo results indicate that ATG5
   negatively regulates SMOX protein expression. To this end, mTECs were
   transfected with siRNA against ATG5 or control siRNA, followed by
   metabolomic analysis to measure intracellular spermine levels in
   response to TGF‐β1 stimulation. Among differential metabolites,
   downregulation of ATG5 significantly decreased TGF‐β1‐induced spermine
   content compared to control siRNA cells (Figure [148]8M,N). In vivo, we
   generated tubule‐specific Atg5‐knockout mice using a Cre‐LoxP
   recombination system to better elucidate the effect of ATG5 expression
   on spermine metabolism in renal TECs. The spermine contents in the
   kidneys of Atg5^+/+ and Atg5^−/− mice, with or without UUO, were
   detected by ELISA. As shown in Figure [149]8O, in the sham‐operated
   group, Atg5^−/− mice exhibited lower spermine contents than their
   wild‐type littermates, denoting that impaired autophagy may lead to
   rapid spermine consumption under physiological conditions. Consistent
   with our observation above (Figure [150]1I,J), spermine levels were
   reduced in Atg5^+/+ UUO mice relative to sham‐operated Atg5^+/+ or
   Atg5^−/− mice, and this response was significantly aggravated in
   Atg5^−/− mice after UUO, suggesting that inhibition of autophagy can
   further exacerbate spermine decline under diseased states. On the other
   side, the effect of spermine was achieved by enhancing autophagy.
   Spermine supplementation in Atg5^+/+ UUO or UIRI mice could reduce the
   accumulation of collagen fibers in the kidneys by half, while this
   effect disappeared in Atg5^−/− mice receiving the same treatment
   (Figure [151]S6, Supporting Information). And the ATG5/SMOX interaction
   may be involved in the occurrence and development of renal fibrosis
   through regulating spermine content.

2.6. Spermine Inhibits TGF‐β1‐Induced Fibroblast Activation and Proliferation

   Fibroblasts are important effector cells in the process of renal
   fibrosis. We noted that SMOX knockdown markedly attenuated UUO‐induced
   expression of the marker α‐SMA in activated fibroblasts (Figure [152]
   9A). We thus evaluated whether exogenous spermine could suppress
   fibroblast activation. To this end, cultured fibroblasts (NRK‐49F) were
   incubated with TGF‐β1 in the absence or presence of spermine. As shown
   in Figure [153]9B, spermine (10 × 10^−6 m) repressed TGF‐β1‐induced
   mRNA expression of α‐SMA and Col I. Immunofluorescent staining also
   showed that TGF‐β1‐triggered α‐SMA expression was reduced by treatment
   with spermine (Figure [154]9C). Further, spermine supplementation
   inhibited TGF‐β1‐induced expression of α‐SMA and vimentin
   (Figure [155]9D). Collectively, these data suggest the ability of
   spermine to abolish the myofibroblastic activation of renal
   fibroblasts.

Figure 9.

   Figure 9
   [156]Open in a new tab

   Spermine inhibits TGF‐β1‐induced fibroblast proliferation and
   activation. A) Representative micrographs and quantitative data with
   immunofluorescence staining of α‐SMA in the kidneys of Smox^+/+ and
   Smox^+/− mice. Scale bar, 50 µm. ***P < 0.001 versus the Smox^+/+ sham;
   ##P < 0.01 versus Smox^+/+ UUO 10d (n = 3). B) NRK‐49F cells were
   preincubated with 10 × 10^−6 m spermine (Spm) for 12 h and then treated
   with TGF‐β1 (5 ng mL^−1) for 24 h and subjected to qRT‐PCR detection.
   C) Representative micrographs with immunofluorescence staining of α‐SMA
   are shown. Scale bar, 10 µm. DAPI denotes the nuclei. D) Western
   blotting and quantitative data of α‐SMA and vimentin in different
   groups as indicated are shown. E–G) Representative micrographs with
   immunofluorescence staining and flow cytometry plots showing the mean
   fluorescence intensity of EdU in different groups. Scale bar, 10 µm.
   Data are expressed as means ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001
   versus Ctrl; #P < 0.05, ##P < 0.01 versus TGF‐β1; ns, means no
   significant (n = 3).

   To investigate the effect of spermine on TGF‐β1‐induced fibroblast
   proliferation, we used an EdU kit to measure the percentage of
   EdU‐positive NRK‐49F cells. As shown in Figure [157]9E, the number of
   EdU‐positive cells was dramatically increased upon TGF‐β1 stimulation,
   however, spermine obviously attenuated this response. Flow cytometry
   analysis revealed that under TGF‐β1 stimulation, the percentage of
   proliferating NRK‐49F cells increased from 58.75% in the ground state
   to 82.52%, while in spermine‐treated cells, this stimulating effect was
   diminished to 62.84% (Figure [158]9F,G), suggesting that spermine also
   alleviated renal fibroblast proliferation.

3. Discussion

   In this study, we demonstrate that SMOX protein expression was
   increased predominantly in tubular epithelium in fibrotic kidneys of
   patients and mice. Upregulated renal SMOX was positively correlated
   with kidney fibrosis and function decline in CKD patients. Notably,
   supplementation with spermine or genetically deficient SMOX improved
   autophagy, repressed cell senescence, and alleviated renal fibrosis.
   Further, ATG5 physically interacted with SMOX that affected cellular
   spermine content. Our findings demonstrate a spermine oxidase/spermine
   axis that links polyamine metabolism to kidney fibrogenesis and suggest
   that targeting spermine pathway is a novel potential therapeutic avenue
   for CKD (Figure [159] 10 ).

Figure 10.

   Figure 10
   [160]Open in a new tab

   Schematic illustration of the potential mechanism of SMOX/spermine axis
   against renal fibrosis. Created with BioRender.com.

   SMOX, a key catabolic enzyme in polyamine metabolism, was easily
   induced by various stimuli, and upregulation of SMOX reduced
   intracellular spermine pools.^[ [161]^12 , [162]^28 ^] However, the
   changes in SMOX expression and spermine content in experimental kidney
   disease remain controversial. It has been suggested that Smox mRNA
   level and spermine content were significantly decreased in fibrotic
   kidneys at day 21 after IRI,^[ [163]^29 ^] while in another study, Smox
   mRNA level was markedly increased on day 14 after UUO without affecting
   spermine levels in the kidneys.^[ [164]^30 ^] Interrogation of
   available single‐cell RNA‐seq datasets (http://humphreyslab.com), on
   the other hand, a recent study demonstrated a distinct proportion of
   tubular epithelial cells with highly and broadly upregulated
   transcription level of Smox in UUO and IRI mouse kidneys.^[ [165]^31 ^]
   Our data were consistent with these findings and showed that SMOX was
   noticeably increased at both mRNA and protein levels in mouse fibrotic
   kidney tissues and that SMOX was predominantly expressed in tubular
   epithelium in fibrotic kidneys of patients and mice. Further, our
   results extend previous studies showing a reduction of spermine content
   in the kidneys of UUO and IRI mice,^[ [166]^29 , [167]^32 ^] as well as
   decreased renal tubular epithelial cellular levels of spermine upon
   TGF‐β1 exposure.^[ [168]^32 ^] In vivo, we also found that the levels
   of spermidine were significantly increased in the kidneys of both
   fibrosis models, while putrescine was elevated only in UUO kidneys. The
   dynamic changes in SMOX and polyamine levels during renal fibrosis may
   differ depending on the model and its observation time points. Since
   spermine is the end‐product of the polyamine metabolic pathway, reduced
   spermine content in fibrotic kidneys may be caused by multiple
   processes, including enzymes that related to spermine metabolism. As
   such, our study focused on increased SMOX promoting spermine catabolism
   in fibrotic kidneys. Apart from endogenous metabolism, food
   consumption, and intestinal microbiota are the other two sources of
   spermine.^[ [169]^10 ^] However, CKD patients often exhibit intestinal
   flora disorders and impaired intestinal barrier,^[ [170]^33 ^] which
   may also lead to reduced intake of spermine. So far, it appears
   uncertain fibrosis‐associated changes in SMOX expression and spermine
   content, more studies are needed to confirm these findings.

   Using proteomics, here we found that the biological process of
   autophagy was affected in fibrotic kidneys of Smox knockdown mice,
   prompting us to further focus on the autophagy signaling pathway in the
   mechanistic exploration of SMOX. Autophagy, as an important regulator
   of cellular metabolism, is essential for maintaining kidney
   homeostasis, structure, and function.^[ [171]^34 ^] Furthermore, the
   reaction products spermine and spermidine of SMOX are potent autophagy
   agonists.^[ [172]^25 , [173]^35 , [174]^36 , [175]^37 ^] Yet, a direct
   relationship between SMOX and autophagy has not been reported. Our
   results demonstrated that either spermine administration or genetically
   deficient SMOX enhanced UUO‐induced autophagy and attenuated renal
   fibrosis, reinforcing the SMOX/Spermine axis and autophagy interaction
   as the core mechanism of CKD. Cellular senescence is stress response
   essential for homeostasis.^[ [176]^38 ^] Tubular epithelial senescent
   cells exhibit a secretory phenotype, which can drive chronic
   inflammation and accelerate the progression of renal fibrosis.^[
   [177]^39 ^] Studies have revealed that spermine is significantly
   decreased in kidneys of aged rats,^[ [178]^40 ^] and exogenous
   polyamine supplementation delays age‐related diseases.^[ [179]^41 ,
   [180]^42 , [181]^43 ^] Moreover, diets rich in polyamines reduce
   age‐related glomerulosclerosis and mortality in elderly mice.^[
   [182]^44 ^] Thus, it is conceivable that action of spermine in vivo may
   go beyond the simple activation of autophagy. Consistent with previous
   studies showing that spermine administration reduced renal p53 and p21
   expression, as well as the activity of SA‐β‐gal in UUO mice.
   Collectively, our findings indicated that spermine protected against
   renal fibrosis through enhancing autophagy activity and inhibiting cell
   senescence.

   It is worth noting that autophagy has an essential role in maintaining
   cellular polyamine metabolic homeostasis.^[ [183]^26 ^] The readily
   interaction between SMOX and ATG5 was identified by
   co‐immunoprecipitation and mass spectrometry. Furthermore, the
   structure of the mouse SMOX sequence with 34‐268 and 309‐544 site
   features for the oxidase site, which overlaps with the ATG5 and SMOX
   binding sites analyzed with the interPro protein structure domain
   database (https://www.ebi.ac.uk/interpro/search/sequence/). Notably,
   downregulation of ATG5 increased SMOX protein expression, whereas
   overexpression of ATG5 reduced SMOX expression. Thus, we speculate that
   the interaction between ATG5 and SMOX may positively affect SMOX
   function, thereby resulting in altered spermine levels. To explore
   whether autophagy is involved in spermine regulation, we suppressed
   autophagy by depleting ATG5, a key regulation of autophagy, in renal
   TECs. Our in vivo and in vitro results revealed that ATG5 knockdown
   aggravated a decrease in spermine during renal fibrosis. Subsequently,
   the protective role of spermine supplementation was disrupted in
   ATG5‐deficient mice, which in part indicates that ATG5‐mediated
   autophagy is required for the protective effects of spermine against
   renal fibrosis. Collectively, these results indicate that the
   SMOX/spermine axis coordinates ATG5‐mediated autophagy to orchestrate
   renal fibrosis.

   Injured tubular cells produce profibrotic factors, such as TGF‐β1,
   leading to fibroblast activation and renal fibrosis.^[ [184]^45 ^] Our
   results demonstrated that spermine supplementation effectively
   inhibited extracellular matrix production by mTECs and fibroblasts.
   Additionally, SMOX knockdown in mice markedly attenuated the
   UUO‐induced fibroblast accumulation and depletion of SMOX in mTECs
   significantly reduced TGF‐β1‐induced FN expression. Notably, increased
   expression of SMOX was only observed in renal TECs but not in
   fibroblasts (data not shown). Therefore, SMOX‐mediated spermine
   reduction is most likely to occur primarily in TECs, thereby promoting
   epithelial phenotype loss and in vivo fibroblast activation, ultimately
   leading to ECM accumulation and renal fibrosis.

   The present study has several limitations and leaves questions
   unanswered. First, our SMOX global knockdown murine model is not
   restricted to renal tubular epithelial cells, further study is needed
   to elucidate the context and cell‐type‐dependent role of SMOX by using
   a cell‐specific inducible model. Second, although we confirmed the
   interaction of SMOX with the potential binding sites of ATG5,
   additional exploration is warranted to ascertain the exact binding
   sites within ATG5. Thirdly, beyond autophagy and cellular senescence,
   whether other signaling pathways are involved in the SMOX/Spermine
   axis’ effects in renal fibrosis as well as their complex interrelated
   mechanisms remain to be explored.

   In conclusion, we have identified the protective roles of downregulated
   SMOX expression and increased spermine content in TGF‐β1‐induced
   fibrogenesis in vitro and kidney fibrosis in vivo, using two
   independent models of kidney fibrosis. Furthermore, we demonstrated
   that ATG5 physically interacted with SMOX to modulate the occurrence
   and development of renal fibrosis through maintaining intracellular
   spermine levels. Our study suggests that targeting spermine metabolism
   may have therapeutic benefits for patients with renal fibrosis.

4. Experimental Section

Human Kidney Samples

   Human kidney specimens collected from patients with CKD in the First
   Affiliated Hospital of Sun Yat‐Sen University, with written informed
   consent were analyzed. The clinical characteristics of those patients
   at the initial biopsy were described in Table [185]S1 (Supporting
   Information). The severity of renal fibrosis was determined by the
   Masson trichrome staining according to the previous report.^[ [186]^46
   ^] The renal fibrosis degree was classified according to Banff
   classification: grade 0 was no fibrosis, grade 1 was mild fibrosis
   (collagen fibers accounted for less than 25% of the kidney tissue),
   grade 2 was moderate fibrosis (accounted for 25%−50%), and grade 3 was
   severe fibrosis (accounted for more than 50%).^[ [187]^47 ^] The linear
   regression between SMOX expression and renal fibrosis was evaluated
   according to previous reports.^[ [188]^48 ^] Non‐tumor renal tissues
   from eight patients with renal cell carcinoma were used as control.
   This study was approved by the First Affiliated Hospital of Sun Yat‐Sen
   University Institutional Review Board (project number: (2023)050).

Mice and Animal Models

   Male wild‐type C57BL/6 mice and Smox‐KO mice(Strain NO.T027342)were
   purchased from GemPharmatech (Nanjing, Jiangsu, China). Atg5 ^flox/flox
   mice were purchased from the RIKEN BioResource Center (Tsukuba,
   Ibaraki, Japan). Proximal tubule‐specific Atg5^−/− mice were generated
   by crossing an Atg5 ^flox/flox mouse with a Kap‐cre mouse, as described
   previously.^[ [189]^49 ^] The Kap‐cre mice expressing Cre recombinase
   under the control of the Kap promoter were from the Jackson Laboratory
   (Bar Harbor, Maine, USA). Genotyping was identified by the polymerase
   chain reaction (PCR) from genomic DNA obtained from tail biopsies, the
   primer sequences were shown in Table [190]S2 (Supporting Information).
   Mice were maintained under pathogen‐free facility of the Laboratory
   Animal Center, Sun Yat‐Sen University. Animal experiments were approved
   by the Ethics Committee for the Use of Experimental Animals of Yat‐Sen
   University (project number: 2022001708).

   Mice aged 8–12 weeks were used for experiments. UUO and UIRI surgery
   were performed as described previously.^[ [191]^46 , [192]^49 ^]
   Briefly, for the UUO injury model, mice were anesthetized, and the
   abdomen was opened with a midline incision, the left ureter was exposed
   and ligated twice with 5‐0 silk sutures. Sham‐operated mice had their
   left ureters exposed and manipulated but ligated. Mice were sacrificed
   7 or 10 d after operation. For the UIRI model, mice were anesthetized
   and the left renal pedicle was clamped for 45 min at 37 °C through a
   left abdominal incision. Mice were sacrificed 7 or 14 d after
   reperfusion. For some experiments, 1 mg kg^−1 of spermine (Cat#S4264,
   Sigma‐Aldrich, St Louis, MO, USA) was injected intraperitoneally as
   indicated in the results section. Blood and kidneys were collected for
   analysis.

Cell Culture and Treatment

   Mouse kidney proximal tubular epithelial cells (mTEC, CRL‐3361), normal
   rat kidney fibroblasts (NRK‐49F, CRL‐1570), and human 293T (293T,
   CRL‐11268) were purchased from American Type Culture Collection. mTEC
   cells and 293T cells were cultured in Dulbecco's modified Eagle
   medium/Nutrient Mixture F‐12 (Cat #C11330500BT, Gibco, Thermo Fisher
   Scientific, Halethorpe, MD), supplemented with 10% fetal bovine serum
   (Cat#10099‐141C, Gibco). NRK‐49F were maintained in Dulbecco's Modified
   Eagle Medium (Cat#C11995500BT, Gibco) with 10% fetal bovine serum.
   Cells were pretreated with or without spermine (5 × 10^−6 or 10 × 10^−6
   m) for 12 h, followed by treatment with 10 ng mL^−1 TGF‐β1 (Cat#240‐B,
   R&D systems, Minneapolis, MN, USA) for an additional 24 h. The siRNA
   transfection was performed using Lipofectamine 2000 (Cat#11668019,
   Invitrogen, Thermo Fisher Scientific, Halethorpe, MD) or plasmids were
   introduced into cells using PolyJet regent (Cat#SL100688, SignaGen
   Laboratories, USA), according to the manufacturer's instructions. The
   MMut plasmid has mutated a total of 10 amino acid binding sites of ATG5
   that potentially bind to SMOX, including T2A, D3A, R9A, D10A, V11A,
   D160A, R161A, H241A, E244A, and M246A, and The SMut plasmid (mutated
   with E244A) were used for transfection and coimmunoprecipitation of
   mTECs. The information of targeted siRNA and plasmids are shown
   in Table [193]S3 (Supporting Information).

Measurement of Spermine Concentration

   ELISA assay was performed to determine the amount of spermine in the
   kidneys or cells by using Mouse spermine ELISA kit (Cat# LV30695,
   Animalunion, Shanghai, China) according to the manufacturer's
   instructions. Briefly, the kidneys or cells were homogenized with PBS
   and centrifuged at 12 000 rpm for 20 min at 4 °C. The supernatant was
   collected to examine the production of spermine, and protein content of
   the supernatant quantified by BCA assay was used to correct spermine
   concentration.^[ [194]^50 ^] Detection of renal tissue spermine by mass
   spectrometry is commissioned by a platform according to standard
   procedures (LipidALL Technologies Co., China). The content of spermine
   was analyzed by Jasper HPLC‐SCIEX 4500 MD. The chromatography
   conditions were as follows: Waters Acquity UPLC SS‐T3 (3×100 mm,
   1.8 µm), A: 0.1% formic acid, mobile phase B: Acetonitrile.

Untargeted Metabolomics Analysis

   mTECs were transfected with either negative control or ATG5 siRNA as
   described above. After exposure to TGF‐β1 (10 ng mL^−1) for 24 h, cells
   were collected for untargeted metabolomics analysis with liquid
   chromatography/mass spectrometer (Q Exactive HF) (Thermo Fisher
   Scientific, USA) via BGI Genomics platform following a standard method
   (BGI, Shenzhen, China). In order to increase the metabolite coverage,
   samples were analyzed under both positive and negative ionization
   modes. The raw metabolome data pretreatment, including peak alignment,
   peak extraction, and compound identification were processed using
   Compound Discoverer 3.1 (Thermo Fisher Scientific, USA), followed by
   metabolite annotation and statistical analysis. Metabolomics data have
   been submitted to MetaboLights
   ([195]www.ebi.ac.uk/metabolights/MTBLS8529).

Transcriptomic Analysis

   RNA isolation, library construction, and sequencing were performed by
   Novogene (Beijing, China). Sequencing was performed on an Illumina
   NovaSeq 6000. The sequencer images were transformed into reads through
   base recognition analysis and clean reads were obtained after raw data
   filtering, error data checking, and GC‐content distribution checking.
   Reference genome and gene model annotation files were downloaded from
   genome website directly. Index of the reference genome was built using
   Hisat2 (v2.0.5) and paired‐end clean reads were aligned to the
   reference genome. Through the Novogene bioinformatics platform,
   differentially expressed genes, KEGG pathway enrichment analysis and
   heatmap analysis were analyzed. For the data generated in this study,
   the RNA sequencing data has been submitted to National Center for
   Biotechnology Information (NCBI) Sequence Read Archive (SRA) database
   with the identifier PRJNA 1015396.

Mass Spectrometry‐Based Protein Identification

   After co‐immunoprecipitation, proteins were separated by SDS‐PAGE,
   extracted from gel slices, and digested with trypsin to extract the
   peptide. The peptide samples were redissolved with mobile phase A (2%
   Acetonitrile, 0.1% formic acid), centrifuged at 20 000 g for 10 min,
   and then the supernatant was taken into the separation by Thermo
   UltiMate 3000 UHPLC. Spectra of peptide fragments were analyzed by
   Q‐Exactive HF X (Thermo Fisher Scientific, USA) and matched to the
   theoretical spectra. After quality control and filtering, protein
   identification results were obtained.

Statistical Analysis

   Statistical processing was performed using GraphPad Prism Software
   (version 8.0). Quantitative values of all experimental data were
   expressed as Mean ± standard error of the mean (SEM). Unpaired
   Student's t‐test (two‐tailed) was conducted for comparison between two
   groups, and one‐way ANOVA with Tukey's test was used for comparison
   between multiple groups. P < 0.05 indicated a statistically significant
   difference.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   D.L., X.H.L., and H.Y.L. contributed equally to this work. D.L. and
   X.H.L. performed major experiments and wrote the original draft. Y.T.W.
   and H.Y.L. performed the experiment design and data analysis. H.Y.L.,
   S.M.J., G.L.L., and Y.P.X. performed the animal experiments and kidney
   histologic examination. K.F.W., X.R.D., and Q.H.L. performed project
   administration and manuscript preparation. W.C., Y.Z., and H.P.M.
   designed the study and revised the manuscript. All authors read and
   approved the final manuscript.

Supporting information

   Supporting Information
   [196]ADVS-11-2306912-s002.docx^ (13.9MB, docx)

   Supporting Information
   [197]ADVS-11-2306912-s001.xlsx^ (17.2KB, xlsx)

Acknowledgements