Establishment of selective ablation of Atg5 in kidney proximal tubular mice.

To further investigate the exact role of autophagy in the renal tubules, we established conditional deletion of Atg5 to selectively ablate autophagy in proximal tubules. We crossed Atg5-floxed (Atg5flox/flox) mice with the PEPCK-Cre mice, establishing renal proximal tubule-specific atg5 knockout (PT-atg5 KO) mice. WT littermates were used as controls. The breeding protocol is illustrated in Supplementary Figure S1A. PCR genotyping was performed for each mouse, with the PT-atg5 KO genotype confirmed by (1) amplification of the 700-bp fragment of the floxed allele, (2) lack of amplification of the 350-bp fragment of the WT allele, and (3) amplification of the 370-bp fragment of the Cre gene (Supplementary Figure S1B). Western blot analysis revealed significantly reduced ATG5 expression in PT-atg5 KO mice compared to WT littermates, as well as markedly decreased LC3-I to LC3-II conversion in the kidney cortex and outer medulla (Fig. 1A, B). Furthermore, western blot analysis demonstrated the accumulation of p62, a selective substrate of autophagy, which was markedly higher in PT-atg5 KO kidney tissues than in WT (Fig. 1A, B). These findings confirm the specific deletion of Atg5 in renal proximal tubular epithelial cells in the PT-atg5 KO model. Without surgical treatment, these mice exhibited comparable blood urea nitrogen (BUN) and serum creatinine levels to WT (Fig. 1C), indicating preserved renal function.

Fig. 1

Atg5 deletion in proximal tubules exacerbates acute renal injury but reduces fibrosis during the recovery phase after I/R. WT and PT-atg5 KO mice were subjected to a sham operation or bilateral renal ischemia followed by reperfusion for up to 4 weeks (n = 5–10 mice per group), and the kidneys were collected at the indicated time points. A, B Representative immunoblots and quantification of ATG5, p62, and LC3 expression in the kidneys (n = 3). C Serum creatinine and BUN levels (n = 10). D Representative images of H&E staining, Masson’s trichrome and Sirius red staining. Scale bars, 50 μm. E, F Quantitative analysis of Masson’s trichrome and Sirius Red staining (n = 6). The values are expressed as mean ± standard deviation (SD), * represents a significant difference from the sham groups, and # represents a significant difference from the relevant wild-type group. # or * P < 0.05, ## or ** P < 0.01, ### or *** P < 0.001, #### or **** P < 0.0001

Atg5 deficiency in proximal tubules aggravates renal injury in the acute phase and attenuates renal fibrosis in the recovery period after AKI

AKI plays an important role in the onset and progression of CKD. The I/R model has been employed to investigate the mechanisms underlying the transformation from AKI to CKD. To analyze the impact of autophagy in the proximal tubules after ischemia-induced AKI, we first examined autophagy at the specified time points. As shown in Fig. 1A, B, WT kidney tissues showed continuous activation of tubular autophagy, whereas PT-atg5 KO kidney tissues showed significantly reduced ATG5 expression, a decreased LC3II/I ratio, and an accumulation of p62 compared with WT littermate mice, suggesting that the activation of tubular autophagy was nearly abolished in PT-atg5 KO mice.

Subsequently, we examined the effects of autophagy deficiency on kidney function and morphology during ischemic AKI in PT-atg5 KO mice. At 24 h post-renal ischemia/reperfusion, the deletion of Atg5 in proximal tubular significantly worsened renal function compared to WT littermates, exhibiting severe brush edge loss and tubular dilation, which resulted in BUN and serum creatinine levels of 68.38 mmol/L and 172.98 μmol/L, respectively (Fig. 1C). In contrast, WT littermate mice demonstrated BUN levels of 59.26 mmol/L and serum creatinine levels of 144.24 μmol/L. At 48 h after ischemia/reperfusion, serum creatinine and BUN levels in both groups increased, but no significant differences were observed compared to the levels at 24 h (Supplementary Figure S2A). At day 4 post renal ischemia/reperfusion, PT-atg5 KO mice had 58.12 mmol/L BUN and 104.50 μmol/L serum creatinine, while WT littermate mice had 39.98 mmol/L BUN and 83.71 μmol/L serum creatinine (significantly higher than the levels of the WT). Furthermore, renal histological examination revealed more severe tubule injury in PT-atg5 KO mice compared with WT littermates (Fig. 1D, Supplementary Figure S2B). At 7 days or 2 weeks after AKI, serum creatinine and BUN levels in PT-atg5 KO were closely comparable to those in WT mice. However, at 4 weeks after AKI, WT mice still exhibited 14.99 mmol/L BUN and 28.96 μmol/L serum creatinine level, whereas PT-atg5 KO mice had 12.90 mmol/L BUN and 21.93 μmol/L serum creatinine level, indicating better renal recovery (Fig. 1C). Consistently, Masson’s trichrome and Sirius red staining revealed decreased collagen deposition in PT-atg5 KO (Fig. 1D–F). PT-atg5 KO mice also had less α-SMA (Fig. 2A, B), VIMENTIN (Fig. 2A, C), PDGFR-β (Fig. 2A, D), FN1 (Fig. 2E, F), and collagen I (Fig. 2G) than control group at day 28 after I/R. Collectively, the results indicate that Atg5 deletion in proximal tubule aggravates renal injury during the acute phase while attenuating renal fibrosis during the recovery period following AKI.

Fig. 2

Atg5 deficiency in proximal tubules suppresses renal interstitial fibrosis after I/R injury. Kidneys were collected from WT and PT- atg5 KO mice after sham surgery or 4 weeks after bilateral renal ischemia/reperfusion. A Representative α-SMA (red, left), VIMENTIN (red, middle), and PDGFR-β (red, right) immunofluorescence staining images of kidney sections. The nuclei were stained with DAPI (blue). LTL was used to label proximal tubules (green). Scale bars, 20 μm. B–D Quantitative analysis of α-SMA, VIMENTIN, and PDGFR-β-positive area (n = 6). E, F The kidney sections were immunostained with antibody against FN1. The FN1-positive areas were quantified to view areas (n = 6). Scale bars, 50 μm. G mRNA levels of collagen I in kidney tissues, as detected by qPCR (n = 3). The values are expressed as mean ± SD. * represents a significant difference from the sham group; # represents a significant difference from the relevant wild-type group. # or * P < 0.05, ## or ** P < 0.01, ### or *** P < 0.001, #### or **** P < 0.0001

Atg5 deletion in proximal tubules changes the ultrastructure of proximal tubules of the ischemic kidney

Kidneys were collected at 1 and 28 days post-ischemia/reperfusion for ultrastructural analysis to investigate the impact of Atg5 on the tubule cell stress response. The results showed that proximal tubule epithelial cells exhibited variable sizes and deformed mitochondria 1 day after I/R. Mitochondria from the PT-atg5 KO group showed more pronounced cristae fragmentation and effacement, along with lysosomal storage, compared with those from the WT group (Fig. 3A). Tubular epithelial cells of the WT group contained autophagosomes and autolysosomes filled with mitochondria, whereas concentric membrane bodies were detected around damaged mitochondria in Atg5 deficient tubule cells (Fig. 3A). After 28 days, the mitochondria in both groups continued to exhibit mild cristae disruption, with collagen deposition observed between the cells (Fig. 3B). Autophagosomes and autolysosomes were still present in WT mice; in the PT-atg5 KO group, concentric membrane bodies remained visible, but lysosomal accumulation was no longer observed.

Fig. 3

Proximal tubule Atg5 deletion changes the ultrastructure of the proximal tubule in ischemic kidney injury. Electron micrographs of proximal tubule cells (A) 1 day after reflow and (B) 28 days after reflow following ischemia. At 1 day of I/R, the mitochondria from the PT-atg5 KO group showed more severe cristae fragmentation and lysosomal storage than those from the WT group. Mitochondria-filled autophagosomes and autolysosomes were present in WT tubule cells, whereas concentric membranes were detected in Atg5 deficient tubule cells. After 28 days, the mitochondria showed slight ridge fragmentation with intercellular collagen deposition. Autophagosomes and autolysosomes were still present in the WT group, and concentric membrane bodies were still detected in the PT-atg5 KO group; however, lysosomal accumulation was no longer observed. The arrows indicate double-membrane autophagosomes and autolysosomes with a single membrane. Arrowheads indicate concentric membranes, asterisks indicate collagen fibers, mitochondria are labeled with “M,” and lysosomes are labeled with “L.” Scale bars, 1 μm

Characteristics of injury and repair in tubular autophagy-deficient mice after ischemic AKI

To comprehensively investigate the effect of tubular Atg5 deficiency on injury and repair mechanisms during AKI following I/R, we meticulously monitored the levels of kidney injury molecule-1 (KIM-1), a biomarker of renal tubular injury and proliferating cell nuclear antigen (PCNA), an indicator of cell proliferation. Immunofluorescent staining revealed that proximal tubule injury and cell proliferation notably increased in PT-atg5 KO mice on day 4 following I/R injury compared to their WT counterparts (Fig. 4A, B), implying that the absence of tubular autophagy amplifies tubular damage and repair simultaneously during the acute injury stage. KIM-1 levels in WT kidneys continued to increase during the AKI repair process, peaking on day 14 (Fig. 4A, C), which may indicate a risk of progression from AKI to CKD [22,23,24]. In contrast, KIM-1 levels in PT-atg5 KO mice remained stable on days 7, 14, and 28 after AKI, significantly lower than that in the WT group (Fig. 4A, C). Similar results were observed in urinary KIM-1 levels (Supplementary Figure S3A), further indicating that tubular Atg5 worsened tubular damage during kidney recovery from AKI. Nevertheless, tubular Atg5 seemed to be ineffective in cell proliferation during the repair stage, with no difference in the number of proliferating cells between the WT and KO groups at 7, 14, or 28 days after AKI. (Fig. 4B, D).

Fig. 4

Dynamic changes of KIM-1 and PCNA in Atg5-deficient proximal renal tubules after I/R injury. WT and PT-atg5 KO mice kidneys were harvested at 1, 4, 7, 14, and 28 days after sham surgery or bilateral ischemia/reperfusion, as described in Fig. 1, to observe the dynamic course of kidney injury and repair. A Immunofluorescence staining of KIM-1 (red) in the kidney sections. B Representative PCNA (red) immunofluorescence staining of kidney sections. Kidney tissue sections were stained with DAPI to visualize cell nuclei (blue) and LTL to visualize proximal tubules (green). Scale bars, 20 μm. C Quantitative analysis of the KIM-1 positive stained areas. D Quantification of PCNA-positive cells per mm2. The dates are shown as mean ± SD (n = 6). * Indicates significant difference from the sham group; # indicates a significant difference from the relevant wild-type group. # or *P < 0.05, ## or **P < 0.01, ### or ***P < 0.001, #### or ****P < 0.0001

Oxidative stress is significantly relevant to renal injuries, primarily during reperfusion [25]. To investigate the effect of renal tubular Atg5 on oxidative stress during AKI progression, immunofluorescence was performed to detect the formation of 4-hydroxynonenal (4HNE) adduct, one of the most commonly used biomarkers for oxidative stress. A marked increase in tubular 4HNE adduct formation was observed on days 1 and 4 after I/R, further exacerbated by Atg5 deletion (Fig. 5A, B). Fourteen and 28 days after I/R, there was no difference in the protein adduct formation of 4HNE between the WT and PT-atg5 KO groups, with levels reduced those similar to the sham group (Fig. 5A, B), suggesting Atg5 depletion in renal proximal tubules exacerbated tubular oxidative stress in the early phase of AKI following I/R, potentially contributing to the more severe kidney injury without autophagy.

Fig. 5

Atg5 deficiency promotes 4HNE production in renal tubules in the acute injury phase of AKI. WT and PT-atg5 KO mice were subjected to a sham operation or bilateral renal ischemia/reperfusion, and the kidneys were collected at the indicated time points for immunofluorescence analysis of 4HNE, a biomarker of oxidative stress. A Representative immunofluorescence staining images with the 4HNE antibody (red). The cell nuclei were labeled with DAPI (blue). Proximal renal tubules were stained with LTL (green). Scale bars, 20 μm. B Quantification of the 4HNE positive stained areas (n = 6). The values are expressed as mean ± SD, * Indicates a significant difference from the sham groups; # indicates a significant difference from the relevant wild-type group. # or *P < 0.05, ## or **P < 0.01, ### or ***P < 0.001, #### or ****P < 0.0001

Atg5 depletion in renal proximal tubules inhibits the production of tubular profibrotic factors after I/R

Damaged tubular cells have the capacity to produce various growth factors, including FGF2, platelet-derived growth factor beta (PDGF-β), connective tissue growth factor (CTGF), and transforming growth factor beta 1(TGFβ1). These factors are well-documented for their role in promoting fibrosis formation during kidney recovery from AKI [18, 26,27,28,29,30]. The levels of theses factors were dynamically assessed during AKI progression. Tubular FGF2 levels in both groups increased rapidly after I/R, peaked on day 4, and then decreased and returned to baseline levels on day 28. Notably, during AKI progression, FGF2 production in the PT-atg5 KO group was consistently significantly lower than that in WT kidneys (Fig. 6A, B), suggesting that Atg5 deletion in tubule cells inhibited the production of tubular FGF2 after I/R. Additionally, the levels of PDGF-β (data not shown), TGFβ1, and CTGF also significantly increased after I/R and remained elevated throughout the AKI phase (Supplementary Figure S4A–D). However, there were no differences in PDGF-β (data not shown), TGFβ1 and CTGF levels between the WT and PT-atg5 KO groups during the development of AKI (Supplementary Figure S4A–D), indicating that the absence of Atg5 in renal tubules did not affect the production of renal PDGF-β, TGFβ1, and CTGF after I/R.

Fig. 6

Atg5 deletion inhibits FGF2 production in renal tubules after I/R. The kidneys of WT and PT-atg5 KO mice were harvested at the indicated time points after sham surgery or bilateral I/R injury to detect the levels of the profibrotic factor FGF2. A Kidney slices were immunostained with antibody against FGF2 (red), DAPI to visualize the cell nuclei (blue), and LTL to visualize the proximal tubules (green). Scale bars, 20 μm. B Quantitative analysis of FGF2 positive areas. The dates are expressed as mean ± SD (n = 6). * represents a significant difference from the sham group; # represents a significant difference from the relevant wild-type group. # or *P < 0.05, ## or **P < 0.01, ### or ***P < 0.001, #### or ****P < 0.0001

These results suggest that Atg5 in renal proximal tubules is involved in regulating the production of profibrotic factors FGF2 after I/R, but not PDGF-β, TGFβ1, and CTGF.

Atg5 in renal proximal tubules regulates fibrosis in the UUO model

To further verify the effect of Atg5 in renal proximal tubules on renal fibrosis, we established a UUO model by ligating the left ureter, a classic model for studying renal interstitial fibrosis. On day 14 post-surgery, the expression levels of ATG5 and LC3B in the PT-atg5 KO group were significantly lower than that in the WT group, while P62 accumulated in large amounts (Fig. 7A), indicating that successful elimination of the proximal tubule Atg5 resulted in an impaired ability to induce autophagy in PT-atg5 KO mice during UUO. The kidneys of the WT mice exhibited more severe structural disorders, characterized by tubular atrophy, intratubular cast formation, and inflammatory cell infiltration (Fig. 7B). Correspondingly, the KO group exhibited reduced collagen deposition (Fig. 7B–D) and lower levels of VIMENTIN, PDGFR-β, and α-SMA (Fig. 7E–H), primarily concentrated in the interstitial region, suggesting that tubule Atg5 loss inhibited the formation of interstitial fibrosis during UUO. We also examined the induction of tubular profibrotic factors in the UUO model. Atg5 deletion in proximal renal tubules inhibited the production of FGF2 (Fig. 8A, B), while PDGF-β (data not shown), CTGF (Supplementary Fig. 5A, C), and TGFβ1 (Supplementary Figure S5B, D) levels remained unaffected. These findings suggest that proximal tubule Atg5 may exacerbate renal fibrosis in the UUO model by promoting the production of the profibrotic factor FGF2.

Fig. 7

Atg5 deficiency in renal proximal tubules inhibited interstitial fibrosis during UUO. WT and PT-atg5 KO mice were subjected to sham operation or unilateral ureteral obstruction (UUO) (n = 6 each group), and the kidneys were harvested 14 days later. A Representative immunoblots and quantification of ATG5, p62, and LC3 expression in the kidneys (n = 3). B Representative images of H&E staining, Masson’s trichrome staining, and Sirius red staining. Scale bars, 50 μm. C, D Quantitative analysis of collagen deposition. E Kidney slices were immune stained with antibodies against α-SMA (red, left), VIMENTIN (red, middle), and PDGFR-β (red, right), labeled by DAPI to visualize cell nuclei (blue) and LTL to visualize proximal tubules (green). Scale bars, 20 μm. F–H Quantitative analysis of α-SMA, VIMENTIN, and PDGFR-β positive areas. The values are expressed as mean ± SD (n = 6). * Indicates a significant difference from the sham group; # indicates a significant difference from the relevant wild-type group. # or * P < 0.05, ## or **P < 0.01, ### or ***P < 0.001, #### or ****P < 0.0001

Fig. 8

Atg5 deficiency suppresses FGF2 production in renal tubules during UUO. Kidneys subjected to sham or UUO surgery were collected on day 14, as shown in Fig. 7. A Representative images of immunofluorescence staining images with FGF2 antibody (red). The cell nuclei were labeled with DAPI (blue). Proximal renal tubules were stained with LTL (green). Scale bars, 20 μm. B Quantification of FGF2 positive stained areas. The values are expressed as mean ± SD (n = 6). * Indicates a significant difference from the sham group; # indicates a significant difference from the relevant wild-type group. # or *P < 0.05, ## or ** P < 0.01, ### or ***P < 0.001, #### or ****P < 0.0001

Development of AKI and CKD from patients correlated with Atg5, oxidative stress, and profibrotic factors FGF2

We sought to verify whether Atg5 plays a similar role in regulating AKI and interstitial fibrosis in human samples. Our study included six patients with minimal change disease (MCD) as the control group, six cases of AKI, and six cases of CKD. Compared with MCD, both AKI and CKD exhibited more severe structural disorders, tubular damage, and increased collagen deposition, especially in CKD, as confirmed by Masson and PASM staining (Fig. 9A–C). Immunofluorescence results revealed persistent activation of ATG5 during AKI and CKD in human samples (Fig. 9D, F). Unlike the findings in the mouse model, the production of the oxidative stress marker 4HNE adduct increased not only during AKI but also during CKD (Fig. 9D, G), and the area co-stained by ATG5 and 4HNE increased significantly in AKI and CKD (Fig. 9D, I). The profibrotic factor FGF2 was found to be elevated exclusively in the renal tubules of patients with CKD (Fig. 9E, H). Similarly, the area co-stained with ATG5 and FGF2 significantly increased in the CKD group (Fig. 9E, J). Biopsy analyses of these patients showed that renal tubule autophagy, oxidative stress, and the profibrotic factor FGF2 were involved in the development of AKI and CKD.

Fig. 9

Increased induction of renal tubule autophagy, 4HNE, and FGF2 in renal biopsies in patients with AKI and CKD. Renal biopsies from patients with minimal change disease (MCD) as controls (n = 6), patients with AKI (n = 6), and patients with CKD (n = 6) were evaluated. A Representative images of H&E staining, Masson’s trichrome staining and PASM staining. Scale bars, 50 μm. B, C Quantitative analysis of Masson’s trichrome and PASM staining. D 4HNE (red) and ATG5 (violet) cells. E FGF2 (red) and ATG5 (violet) double immunofluorescence staining of the kidney sections. Cell nuclei were counterstained with DAPI (blue). Proximal renal tubules were stained with LTL (green). Scale bars, 20 μm. F Quantification of ATG5 positive area fraction. G Quantification of the 4HNE positive stained areas. H Quantification of FGF2 positive stained areas. I Quantification of 4HNE and ATG5 co-stained areas. J Quantification of FGF2 and ATG5 co-positive areas. The values are expressed as mean ± SD, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001