The downregulation of Insig1 is associated with CKD
To identify the co-occurrences with the fibrosis of PTCs, we performed scRNA-seq with dataset GSE182256 from the GEO database. Following a quality control process, the main kidney cell types were clustered (Supplementary Fig. 1A,B), and the PTCs were reclustered to identify PT subsets based on canonical markers from the original publication. Uniform manifold approximation and projection (UMAP) allowed two-dimensional visualization of the main cell types and the PT subset (Fig. 1A and Supplementary Fig. 1C). The pseudo-time trajectory intriguingly revealed that S1 subset were transformed into S3 and profibrotic PT subgroups (Fig. 1B). A heatmap showed that after UUO, the profibrotic PT subgroup exhibited the highest level of ER stress of all the PT subsets (Fig. 1C). Then, in samples from patients with CKD, we observed an apparent ER expansion and vesiculation in PTCs using an electron microscope and immunofluorescence (IF) staining for the ER marker calreticulin (Fig. 1D). Additionally, PTCs from mice with UUO and 5/6 Nx-induced CKD displayed comparable ER alterations and increased ER stress (Fig. 1E and Supplementary Fig. 1D,E). According to these results, the pathophysiological pathways underlying PT fibrosis involved ER stress.
To identify the genes that may prevent the transition of the S1 subset into the profibrotic PT subset, we applied the BEAM technique to branch point 2. The expression of these genes along with the pseudo-time were shown in the heatmap (Fig. 1F). Then, we employed hdWGCNA to cluster branch-dependent genes. Branch point 2 involved three coexpressed gene modules (Fig. 1G). Three gene modules' eigengene levels from four PT subgroups were shown in the boxplot. We chose the blue module after comparing the gene expression variations between the S1 subset transiting into the S3 subset and the S1 subset transiting into profibrotic PT subset. Compared to those in the S2 subset, the eigengene levels gradually increased in the S1 and S3 subsets but decreased in the profibrotic PT subset (Fig. 1H). There were 120 genes in the blue module, and to identify target genes, we compared these genes with genes that were downregulated in both the GSE118341 dataset and our bulk RNA-seq dataset (Fig. 1I). Next, we combined GO analysis of these 17 similar genes with ER stress related genes. Insig1 was the only gene identified in the ER unfolded protein response (Fig. 1J).
Insig1 was widely expressed in PTCs and with its expression was reduced in the UUO group (Fig. 1K,L and Supplementary Fig. 1F-I). Consistent with its alteration in the blue module, Insig1 was downregulated in kidney biopsies of CKD patients compared to controls (Fig. 1M,N). Table S1 presents a list of the clinical details of CKD patients. It was positively correlated with the estimated glomerular filtration rate (eGFR) and negatively correlated with serum creatinine (SCr) level (Fig. 1O). These results demonstrated that the downregulation of Insig1 in CKD may contribute to the progression of renal fibrosis.
Insig1 deficiency in PTCs aggravated UUO-induced CKD
Next, we produced a PTCs-specific Insig1-knockout (Insig1ΔKap) mice (Fig. 2A). The protein expression in isolated renal tubule tissue confirmed the reduction in Insig1 expression in Insig1ΔKap mice compared to Insig1flox/flox littermates without Cre recombinase activity (Fig. 2B,C). To elucidate the role of Insig1 in CKD renal fibrosis, we subjected Insig1ΔKap and Insig1flox/flox mice to UUO for 14 days to generate classical renal fibrosis model mice. The Insig1ΔKap mice showed worsened collagen and FN deposition in the kidney compared to that in the Insig1flox/flox mice, as determined by Sirius Red staining and Immunohistochemical (IHC) staining (Fig. 2D). Insig1ΔKap mice showed significantly higher expressions of the fibrotic indicators (Fn, α-SMA (Acta2), COL-I, and Postn) in the kidneys after UUO (Fig. 2E, F and Supplementary Fig. 1J), as well as higher expression of genes (Pdgfb, Cd74) representative of the profibrotic PT subset and lower expression of genes (Slc6a13, Slc7a13, Cyp4b1) representative of the S3 subset (Fig. 2G and Supplementary Fig. 1K). Additionally, Insig1 deletion in PTCs increased expressions of ER stress markers (Calreticulin and Atf6) and exacerbated UUO-induced ER expansion and vesiculation (Fig. 2H-J), while exerting no effect on the inflammatory response (Supplementary Fig. 1L). These results revealed that Insig1 deficiency in PTCs aggravated UUO-induced renal fibrosis and ER stress.
Insig1 deficiency in PTCs aggravated 5/6 nephrectomy-induced CKD
Then, we investigated the role of Insig1 in another renal fibrosis models prepared with Insig1ΔKap mice treated with 5/6 Nx for 16 weeks. In accordance with the findings from the UUO-induced CKD model, the kidneys of Insig1ΔKap mice challenged with 5/6 Nx showed higher levels of urine albumin, SCr, and blood urea nitrogen (BUN) levels than Insig1flox/flox mice (Fig. 3A,B). Moreover, compared to the Insig1flox/flox mice, Insig1ΔKap mice exhibited exacerbated collagen and FN deposition rates and increased expressions of the fibrotic indicators (Fn, Col-I and Col-III) after 5/6 Nx challenge (Fig. 3C, D). Additionally, after 5/6 Nx challenge, Insig1 deficiency in PTCs exacerbated ER expansion and vesiculation and increased expressions of ER stress markers (Calreticulin, Atf6 and Trib3), while decreasing Slc6a13 expression and increasing Cd74 expression (Fig. 3E-H). These results revealed that Insig1 deficiency in PTCs aggravated 5/6 Nx-induced renal fibrosis and ER stress.
Insig1 overexpression alleviated UUO-induced renal fibrosis in mice
In addition to loss-of-function models, gain-of-function model was established by injecting Insig1 and Ctrl plasmids under high pressure into the mouse tail vein. Supplementary Fig. 2A demonstrates that 36 hours after injecting the Insig1 plasmids, the protein expression of Insig1 was significantly increased in the kidneys. These mice underwent UUO surgery 36 hours after receiving the Insig1 plasmids. In contrast to the findings with Insig1ΔKap mice, injecting the Insig1 plasmids markedly improved renal fibrosis in the UUO-treated mice compared to that in the mice injected with Ctrl plasmids, as evidenced by reduced collagen deposition and fibrotic indicator (COL-I and α-SMA) protein expression, as well as similar changes in the mRNA expression of Fn, Col-I, and Col-III in the kidneys (Supplementary Fig. 2B-D). These results considerably demonstrated Insig1's protective function in renal fibrosis, which was in line with the beneficial role in the blue module.
Silencing Insig1 aggravated TGF-β1-induced fibrotic responses in vitro
To examine the direct effect of Insig1 on renal tubular cells, we transfected HK2 cells with Insig1 siRNA or Ctrl siRNA (Supplementary Fig. 2E) and then treated the cells with TGF-β1 for 24 h. In accordance with our in vivo findings, silencing Insig1 worsened TGF-β1-induced fibrotic responses, as shown by increased expression of Fn and Acta2, as well as worsened ER stress compared to that of the control group (Supplementary Fig. 2F,G). We also found that silencing Insig1 boosted TGF-β1-induced expression of FN, CTGF, Acta2 and Col-I in mouse renal tubular epithelial cells (TKPTS) (Supplementary Fig. 3A-E). Therefore, we hypothesized that silencing Insig1 exacerbated TGF-β1-induced fibrotic responses in HK2 and TKPTS cells.
In contrast, overexpression of Insig1 reduced TGF-1-induced expression of FN, CTGF, Acta2, and Col-I compared to that of the control group (Supplementary Fig. 3F–J). Collectively, these findings showed that Insig1 significantly inhibited the synthesis of extracellular matrix components and ER stress caused by TGF-β1, showing that Insig1 functions as a protective factor against renal fibrogenesis.
Aldh1a1 is a transcriptional targert of Insig1 in the kidney
As Insig1 regulates lipid metabolism, we first examined whether Insig1 deletion exacerbates UUO-induced renal fibrosis mediated through its effect on lipid metabolism. We observed that UUO-induced lipogenesis was unaffected by Insig1 deletion (Supplementary Fig. 4A). Additionally, TOFA (an inhibitor of acetyl-CoA carboxylase-α) and C75 (an inhibitor of fatty-acid) were applied to an in vitro CKD model. When TGF-β1 was present, silencing Insig1 enhanced the mRNA expression of Fn and Ctgf, and neither TOFA nor C75 reversed this trend (Supplementary Fig. 4B). These findings suggested that lipogenic activity may not be related to the profibrotic effect of Insig1 deletion in PTCs.
Then, utilizing isolated renal tubule tissue, we conducted a bulk RNA-seq analysis to examine differences in gene expression between Insig1flox/flox and Insig1ΔKap mouse PTCs to better understand the underlying mechanism by which Insig1 prevents renal fibrosis. Insig1ΔKap mice showed 179 differentially expressed genes, 109 downregulated genes and 70 upregulated genes, in comparison to Insig1flox/flox mice (Fig. 4A). The retinol metabolism pathway was one of the most significantly altered pathways, according to KEGG pathway enrichment results (Fig. 4B). Aldh1a1, a crucial enzyme in retinol metabolism, was markedly increased in the kidneys of Insig1ΔKap mice in contrast to Insig1flox/flox mice (Fig. 4C). Aldh1a1 was also noticeably elevated in Insig1-silenced TKPTS cells (Fig. 4D). Additionally, the luciferase reporter gene data showed that Insig1 negatively modulated Aldh1a1 transcriptional activity (Fig. 4E,F).
According to the literature, after deletion, Insig1 was unavailable to interact with SCAP, causing the SCAP-SREBP1 complex to translocate to the Golgi apparatus and increasing the amount of SREBP1 entering the nucleus 8. Moreover, SREBP1 protein expression was markedly decreased in the cytoplasm but increased in the nucleus of Insig1-silenced TKPTS cells (Fig. 4G). Aldh1a1 transcriptional activity was significantly elevated when SREBP1 was overexpressed (Fig. 4H). A chromatin immunoprecipitation (ChIP) assay revealed that SREBP1 bound to the Aldh1a1 promoter region (Fig. 4I). Additionally, fatostatin (an SREBP1 inhibitor) markedly reversed the increase in transcriptional activity of Aldh1a1 in Insig1-silenced TKPTS cells (Fig. 4J). Although the active form of SREBP1 (SREBP1 1-480 bp) increased Aldh1a1 transcriptional activity, fatostatin was failed to reverse this trend (Fig. 4K). These findings demonstrated that Insig1 prevented SREBP1 protein entry into the nucleus, thereby lowering Aldh1a1 transcriptional activity (Fig. 4L).
Inhibiting Aldh1a1 alleviated renal fibrosis in vivo and in vitro
The Aldh1a1 level was found to be significantly higher in the UUO model mice (Fig. 5A and Supplementary Fig. 4C–E), and it positively correlated with the SCr level and negatively correlated with the eGFR (Fig. 5B). To verify the role of Aldh1a1 in CKD, we subsequently constructed Aldh1a1-KO mice (Supplementary Fig. 4F,G). In contrast to wild-type (WT) mice, Aldh1a1 deletion markedly decreased collagen deposition and FN expression in the kidneys after UUO (Fig. 5C). The fibrotic index (α-SMA, COL-I, Col-III, Postn, and Ctgf) and the profibrotic PT maker (Pdgfb) level were also significantly lower in the Aldh1a1-KO animals after UUO, while the S3 marker (Slc6a13, Slc7a13 and Cyp4b1) levels were significantly higher (Fig. 5D–F and Supplementary Fig. 4H). Moreover, Aldh1a1 deletion significantly decreased ER dilation and ER stress makers (Calreticulin, Atf6 and Perk) (Fig. 5G–I).
In TKPTS cells, overexpression of Aldh1a1 (Supplementary Fig. 4I) worsened TGF-β1-induced fibrotic responses, as shown by increased expression of FN, Acta2 and Ctgf (Fig. 5J,K). In contrast, Aldh1a1-silencing reduced ER stress and the TGF-β1-induced expression of FN, α-SMA and Col-I in TKPTS cells compare to the levels in the control group (Fig. 5L and Supplementary Fig. 4J–L). Additionally, modifying the protein's active site may be able to block Aldh1a1's profibrotic effects (Fig. 5M).
In line with these findings, pharmacological intervention using the Aldh1a1 inhibitor NCT501 reduced collagen deposition and the expression of FN, Acta2, and COL-I in kidneys after UUO (Supplementary Fig. 5A–C). NCT501 also reduced the veracity of TGF-β1-induced fibrotic responses in TKPTS and HK2 cells (Supplementary Fig. 5D-F). These results together demonstrated that Aldh1a1 overexpression exacerbated renal fibrosis.
Insig1/Aldh1a1 activation mitigated renal fibrosis by increasing NAD+ levels
An untargeted metabolomics analysis was carried out to determine the mechanism of Aldh1a1 in renal fibrosis. As expected, we discovered a considerable decrease in the levels of NAD+ and its precursors (nicotinamide riboside and nicotinamide ribotide) in kidneys after UUO (Supplementary Fig. 5G–I), and this decline was even more pronounced following the loss of Insig1 in PTCs (Supplementary Fig. 5J, K and Fig. 6A,B). Consistent with the metabolomic analysis results, after UUO or 5/6 Nx challenge, the NAD+ concentration was considerably lower in the kidneys of the Insig1ΔKap mice than in the kidneys of the Insig1flox/flox mice (Fig. 6C). Additionally, compared to WT mice, Aldh1a1-KO mice showed significantly higher levels of NAD+ in the kidneys after UUO (Fig. 6C). NCT501 treatment increased the NAD+/NADH ratio in TKPTS cells despite Insig1 silencing or Aldh1a1 overexpression decreasing it. Additionally, Aldh1a1 overexpression worsened TGF-β1-induced NAD+ consumption (Fig. 6D,E). In hiPSC-derived kidney organoids, we also found that NAD+ reduced TGF-β1-induced extracellular matrix component production (Supplementary Fig. 5L). These findings suggested that the antifibrotic activity of the Insig1/Aldh1a1 pathway may depend on NAD+ homeostasis.
Recent research suggested that NAD+ controls ER membrane expansion by deribosylating the ER sensor GRP78/BIP 28 (Fig. 6F). In TKPTS cells, NCT501 treatment or Insig1 overexpression significantly increased ADP-ribosylated BIP levels (Fig. 6G,H). Insig1 silencing also resulted in an expansion of the ER membrane and reduced IRE1α and BIP binding. NAD+ or Aldh1a1 silencing, however, increased BIP and IRE1α binding (Fig. 6I,J and Supplementary Fig. 5M). Together, Insig1 overexpression or Aldh1a1 silencing restored the ADP ribosylation of BIP, enhanced the binding of IRE1α and BIP, and maintained ER homeostasis by increasing NAD+ levels.
Identification of a novel human Aldh1a1 inhibitor that alleviated UUO-induced CKD in a Aldh1a1-dependent manner
The TargetMol-Approved Drug Screening Library (a total of 1600 compounds) and TargetMol-Anti Cancer Compound Library (a total of 3145 compounds) were used in our study for virtual screening (Fig. 7A). On the basis of the virtual screening results, nineteen compounds were chosen for additional examination by in vitro enzymatic activity assays. The human recombinant Aldh1a1 (hAldh1a1) protein was profoundly inhibited by nicardipine (13-F9), and its docking score was − 9.4128 (Fig. 7B,C). Then the BIAcore system was used to directly evaluate the binding of nicardipine and Aldh1a1. The equilibrium dissociation constant (KD) value of Aldh1a1 with nicardipine was 1.98E-06 M, indicating strong binding between nicardipine and Aldh1a1, while the KD value of Aldh1a1 with NCT501 was 3.31E-05 M (Fig. 7D and Supplementary Fig. 5N). In vivo, nicardipine treatment decreased Aldh1a1 protein expression in the kidney but exerted no effect on Aldh1a1 mRNA expression (Fig. 7E and Supplementary Fig. 6A). Consistent with in vivo results, nicardipine lowered Aldh1a1 protein expression in TKPTS cells in a dose-dependent manner but exerted no effect on Aldh1a1 mRNA expression or transcriptional activity (Fig. 7F and Supplementary Fig. 6B,C). Additionally, the protease inhibitor MG132 profoundly reversed the nicardipine-induced reduction in Aldh1a1 protein expression (Supplementary Fig. 6D). Therefore, nicardipine inhibited Aldh1a1 expression via a proteasomal degradation mechanism. Thus, nicardipine may function as a novel Aldh1a1 inhibitor.
For further analysis of the effects of nicardipine on UUO-induced CKD, mice were given nicardipine (1 mg/kg/day intragastrically (i.g.)) for 2 days before UUO surgery. Nicardipine administration decreased renal fibrosis in the mice after UUO, which was consistent with the outcomes observed in mice treated with NCT501. Following nicardipine treatment, collagen deposition and FN expression were reduced in the mice after UUO (Fig. 7G), and profibrotic marker expression (FN, CTGF, Col-I, and Col-III), ER stress index (Atf6, Trib3), and ER expansion, were decreased in the kidneys compared to UUO-only exposed animals (Fig. 7H–K). Similarly, nicardipine markedly reduced the fibrotic reactions caused by TGF-β1 in TKPTS cells, HPTCs, and hiPSC-derived kidney organoids (Supplementary Fig. 7A–F). Mechanistically, nicardipine greatly elevated the NAD+ concentration and the NAD+/NADH ratio, which had been decreased by TGF-β1, significantly boosting ADP-ribosylated BIP levels and increasing IRE1α and BIP binding, which prevented ER expansion (Fig. 7L–N).
To determine whether the protective effect of nicardipine on UUO-induced CKD involved an Alhd1a1-dependent mechanism, we administered nicardipine to WT and Aldh1a1-KO mice for 14 days. As expected, the protective effects of nicardipine on UUO-induced CKD were abolished in Aldh1a1-KO mice (Fig. 7O–Q), showing that Aldh1a1 is involved in the effect of nicardipine on UUO–induced CKD. According to these findings, nicardipine might be a new highly targeted Aldh1a1 inhibitor.