GSTK1 and RETREG1/FAM134B-mediated reticulophagy attenuates tubular injury in diabetic nephropathy through endoplasmic reticulum stress and apoptosis

ABSTRACT

Reticulophagy is a key process to recovery from endoplasmic reticulum (ER) stress and for maintaining ER homeostasis by selectively removing damaged ER and its components. However, its precise mechanisms in diabetic nephropathy (DN) remain unclear. Here, we found that the expression of RETREG1/FAM134B (reticulophagy regulator 1) was decreased in the tubular cells in DN patients and animal models, which was positively correlated with estimated glomerular filtration rate (eGFR) and negatively associated with tubulointerstitial damage. Proximal tubule-specific knockout of Retreg1 exacerbated reticulophagy abnormalities in diabetic mice induced by high-fat diet (HFD) combined with streptozotocin (STZ), which was accompanied by increased ER stress, apoptosis of tubular cells and tubulointerstitial fibrosis. In vitro, overexpression of RETREG1 notably restored reticulophagy, and alleviated ER stress and apoptosis in HK-2 cells, a human proximal tubular cell line, treated with high glucose. Mechanistically, immunoprecipitation coupled with mass spectrometry (IP-MS) suggested that RETREG1 could interact with GSTK1 (glutathione s-transferase kappa 1). Silencing of GSTK1 further aggravated the reduction of reticulophagy and tubular injury both in vivo and in vitro. These effects in in vitro were partially blocked by overexpressing RETREG1. Collectively, these findings suggest that GSTK1 and RETREG1 exert a protective role in tubular injury through restoring reticulophagy and mitigating ER stress of tubular cells in DN.

Abbreviation: ACTB: actin beta; cCASP3: cleaved caspase 3; CANX: calnexin; CASP: caspase; Co-IP: co-immunoprecipitation; DDIT3: DNA damage-inducible transcript 3; DN: diabetic nephropathy; ER: endoplasmic reticulum; FN1: fibronectin 1; GSTK1: glutathione S-transferase kappa 1; HFD: high-fat diet; HG: high glucose; HK-2: human tubular cell; HSPA5: heat shock protein family A (Hsp70) member 5; IHC: immunohistochemistry; IF: immunofluorescence; IP MS: immunoprecipitation coupled with mass spectrometry; LIR: LC3-interacting region; LTL: Lotus tetragonolobus lectin; MAP1LC3/LC3: microtubule associated protein 1 light chain 3; PACS2: phosphofurin acidic cluster sorting protein 2; PTCs: proximal tubular cells; PT: proximal tubule; RETREG1/FAM134B: reticulophagy regulator 1; RHD: reticulon homology domain; RT-qPCR: real time-quantitative PCR; SQSTM1/p62: sequestosome 1; STZ: streptozotocin; TECs: tubular epithelial cells; TEM: transmission electron microscopy; TUNEL: terminal deoxynucleotidyl transferase dUTP nick-end labeling; UACR: urine albumin creatine ratio; UPR: unfolded protein response.

Introduction

Diabetic nephropathy (DN) is the leading cause of end-stage renal disease worldwide [1,2]. In recent years, accumulating evidence has shown that tubular epithelial cells (TECs) injury is an early and initial feature of DN that might be independent of glomerulopathy, which plays a key role in the pathogenesis of DN [3–5]. The mechanism of DN tubular injury is complex. It was suggested that high glucose (HG), advanced glycation end products, and hemodynamics lead to increased mitochondrial reactive oxygen species/ROS production, endoplasmic reticulum (ER) stress, and decreased macroautophagy/autophagy, ultimately leading to TECs injury [6–8]. Among these, ER stress is well recognized as a key pathogenesis leading to TECs injury and DN progression [9,10].

ER stress is activated when misfolded and unfolded proteins accumulate in the ER [11], and sustained ER stress ultimately leads to cell apoptosis [12,13]. Several studies have demonstrated the pivotal role of ER stress in tubular apoptosis and tubulointerstitial injury under DN conditions [9,14–16]. It was shown that AdipoRon can inhibit ER stress through ADIPOR1-p-PRKAA/AMPK and ameliorate tubular injury in DN [9]. Dapagliflozin, a medication that belongs to a class of drugs called SLC5A2/sodium-glucose co-transporter 2 inhibitors, was suggested to protect tubular cell apoptosis partially through decreased ER stress in db/db mice [16]. In addition, the expression of unfolded protein response (UPR) genes, such as HSPA5 (heat shock protein family A (Hsp70) member 5) and XBP1 (X-box binding protein 1) are increased in renal biopsies of DN patients [17]. Thus, targeting ER stress might be a crucial therapeutic strategy to alleviate tubular injury in DN.

To recover from ER stress and maintain ER homeostasis, signaling system including the UPR, ER-associated degradation/ERAD, and the recently discovered reticulophagy are activated in eukaryotic cells [10,18]. Reticulophagy is a selective autophagy process mediated by specific reticulophagy receptors, which removes damaged ER subdomains and abnormally clustered ER proteins to maintain ER homeostasis [19–21]. Emerging research has shown that reticulophagy dysfunction plays a critical role in the pathogenesis of cancer [22], neurodegenerative diseases [23], and metabolic diseases [20,24]. Notably, it was shown that the reduction in reticulophagy leads to apoptosis of renal TECs in acute kidney injury/AKI and streptozotocin (STZ)-induced diabetic mice [25,26]. However, the precise role and regulation of reticulophagy in DN conditions remains largely unknown.

RETREG1/FAM134B (reticulophagy regulator 1) is the first discovered reticulophagy receptor, which exerts a vital effect on ER membrane remodeling and turnover. RETREG1 mainly consists of a N-terminal cytoplasmic domain, a reticulon homology domain (RHD), and a MAP1LC3B/LC3B (microtubule associated protein 1 light chain 3 beta)-interacting region (LIR). Among these, the RHD domain and LIR domain are responsible for inducing ER membrane curvature and binding to the autophagosome marker protein LC3 to mediate reticulophagy, respectively [27,28]. It was found that RETREG1-mediated reticulophagy exerts a beneficial effect on myocardial injury by reducing inflammation and apoptosis in sepsis mice [29]. In addition, Kim et al. reported that RETREG1 has neuroprotective effects in the Parkinson disease animal models through binding to the ER resident lectin chaperone CANX (calnexin) and promoting the degradation of SNCA/α-synuclein via reticulophagy [30]. Khaminets et al. found that the absence of RETREG1 promotes ER expansion and ER stress, which leads to degeneration of sensory neurons in an in vivo study [31]. Furthermore, we previously showed that PACS2 (phosphofurin acidic cluster sorting protein 2) gene deficiency blocked reticulophagy, and exacerbated tubulointerstitial inflammation and fibrosis through a TFEB (transcription factor EB)-RETREG1 pathway in STZ-induced diabetic mice [26]. However, the regulatory mechanisms of RETREG1-mediated reticulophagy in proximal tubular cells (PTCs) in DN remain to be elucidated.

In this study, we investigated the role of RETREG1 in reticulophagy and whether abnormal RETREG1 expression could be involved in glucotoxicity-mediated renal tubular injury both in vitro and in vivo. In addition, immunoprecipitation coupled with mass spectrometry (IP-MS) showed that GSTK1 (kappa 1 glutathione S-transferase kappa 1) could interact with RETREG1. Thus, we further explored the mechanism of GSTK1-RETREG1-mediated reticulophagy and tubular cell injury under DN conditions.

Results

The expression of RETREG1 was significantly downregulated in tubular cells under diabetic conditions and was negatively correlated with tubulointerstitial damage

We first investigated the expression of RETREG1 in the kidney tissues of DN animal models and patients. Western blot and real time-quantitative PCR (RT-qPCR) results indicated that RETREG1 expression decreased in the kidneys of db/db mice compared to db/m mice, and was further decreased from 8 to 32 weeks (Figure 1A–C). Meanwhile, immunohistochemistry (IHC) staining revealed that the expression of RETREG1 was notably downregulated in the kidneys of db/db mice, especially in the tubules (Figure 1D). Encouragingly, it was shown that RETREG1 was also significantly downregulated in the renal section of DN patients, and this became more pronounced with the increase of pathological stage (Figure 1D), which was positively correlated with estimated glomerular filtration rate (eGFR), and negatively correlated with the interstitial fibrosis and tubular atrophy (IFTA) scores (Figure 1E,F). Furthermore, we used tubular-specific (Lotus Tetragonolobus Lectin, LTL) and glomerulus-specific (Podocin) antibodies to detect the location of RETREG1 in the kidneys of db/db mice with immunofluorescence (IF) assay. It was found that RETREG1 was mainly expressed in tubules, and slightly expressed in glomeruli in db/db mice (Figure 1G,H). We further confirmed the expression of RETREG1 in human tubular cell (HK-2) cells incubated with HG at various concentrations (0 ~ 45 mM), and at different time points. The results showed that the protein expression of RETREG1 was significantly reduced with the treatment of 30 mM glucose, and further decreased with 45 mM HG incubation compared to control. In addition, RETREG1 protein expression decreased after 12 h HG (30 mM) treatment and further decreased at 24 h and 48 h in a time-dependent manner (Figure 1I–K).

Figure 1.

RETREG1 was dramatically downregulated in tubular cells under diabetic conditions. (A and B) western blot analysis and quantification of RETREG1 expression in the kidney of db/db mice (8 ~ 32 w), db/m as control (n = 4). (C) mRNA expression of Retreg1 assayed by real-time qPCR (n = 4). (D) the expression of RETREG1 was determined by IHC staining in the kidney of db/db mice (8 ~ 32 w) and DN patients (n = 4). Scale bars: 50 μm. (E and F) correlation between RETREG1 expression and eGFR and IFTA. (G and H) fluorescence assay of colocalization of RETREG1 with tubular marker (LTL) and podocyte marker NPHS2/podocin, respectively, in the kidneys of db/db mice (db/m as control). (I and K) western blot analysis and quantification of RETREG1 expression in HK-2 cells stimulated by various concentrations of HG (0 ~ 45 mM) for 24 h, or HG (30 mM) at different time points (n = 3). ACTB was used as a loading control. Values are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant. HG, high glucose; ACTB, actin beta; GS, glucose; NG, normal glucose; eGFR, estimated glomerular filtration rate; IFTA, interstitial fibrosis and tubular atrophy; if, immunofluorescence; IHC, immunohistochemistry; LTL, Lotus tetragonolobus lectin.

Proximal tubular cell-specific Retreg1 gene deficiency aggravated renal injury in HFD- and STZ-induced diabetic mice

To further elucidate the role of RETREG1 in tubular cells under diabetic conditions, we generated mice with a targeted deletion of the Retreg1 gene in the proximal tubule (PT) by crossing Retreg1 exon 4-floxed mice (Retreg1^fl/fl) with Ggt1-Cre mice to obtain the retreg1 ptKO strain (Figure 2A). As shown in Figure 2B, PCR amplification of the tail genome DNA produced 187 bp from wild-type (Retreg1^+/+) mice, whereas bands of 187 bp and 247 bp were observed with the DNA from Retreg1 heterozygous (Retreg1^fl/+) mice and 187 bp amplicon was used to obtain Retreg1 homozygous (Retreg1^fl/fl) mice. The mice with the retreg1^fl/fl Ggt1-Cre genotype were referred to as retreg1 ptKO, while those without Cre amplification were used as littermate control (referred to as Retreg1 Ctrl) (Figure 2B). The deletion of the Retreg1 gene was also confirmed by western blot, IF staining, and RT-qPCR (Figure 2C,F). The DN model was established with Retreg1 Ctrl and retreg1 ptKO mice by high-fat diet (HFD) and streptozotocin (STZ) injection. Compared with control mice, it exhibited significantly increased levels of blood glucose in both HFD- and STZ-induced Retreg1 Ctrl and retreg1 ptKO mice. There was no significant difference in blood glucose and body weight between diabetic Retreg1 Ctrl and retreg1 ptKO mice (Figure 2G,H). The level of urine ALB (albumin) creatine ratio (UACR) was significantly increased in HFD- and STZ-induced Retreg1 Ctrl mice, and further increased in retreg1 ptKO DN mice (Figure 2I). In addition, hematoxylin-eosin (HE), periodic acid Schiff (PAS), and Masson’s staining of diabetic mice kidneys showed dilated cortical proximal tubules with loss of brush borders, mesangial matrix proliferation, and tubulointerstitial fibrosis (Figure 2J). Apoptosis in diabetic kidneys was also substantiated by immunostaining of cleaved (c) CASP3 (caspase 3) and terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining. These lesions were further exacerbated in diabetic retreg1 ptKO mice (Figure 2J). Aggravated interstitial fibrosis was further verified in diabetic retreg1 ptKO mice by western blot assay of FN1 (fibronectin 1) (Figure 2K and L). These data suggested that RETREG1 plays a crucial role in regulating tubular injury under DN conditions.

Figure 2.

PT-specific Retreg1 gene deficiency aggravated tubular injury in diabetic mice. (A) generation of conditional knockout mice in which Retreg1 is specifically ablated in tubular cells by using Ggt1-Cre recombination system. (B) PCR with genomic DNA from tail tissues as templates for verification of the floxed mouse using primer pairs. (C and D) western blot analysis and quantification of RETREG1 protein expression in cortical tissue of Retreg1 Ctrl and retreg1 ptKO mice (n = 4). ACTB was used as a loading control. (E) Representative immunofluorescence images of RETREG1 in kidney tissues of Retreg1 Ctrl and retreg1 ptKO mice. (F) real-time qPCR analysis of RETREG1 in the renal cortex of mice in each group (n = 4). ***p < 0.001. (G and I) body weight, blood glucose, and UACR levels of mice in different groups (n = 4). (J) HE, PAS, Masson staining, and IHC analysis of cCASP3 and TUNEL in each group. Scale bar: 50 μm. (K and L) western blot analysis and quantification of RETREG1 and FN1 protein levels (n = 4). Values are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001; PT, proximal tubule; UACR, urine albumin creatine ratio; ACTB, actin beta; HE, hematoxylin-eosin; PAS, periodic acid schiff; IHC, immunohistochemical; cCASP3, cleaved caspase 3; TUNEL, terminal deoxynucleotidyl transferase dUTP nick-end labeling; FN1, fibronectin 1.

Retreg1 gene ablation in PTCs blocked reticulophagy and resulted in increased ER stress in HFD- and STZ-induced diabetic mice

We then used transmission electron microscopy (TEM) to examine the changes of kidneys in diabetic mice. As shown in Figure 3A, the renal TECs of the HFD- and STZ-induced Retreg1 Ctrl exhibited a decreased number of ER-phagosomes, a phenomenon wherein the fragmented ER was encapsulated within double-membrane autophagosomes, compared to Retreg1 Ctrl mice, and the formation of ER-phagosomes was further decreased in tubular cells from diabetic retreg1 ptKO mice (Figure 3A). Further, co-staining of autophagosome marker LC3B and ER membrane protein CANX was observed by IF. As shown in Figure 3B, the colocalization of punctate LC3 and CANX was significantly decreased in tubules from HFD- and STZ-induced Retreg1 Ctrl mice compared to Retreg1 Ctrl mice, and a more pronounced decrease was found in HFD- and STZ-induced retreg1 ptKO mice (Figure 3B,C). Additionally, the amount of LC3 cleavage and the accumulation of SQSTM1/p62 in the tubules of HFD- and STZ-induced diabetic mice were further aggravated in diabetic retreg1 ptKO mice (Figure 3E,G). IHC revealed that the expression levels of the ER stress proteins, including HSPA5/BiP/GRP78 and DDIT3/CHOP (DNA damage inducible transcript 3) was obviously increased in the kidney tissues of HFD- and STZ-induced Retreg1 Ctrl mice compared with Retreg1 Ctrl mice and these effects were further aggravated in HFD- and STZ-induced retreg1 ptKO mice (Figure 3D). These data were further confirmed by western blot analysis (Figure 3E,I). Taken together, these observations indicated that Retreg1 ablation leads to decreased reticulophagy in PTCs, and reticulophagy deficiency might be closely related to the tubulointerstitial injury of DN.

Figure 3.

Ptcs-specific-Retreg1 deficiency reduced the reticulophagy and exacerbated ER stress in tubular cells of diabetic mice. (A) Representative transmission electron microscopy images of renal TECs in each group. Reticulophagy were indicated by red arrows. Third panel means localization patterns of autophagosomes of reticulophagy. Scale bar: 2 μm. (B) Representative immunofluorescence images of CANX (red) and LC3B (green) in kidney tissues from each group. Scale bar: 50 μm. (C) quantification of the colocalization levels between CANX and LC3B by Pearson’s correlation coefficient. (D) IHC analysis of HSPA5 and DDIT3 in each group. Scale bar: 50 μm. (E and I) western blot analysis and quantification of SQSTM1/p62, LC3, HSPA5 and DDIT3 expression in kidney tissues. ACTB was used as a loading. Values are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001; n = 4. PTCs: proximal tubular cells; CANX: calnexin; LC3B, light chain 3 beta; IHC, immunohistochemical; SQSTM1/p62, sequestosome 1; HSPA5, heat shock protein family a (Hsp70) member 5; DDIT3, DNA damage inducible transcript; ACTB, actin beta.

Overexpression of RETREG1 restored defective reticulophagy, and mitigated ER stress and tubular injury under hyperglycemic conditions

We next conducted a series of in vitro experiments to delineate the effect of RETREG1 on reticulophagy, ER stress, and tubular cell injury under hyperglycemic conditions. We selected the si-RNA03 with the best effect on RETREG1 interference (Figure 4A), and transfected it into HK-2 cells. It was found that the expression levels of cCASP3 (a cell apoptosis marker) and FN1 (an extracellular matrix marker) were notably increased in HK-2 cells exposed to HG, and were further increased by transfection of RETREG1 siRNA (Figure 4B and C). Conversely, upregulated expression of cCASP3 and FN1 induced by HG in HK-2 cells was dramatically inhibited with transfection of a RETREG1 overexpression plasmid (RETREG1 O/E) by western blot (Figure 4D,F). Moreover, compared with control, HG treatment increased the expression of ER stress-related proteins, including HSPA5 and DDIT3, which was further increased by transfecting with RETREG1 siRNA (Figure 4G,H), while these increases were inhibited by transfection with the RETREG1 O/E plasmid (Figure 4I,J).

Figure 4.

RETREG1 increased reticulophagy, mitigated ER stress, and apoptosis in HG-incubated HK-2 cells. (A) three RETREG1 siRnas were transfected into HK-2 cells, and the RETREG1 protein level was detected by western blot (WB) to screen out the appropriate siRnas. ACTB was used as a loading control. (B and C) WB analysis and quantification of FN1 and cCASP3 expression in HK-2 cells transfected with si-RETREG1 under HG conditions. (D) WB analysis of RETREG1 levels in HK-2 cells transfected with empty vector (EV) or RETREG1 overexpression plasmid. (E and F) WB assay and quantification of FN1 and cCASP3 expression in HK-2 cells transfected with RETREG1 overexpression plasmid under HG conditions. (G and H) WB assay and quantification of SQSTM1/p62, LC3, HSPA5 and DDIT3 in HK-2 cells incubated with HG and co-transfected with si-RETREG1. (I and J) WB assay and quantification of SQSTM1/p62, LC3, HSPA5 and DDIT3 proteins expressions in HK-2 cells incubated with HG and transfected with RETREG1 overexpression (O/E) plasmids. (K) Representative immunofluorescence images of LC3 (red) and CNX (green) in HK-2 cells incubated with HG, and co-transfected with si-RETREG1 or RETREG1 O/E plasmids. (n = 3). Scale bar: 10 μm. (L) relative fluorescence intensity analysis was performed by calculating the ratio of the colocalization intensity of CANX and LC3 to CANX fluorescence intensity (Merge:canx). Values are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001; n = 3. FN1, fibronectin 1; cCASP3, cleaved caspase 3; EV, empty vector; NG, normal glucose; HG, high glucose; HK-2, human tubular cell; SQSTM1/p62, sequestosome 1; MAP1LC3/LC3, microtubule associated protein 1 light chain 3; HSPA5, heat shock protein family a (Hsp70) member 5; DDIT3, DNA damage inducible transcript 3; CANX: calnexin.

Next, we explored the effect of RETREG1 on reticulophagy of HK-2 cells in vitro. RETREG1 siRNA downregulated LC3-II:I and degradation of SQSTM1/p62 expression in HK-2 cells incubated with HG (Figure 4G-H), whereas the RETREG1 overexpression plasmid dramatically blocked the effect of HG on LC3-II:I and SQSTM1/p62 protein expression (Figure 4I,J). In addition, the reduction of reticulophagy levels was found in HK-2 cells cultured with HG compared to control (5.6 mM dextrose glucose, NG), as reflected by the colocalization between punctate LC3 and CANX, which was further reduced by RETREG1 siRNA, but was abolished with RETREG1 O/E plasmid (Figure 4K-L). These results suggested that RETREG1 promotes reticulophagy, ultimately leading to mitigated ER stress, fibrosis, and apoptosis in HK-2 cells under HG conditions.

Immunoprecipitation coupled with mass spectrometry (IP-MS) indicated that RETREG1 interacted with GSTK1

To further investigate the mechanism of reticulophagy induction and renoprotection by RETREG1, we next examined the RETREG1-interacting proteins by mass spectrometry of proteins that co-immunoprecipitated with FLAG-tagged full-length RETREG1 in HK-2 cells. We selected twenty-four proteins related to ER homeostasis (Figure 5A). Among them, we were interested in GSTK1, as it has been shown that GSTK1 localized to the ER and suppressed ER stress in adipocytes [32]. Reciprocal co-immunoprecipitation (co-IP) analysis validated the interaction between RETREG1 and GSTK1 in HK-2 cells (Figure 5B). Interestingly, it was shown that GSTK1 interacted with RETREG1 in HEK293T cells, which was independent of the RHD domain (Figure 5C). Moreover, colocalization of RETREG1 and GSTK1 in the kidney was further confirmed by IF, and it was found that the intensity was deceased in DN mice (Figure 5D,E) and patients as compared to control (Figure 5F,G).

Figure 5.

Interaction of RETREG1 with GSTK1 was detected by IP‑MS, Co-IP, and immunofluorescence assay. (A) IP‑MS based protein – protein interaction network of the twenty-four proteins related to ER homeostasis. (B) Co-IP analysis confirmed the interaction between RETREG1 and GSTK1 in HK-2 cells. (C) Co-IP analysis confirmed the interaction between GSTK1 and RETREG1 mutants in HEK293T cells. (D and F) Representative immunofluorescence images of RETREG1 (green) and GSTK1 (red) in kidney tissues from diabetic mice and DN patients. Nuclei were counterstained by DAPI (blue). Scale bar: 50 μm. (E and G) quantification of the colocalization levels between RETREG1 and GSTK1 by Pearson’s correlation coefficient. Valves are presented as mean ± SD. **p < 0.01. n = 4. IP‑MS, immunoprecipitation coupled with mass spectrometry; Co-IP, coimmunoprecipitation.

Gstk1 gene deficiency downregulated RETREG1 expression, and exacerbated reticulophagy abnormality and tubular injury in diabetic mice

First, we conducted a cellular component Gene Ontology/GO analysis of GSTK1 using the BioGRID website, which suggested that GSTK1 is enriched in ER (Figure 6A). The interaction between GSTK1 and ER was further validated by IF with colocalization of GSTK1 and CANX in HK-2 cells (Figure 6B), suggesting that GSTK1 might play an important role in ER homeostasis. To further elucidate the role of GSTK1 in tubular injury of DN in vivo, we hybridized Pck2/Pepck-Cre mice with Gstk1^fl/fl mice to generate proximal tubular cell-specific gstk1 knockout mice (gstk1 ptKO). The DN model was established with Gstk1 Ctrl (Gstk1^fl/fl Cre⁻) and gstk1 ptKO (gstk1^fl/fl Cre⁺) mice by STZ injection as described previously [33]. HE, periodic acid silver methenamine (PASM) and Masson staining of diabetic mice kidneys indicated that dilated cortical proximal tubules with loss of brush borders, as well as tubular basement membrane thickening and interstitial fibrosis. These lesions were more severe in the diabetic gstk1 ptKO group (Figure 6C), indicating that the deletion of Gstk1 aggravated tubular injury. Moreover, fewer ER-phagosomes were presented in tubular cells of diabetic Gstk1 Ctrl mice compared to Gstk1 Ctrl with TEM, while further decreased in that of diabetic gstk1 ptKO mice (Figure 6C). Western blot assay showed that both GSTK1 and RETREG1 protein expression were significantly reduced in the renal tissues of diabetic mice, and RETREG1 was further decreased in diabetic gstk1 ptKO mice (Figure 6D,E). Moreover, colocalization of RETREG1 and punctate LC3 was decreased in tubules of diabetic Gstk1 Ctrl mice as compared to control, and was further reduced in diabetic gstk1 ptKO mice (Figure 6F,G). Indicating that Gstk1 gene deficiency exacerbated reticulophagy abnormality and tubular injury of DN, might through downregulating RETREG1.

Figure 6.

Ptcs-specific Gstk1 gene deficiency exacerbated tubular injury and decreased the expression of reticulophagy receptor RETREG1 in diabetic mice. (A) cellular component gene ontology (GO) analysis of GSTK1. (B) Representative confocal images of CANX (green) and GSTK1 (red) in HK-2 cells. Nuclei were counterstained by DAPI (blue). Scale bar: 50 μm. (C) HE, PASM, and Masson staining of renal tubular cells in each group (scale bar: 50 μm), and representative transmission electron microscopy images of renal tubular cells in each group of mice. Reticulophagy were indicated by red arrows (scale bar: 2 μm). (D and E) western blot analysis and quantification of GSTK1 and RETREG1 protein expression in cortical tissue of different groups of mice. ACTB was used as a loading control. (F) Representative immunofluorescence images of RETREG1 (green) and LC3B (red) in the kidney tissues from each group. Nuclei were counterstained by DAPI (blue). Scale bar: 50 μm. (G) quantification of the colocalization levels between RETREG1 and LC3B by Pearson’s correlation coefficient. Values are presented as mean ± SD. *p < 0.05, ***p < 0.001. n = 4. PTCs, proximal tubular cells; HK-2, human tubular cells; HE, hematoxylin-eosin; PASM, periodic acid silver methenamine.

GSTK1 deficiency exacerbated tubular injury by inhibiting RETREG1-induced reticulophagy in HK-2 cells under hyperglycemia conditions

The expression of GSTK1, RETREG1, and LC3-II:I was decreased in HK-2 cells incubated with HG, which was further decreased by transfection with GSTK1 siRNA (Figure 7A-C and E). However, decreased LC3-II:I expression induced by HG with or without GSTK1 siRNA, was restored with RETREG1 O/E plasmid transfection (Figure 7A,E). Contrary results were observed for FN1 expression in cells transfected with RETREG1 O/E plasmid or GSTK1 siRNA in HK-2 cells with HG conditions (Figure 7A,D). We further monitored reticulophagy using ssRFP-GFP-KDEL, a fluorescent reticulophagy tandem reporter [34,35]. Delivery of ER membrane fragments into lysosomes leads to fluorescence quenching of GFP in acidic lysosomal environments, while the fluorescence of RFP is preserved. Therefore, reticulophagy activity was reflected by red-only fluorescence. It was observed that the intensity of red-only puncta decreased in HK-2 cells incubated with HG, and further downregulated by transfection of GSTK1 siRNA, while this reduction was reversed by RETREG1 O/E (Figure 7F,G). In addition, we also monitored reticulophagy of tubular cells with mCherry-EGFP-RETREG1 immunofluorescence reporter [36,37]. It was shown that under HG conditions, the intensity of red-only (mCherry⁺ EGFP⁻) puncta in HK-2 cells decreased, and was restored with the transfection of GSTK1 O/E plasmid, while the effect was abolished with RHD deletion of RETREG1 (RETREG1ΔRHD). These findings indicated that RHD domain of RETREG1 is critical for reticulophagy in tubular cells under HG stress (Figure 7H,I). Considering that the RHD domain of RETREG1 is important for its oligomerization and reticulophagy [36], then we observed whether the oligomerization of RETREG1 was affected by GSTK1. The result showed that the oligomerization of RETREG1 increased by the overexpression of GSTK1, but was abolished with transfection with RETREG1ΔRHD plasmid (Figure 7K). Besides, we also performed a ubiquitination assay, it was found that the ubiquitination level of RETREG1 significantly decreased in HEK293T cells transfected with GSTK1 overexpression plasmid (Figure 7J). Suggesting that GSTK1 and RETREG1 is critical to regulate reticulophagy and tubular injury under hyperglycemic conditions.

Figure 7.

GSTK1 enhanced RETREG1 expression, oligomerization, and its activity in ER-phagy in HK-2 cells under hyperglycemia conditions. (A–C and E) western blot analysis and quantification of GSTK1, RETREG1, FN1 and LC3-II:I in HK-2 cells incubated with GSTK1 siRNA or RETREG1 overexpression plasmid under HG conditions. (F) fluorescence detection of HK-2 cells transfected with ssRFP-GFP-KDEL plasmid by confocal microscopy in each group. Scale bar: 10 μm. (G) quantification of the average of red-only puncta per cell in (F). (H) fluorescence detection of HK-2 cells transfected with mcherry-EGFP-RETREG1 by confocal microscopy in each group. Scale bar: 5 μm. (I) lysosomal mCherry-positive but GFP-negative puncta were quantified for each cell in (H). (J) RETREG1 oligomerization assay in HK-2 cells transfected with GSTK1 O/E, RETREG1 O/E, and RETREG1 ΔRHD mutant. (K) RETREG1 ubiquitination assay in HEK293T cells transfected with HA-ubiquitin, Flag-RETREG1 and MYC-GSTK1. Values are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant. n = 3.

Discussion

In the present study, the expression of RETREG1 was decreased in HG-induced HK-2 cells, the kidneys of db/db mice, and patients with DN, which was tightly correlated with tubular injury and renal function deterioration. Furthermore, we applied the proximal tubule-specific retreg1 knockout mice and found that Retreg1 gene deficiency blocked reticulophagy, aggravated ER stress, and tubular injury of diabetic mice. In vitro studies suggested that overexpression of RETREG1 increased reticulophagy and mitigated ER stress, and cell apoptosis in HK-2 cells under hyperglycemic conditions, whereas silencing of RETREG1 showed the opposite effects. The potential mechanism was that RETREG1 improved reticulophagy of tubular cells mediated by GSTK1. These data indicated that GSTK1- and RETREG1-mediated reticulophagy contributed to protecting tubular cell injury via ameliorating ER stress in DN. Further enriched the understanding of the role and mechanism of reticulophagy in the development of DN.

The role of ER homeostasis has been increasingly recognized in the progression of DN. Accumulating studies have confirmed that the activation of ER stress and disorder in autophagy results in the accumulation of damaged proteins and organelles in diabetic renal cells, thereby promoting the development of DN [38,39]. Our previous studies observed decreased lipophagy, reduced mitophagy, and enhanced ER stress in renal tubular cells of DN patients and animal models [40,41]. Reticulophagy has been considered as a novel process to regulate cellular ER quality control and ER homeostasis [42]. RETREG1 is the first identified and the best characterized mammalian reticulophagy receptor, and mediates the recognition and removal of the ER for lysosomal degradation [21,27,28]. It has been demonstrated that dysfunction of RETREG1 is involved in various diseases including neurodegenerative disorders, cancer, viral infection, and vascular disease [27,28]. In addition, Khaminets et al. found that downregulation of RETREG1 blocked reticulophagy, leading to ER scission, ER stress and subsequent neuropathies [31]. However, the role of RETREG1-mediated reticulophagy in diabetes-induced tubular ER stress remains unknown.

In this study, we demonstrated that RETREG1 is significantly downregulated both in tubules from db/db mice and HK-2 cells incubated with HG (Figure 1). In particular, we also identified that the level of RETREG1 decreased in the renal tubules of DN patients, which negatively correlated with IFTA scores but positively correlated with eGFR, indicating that RETREG1 expression was tightly associated with tubular injury of DN (Figure 1). To verify the role of RETREG1 in tubular cells of DN, we generated PT-specific retreg1 knockout mice, and then induced to DN model by HFD and STZ. It was shown that Retreg1 deficiency exacerbated the disruption of reticulophagy in tubular cells, accompanied by increased ER stress, apoptosis and tubulointerstitial fibrosis in diabetic mice (Figures 2 and 3). In vitro experiment suggested that overexpression of RETREG1 dramatically enhanced reticulophagy and ameliorated ER stress, and cell apoptosis in HK-2 cells incubated with HG, and RETREG1 siRNA further increased the effect of HG (Figure 4). These findings indicated that RETREG1 plays a beneficial role in ER stress and tubular injury by enhancing reticulophagy in DN state.

To further clarify the potential mechanisms, we performed an IP-MS analysis, and the results showed that RETREG1 interacts with GSTK1 (Figure 5). GSTK1/DsbA-L, is an approximately 25-kDa protein that is expressed in renal PTCs of kidneys and primarily localized in mitochondria and endoplasmic reticulum, etc [32,43]. Our previous studies demonstrate that Gstk1 deficiency aggregated ectopic fat disposition, mitophagy, and the loss of mitochondria-associated ER membrane/MAM integrity in the renal tubules of diabetic mice [44–46]. Notably, besides to mitochondrial regulation, it is demonstrated that the N-terminal amino acid residues of GSTK1 allow its localization to the ER, and interaction with the ER membrane-associated ERO1A/ERO1-Lα (endoplasmic reticulum oxidoreductase 1 alpha), thereby mitigating ER stress and promoting ADIPOQ/adiponectin multimerization [32]. However, whether GSTK1 mediated reticulophagy remains unknown. In this study, it was found that conditional gstk1 deletion in PTCs induced a further decrease of reticulophagy and aggravated tubulointerstitial lesions by downregulating RETREG1 in diabetic mice (Figure 6). Moreover, GSTK1 siRNA transfection further decreased RETREG1 expression and reticulophagy, and thus aggravated cell apoptosis in HK-2 cells under HG ambience, which was reversed by RETREG1 overexpression plasmid transfection (Figure 7). These findings suggested that RETREG1-mediated reticulophagy was regulated by GSTK1, thus exerting a significant impact on tubular injury in DN.

It is reported that ubiquitination of RETREG1 within its RHD domain promotes RETREG1 aggregation and binding to lipidated LC3B, thereby stimulating reticulophagy [47]. Another research showed that CAMK28 phosphorylates the RHD domain of RETREG1, which promotes its oligomerization and reticulophagy [36]. Based on these findings, we performed experiments to investigate whether the ubiquitination and oligomerization of RETREG1 were affected by GSTK1 and the possible role of RHD domain in this process. It was shown that GSTK1 overexpression significantly decreased the ubiquitination level of RETREG1 in HEK293T cells. In addition, GSTK1 overexpression increased the oligomerization of RETREG1 in tubular cells, while this effect was abolished in that of RHD domain deletion (Figure 7), which is in accordance with the result reported previously [36]. This indicated that GSTK1 could regulate RETREG1 and subsequent reticulophagy partially through RHD domain. However, the specific binding sites remain to be further clarified.

In summary, in this study, we found that reticulophagy plays an important role in maintaining ER homeostasis of tubular cells in DN. Furthermore, we revealed that Retreg1 deficiency might exacerbate diabetic tubular injury through reticulophagy regulated by GSTK1. This indicates that GSTK1 regulates reticulophagy to attenuate endoplasmic reticulum stress and apoptosis in DN via the RETREG1/FAM134B pathway (Figure 8). Although there is no direct evidence of drug research targeting GSTK1 and RETREG1, some compounds, such as silymarin and curcumin, have been shown to increase GST activity (GSTK1 is the main GST isoform) and play a protective role in renal tubular damage in DN mice [48–50]. Together, these data might suggest GSTK1 and RETREG1/FAM134B mediated reticulophagy could serve as a potential new target of tubular injury in DN.

Figure 8.

A proposed model illustrating the role of GSTK1 and RETREG1 in regulating reticulophagy. The expression of GSTK1 in renal tubular cells is reduced by high glucose, GSTK1 interacted with RETREG1 and down regulated expression of RETREG1. Subsequently, this molecular cascade leads to impaired reticulophagy process, resulting in reduced reticulophagy, aggravated ER stress and tubular cell injury under DN conditions.

Materials and methods

Antibodies and reagents

Antibodies against the following proteins were used: anti-RETREG1 (83414S), anti-HSPA5/BiP (3177S), anti-DDIT3 (2895) and anti-cleaved CASP3 (9661S) were purchased from Cell Signaling Technology. Anti-RETREG1 (21537–1-AP), anti-GSTK1 (14535–1-AP), anti-CANX (10427–2-AP 66,903–1-Ig), anti-MAP1LC3B (81004–1-RR), anti-ACTB/β-ACTIN (66009–1-Ig), anti-IgG (B900620, 30000–0-AP), anti-MYC (60003–2-Ig) and anti-HA (51064–2-AP) were purchased from Proteintech. Anti-FN1 (ab2413) was purchased from Abcam. Anti-Flag (F1804) was purchased from Sigma. MAP1LC3B (NB100–2220) was purchased from Novus Biologicals. Anti-SQSTM1/p62 (GB11531–100) was purchased from Sevier Biotechnology Co., Ltd. Protein A/G agarose resin was purchased from Yeasen (36403ES03). STZ (S0130) was acquired from Sigma. The high-fat diet (containing 35 kcal% carbohydrate, 45 kcal% fat, and 20 kcal% protein; D12451) was obtained from Research Diets, Inc. Lipofectamine 2000 was purchased from Invitrogen (11668019). siRNAs against RETREG1 and GSTK1 were synthesized by RiboBio, Inc.

Plasmids

The Flag-RETREG1, MYC-GSTK1, and HA-UB plasmid was obtained from MiaoLingBio. The pCW57-CMV-ssRFP-GFP-KDEL plasmid (128257) was obtained from Addgene (deposited by Xiao lab, the Second Xiangya Hospital of Central South University). RETREG1 plasmids (Flag-RETREG1 WT, RHD, ΔRHD) and mCherry-EGFP-FRETREG1 (WT, RHD, ΔRHD) were gifts from Professor Qiming Sun (Zhejiang University, China).

General information on patients with DN

Twelve patients with DN diagnosed by renal biopsy were selected for this study. In addition, patients with glomerular minor lesions (n = 4) were recruited as control, as previously described [33,44]. The DN patients were categorized into classes II, III and IV according to the pathological classification of DN [44]. The baseline clinical data of DN patients and control in this study were shown in Table 1. The human protocol was approved by the Ethics Committee of Second Xiangya Hospital, Central South University (human ethics approval number: LYEC2025-K0052).

Table 1.

Clinical characteristics.

	Control (n = 4)	II (n = 4)	III (n = 4)	IV (n = 4)
Age (year)	27.5 ± 7.77	44.0 ± 7.83a	44.5 ± 7.14a	51.0 ± 7.79a
Sex (M/F)	1/3	3/1	3/1	2/2
SCr (μmol/l)	61.18 ± 10.6	114.70 ± 27.55a	150.73 ± 17.56a	226.30 ± 40.66abc
BUN (mmol/L)	3.77 ± 1.12	6.29 ± 0.57	8.33 ± 2.22	13.77 ± 4.94ab
UA (μmol/l)	329.10 ± 111.45	338.78 ± 88.30	390.4 ± 53.61	434.05 ± 40.78
Proteinuria (g/24 h)	0.51 ± 0.54	5.52 ± 3.29	8.64 ± 4.03a	7.83 ± 3.87a
eGFR (mL/min/1.73 m2)	124.73 ± 9.00	64.23 ± 6.44a	47.97 ± 10.34a	28.77 ± 6.13abc
Glucose (mmol/l)	4.25 ± 0.27	6.99 ± 0.95	8.44 ± 1.78a	8.73 ± 2.17ab
HbA1C (%)	5.67 ± 0.45	7.20 ± 1.70	6.53 ± 1.00	8.07 ± 1.17
TG (mmol/L)	1.44 ± 0.62	1.64 ± 0.36	2.92 ± 1.86	2.48 ± 1.22
CHOL (mmol/L)	5.26 ± 0.95	5.21 ± 2.88	5.50 ± 1.95	5.87 ± 1.48
ALB (g/L)	41.90 ± 6.44	29.75 ± 12.84	29.60 ± 4.31	32.53 ± 7.73

F, female; M, male; SCr, serum creatinine; BUN, blood urea nitrogen; UA, uric acid; eGFR, estimated glomerular filtration Rate; HbA1C, glycosylated hemoglobin; TG, triglycerides; CHOL, total cholesterol; ^ap < 0.05 versus control; ^bp < 0.05 versus the II group. ^cp < 0.05 versus the III group. Values are Mean ± SD.

Generation of proximal tubule-specific retreg1-knockout mice and proximal tubule-specific gstk1-knockout mice

Heterozygote Retreg1 mice (Retreg1^flox/+; C57BL/6J background) were generated using CRISPR-Cas9 gene-editing technology by Cyagen Inc., China (Gene ID: S-CKO-12985). Ggt1-Cre (C57BL/6N background) mice were also purchased from the Cyagen Inc., China (Gene ID: C001028). The Retreg1^flox/flox mice were crossed with Ggt1-Cre mice to produce the proximal tubule-specific retreg1 KO mice (retreg1 ptKO) and their littermate control (Retreg1 Ctrl). The mice were genotyped by polymerase chain reaction (PCR) analysis of genomic DNA from tail tissues using a Mouse Direct PCR kit (Bimake, USA), as described previously [33]. The following primers were used as follows: forward primer: Retreg1 forward primer (5’–3’), ATAGAGATTTACTTAGCCCACTAC; Retreg1 reverse primer (5’–3’), TTTTAGACTACAAGTGAGACTCAG; Cre forward primer (5’–3’), TGTGCTGTGACTTCTTATTCTTAG; Cre reverse primer (5’–3’), TTTTAGACTACAAGTGAGACTCAG. In addition, Gstk1^flox/flox mice and Pck2/Pepck Cre mice were provided by Professor Lin Sun from the Second Xiangya Hospital of Central South University as described in a previous study [33].

Mouse model

Male mice aged 6 weeks were randomly divided into 4 groups: Retreg1 Ctrl group, retreg1 ptKO group, HFD- and STZ-induced Retreg1 Ctrl group and HFD- and STZ-induced retreg1 ptKO group. Both Retreg1 Ctrl and retreg1 ptKO groups were given a normal diet, whereas the other two groups were fed with an HFD for four weeks and then injected intraperitoneally with STZ (100 mg/kg; Sigma, S0130) to induce hyperglycemia, as described in a previous study [44]. Mice with random blood glucose levels > 16.7 mmol/L were selected for the experiment, and they were maintained on HFD feeding for another 20 weeks. In addition, male mice aged 6 weeks were randomly divided into 4 groups: Gstk1 Ctrl group, gstk1 ptKO group, diabetic group, and diabetic gstk1 ptKO group. Diabetic group mice were injected intraperitoneally with STZ (50 mg/kg) consecutively for 5 days, as previously described [51]. Three days after the STZ injection, mice with random blood glucose levels > 16.7 mmol/L were selected for the experiment, and were euthanized after 12 weeks. The body weights and fasting blood glucose were monitored every 2 weeks, and at the end of the last week, all mice were anesthetized with an intraperitoneal injection of 50 mg/kg body weight sodium pentobarbital and sacrificed. The samples of urine, serum, and kidney tissues were collected for further studies. The Medical Ethics Committee of Central South University approved all animal procedures (animal ethical approval number: 2021086).

Measurement of urine ALB and creatinine

Urine ALB concentrations were determined using a mouse urine albumin ELISA kit (Bethyl Laboratories, G-AEFI00580.96) following the manufacturer’s instructions. Serum and urine creatinine levels were determined using a QuantiChrom Creatinine Assay Kit (BioAssay Systems, DICT-500) according to the descriptions of this kit.

Morphological analysis of kidneys

Four-μm-thick kidney tissue sections were stained with HE, PAS and Masson staining. IFTA was scored as previously described: 0 = no glomerular/tubular damage, 1 < 25% of the glomerular/tubular area was affected, 2 = 25%–50% of the glomerular/tubular area was affected, and 3 > 50% of the glomerular/tubular area was affected [52].

Immunohistochemistry (IHC)

Paraffin-embedded kidney tissue sections were dewaxed, rehydrated and incubated with primary antibodies at 4°C overnight. After washing, the sections were incubated with secondary antibodies and diaminobenzidine (DAB) substrate (Servicebio, G1212-200T). The sections were analyzed with a Leica microscope and quantified using ImageJ software.

Transmission electron microscopy (TEM)

Fresh kidney tissues (1 mm³) were immediately fixed with 2.5% glutaraldehyde and immersed in 1% osmium tetroxide solution. After washing, the samples were dehydrated and embedded. Ultrathin sections were cut and stained with uranyl acetate and lead citrate. The structure of double-membrane autophagosomes encapsulating the ER was finally observed by transmission electron microscopy.

Real time-quantitative PCR (RT-qPCR)

Total RNA was extracted from renal cortex using RNAiso Plus (TaKaRa, 9108). RNA was reverse-transcribed into cDNA using the PrimeScript Reagent Kit (TaKaRa, RR047A) and qPCR was performed with TB GreenTM Premix Ex Taq II reagent (TaKaRa, RR420L) on a 7300 Real-Time PCR System (Applied Biosystems). The primer sequences used were as follows: Retreg1 forward primer (5’–3’), CAAGCAGGAGTACGATGAGTC; Retreg1 reverse primer (5’–3’), AACGCAGCTCAGTAACAGTC; Actb forward primer (5’–3’), CAAGCAGGAGTACGATGAGTC, Actb reverse primer (5’–3’), AACGCAGCTCAGTAACAGTC.

Cell culture

The HK-2 and HEK293T cell lines were purchased from the American Type Culture Collection (CRL-2190, CRL-3216). HK-2 cells were cultured as previously described [52]. HEK293T cells were cultured in DMEM (Gibco 11,965,092) supplemented with 10% FBS (Gibco 10,270,106), 2 mM L-glutamine (Gibco 25,030,081), and 1% penicillin-streptomycin (Gibco 15,140,122) in a humidified incubator at 37°C with 5% CO₂. Time-dependent experiments were performed using 30 mM D-glucose (Sigma, G6152) for 0–48 h. Furthermore, HK-2 cells were transiently transfected with GSTK1 siRNA, RETREG1 siRNA, or an overexpression plasmid using Lipofectamine 2000 reagent according to the manufacturer’s instructions.

Western blot analysis

Total proteins of renal cortices and HK-2 cells were extracted using radioimmunoprecipitation assay (RIPA) buffer (CoWin Biosciences, CW2333S) containing protease inhibitors (CoWin Biosciences, CW2200S) and phosphatase inhibitors (CoWin Biosciences, CW2383S). A BCA protein assay kit (Thermo Fisher Scientific 23,225) was used to determine the protein concentration. Total proteins were subjected to 8% and 12% SDS-PAGE and then transferred onto PVDF membranes, which were probed with primary antibodies at 4°C overnight. After incubation with a suitable secondary antibody, the membrane blots were measured with enhanced chemiluminescence (ECL; Amersham Pharmacia Biotech, RPN2232) as described previously [52].

Immunofluorescence (IF)

Paraffin-embedded renal sections were dewaxed, hydrated, antigen repaired and blocked in 5% BSA (Sigma, A9647). Meanwhile, HK-2 cells were fixed, infiltrated and blocked. Then, the cells or slides were incubated with primary antibody overnight at 4°C. After washing with PBS (Servicebio, G4202-500 ML), sections were incubated with Alexa Fluor 488 (green; Abcam, ab150113) or 594 (red; Abcam, ab150080) at 37°C for 1 h. The nuclei were stained with 4,’6-diamidino-2-phenylindole (DAPI; SouthernBiotech, 0100–20) as previously described [33,52].

Co-immunoprecipitation (co-IP)

HK-2 and HEK293T cells were lysed with RIPA buffer containing protease and phosphatase inhibitors. The supernatants of lysis were incubated overnight at 4°C with the indicated antibodies and then precipitated by the Protein A/G PLUS-Agarose (Yeasen, 36403ES03). After washing, the precipitated materials were used for western blot analysis.

Ubiquitination assay

As previously described [53], the ubiquitination of RETREG1 was assessed in HEK293T cells, transfected with HA-ubiquitin, Flag-RETREG1 and MYC-GSTK1. Cells were lysed in lysis buffer (RIPA buffer containing protease and phosphatase inhibitors). The lysates were then incubated on ice for 30 min and centrifuged at 12,000 g at 4°C for 30 min. Next, 40 µl of the supernatant was collected, mixed with loading buff, boiled for 10 min at 100°C, and stored at −20°C as input control. The supernatants of lysis were incubated overnight at 4°C with the anti-Flag antibodies and then precipitated by the Protein A/G PLUS-Agarose. Beads were washed three times with lysis buffer, heated at 100°C for 10 min, subjected to SDS-PAGE and analyzed by immunoblotting to detect the HA-Tag.

Oligomerization assay

HK-2 cells were transfected with GSTK1 O/E, RETREG1 WT O/E, and RETREG1 ΔRHD mutants to detect the effect of GSTK1 on RETREG1 oligomerization. As described previously [36], RETREG1 oligomers are partially resistant to denaturing solutions containing SDS and DTT. Cell pellets were homogenized in non-reducing buffer (Epizyme Biomedical Technology Co., Ltd., LT103), loaded onto 8% SDS-PAGE gels, and immunoblotted using an anti-RETREG1 antibody to detect.

Statistical analysis

All statistical analyses were analyzed using GraphPad Prism 8.0 software and SPSS 23.0 software. Values were presented as means ± SD and assessed by Student’s t-test or one-way analysis of variance. Correlation analysis was performed using Pearson’s correlation analysis. p < 0.05 was considered statistically significant.

Funding Statement

This work was supported by the National Science Foundation of China (No.82170744, No.82000697), Natural Science Foundation of Hunan Province (2025JJ50604), Hunan Provincial Health Commission high-level talents major scientific research project (R2023083). Hunan Provincial Innovation Foundation for Postgraduate (CX20220340), Fundamental Research Funds for Central Universities of the Central South University (2022ZZTS0258).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data from this study are available from the Lead Contact, Dr. Li Xiao (xiaolizndx@csu.edu.cn).