Endoplasmic reticulum protein TXNDC5 promotes renal fibrosis by enforcing TGF-β signaling in kidney fibroblasts

Abstract

Renal fibrosis, a common pathological manifestation of virtually all types of chronic kidney disease (CKD), often results in diffuse kidney scarring and predisposes to end-stage renal disease. Currently, there is no effective therapy against renal fibrosis. Recently, our laboratory identified an ER-resident protein, thioredoxin domain containing 5 (TXNDC5), as a critical mediator of cardiac fibrosis. Transcriptome analyses of renal biopsy specimens from patients with CKD revealed marked TXNDC5 upregulation in fibrotic kidneys, suggesting a potential role of TXNDC5 in renal fibrosis. Employing multiple fluorescence reporter mouse lines, we showed that TXNDC5 was specifically upregulated in collagen-secreting fibroblasts in fibrotic mouse kidneys. In addition, we showed that TXNDC5 was required for TGF-β1-induced fibrogenic responses in human kidney fibroblasts (HKFs), whereas TXNDC5 overexpression was sufficient to promote HKF activation, proliferation, and collagen production. Mechanistically, we showed that TXNDC5, transcriptionally controlled by the ATF6-dependent ER stress pathway, mediated its profibrogenic effects by enforcing TGF-β signaling activity through posttranslational stabilization and upregulation of type I TGF-β receptor in kidney fibroblasts. Using a tamoxifen-inducible, fibroblast-specific Txndc5 knockout mouse line, we demonstrated that deletion of Txndc5 in kidney fibroblasts mitigated the progression of established kidney fibrosis, suggesting the therapeutic potential of TXNDC5 targeting for renal fibrosis and CKD.

Keywords: Cell Biology; Fibrosis; Nephrology; Protein misfolding.

Figure 1. TXNDC5 was significantly upregulated in mouse fibrotic kidneys and kidney specimens from patients with CKD.

(A) IHC staining (n = 3) and (B) immunoblots (n = 6) showed protein expression of TXNDC5 was upregulated in mouse fibrotic kidneys induced by UUO or uIRI, compared with contralateral kidneys (CL). Scale bar: 50 μm. (C) Quantitative RT-PCR showed Txndc5 transcript was upregulated in mouse fibrotic kidneys induced by UUO and uIRI (n = 4–7). (D) Reanalyses of microarray data on human kidney specimens from patients with CKD (GSE66494) showed that TXNDC5, COL3A1, and FN1 were significantly upregulated in the kidney tissues from CKD patients (healthy control n = 8, CKD n = 51). For A–C, data are representative of 3 or more independent experimental replicates. For all panels, data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 by 2-sided t test (A–C) or Mann-Whitney test (D).

Figure 2. TXNDC5 was highly upregulated in renal fibroblasts, but not in TECs, endothelial cells, or podocytes, of the fibrotic kidneys.

(A and C) IF staining of TXNDC5 (red) on sections of fibrotic kidneys induced by UUO in Col1a1-GFP^Tg mice showed TXNDC5 was mainly expressed in collagen-secreting renal fibroblasts (green), both in renal cortex and medulla (n = 6). Cell nuclei were stained with DAPI (blue). Scale bar: 100 μm. (B) IF staining of TXNDC5 (green) on section of fibrotic kidneys induced by UUO in Cdh16-Cre, NPHS2-Cre, and Tie2-Cre/ERT2 tdTomato mice. Cell nuclei were stained with DAPI (blue). Scale bar: 100 μm. (D) A pie chart to illustrate the proportion of TXNDC5⁺ cells in different types of kidney cells in the UUO-induced fibrotic kidneys. Data are representative of 3 or more independent experimental replicates. Data in C are presented as mean ± SEM.

Figure 3. Knockdown of TXNDC5 attenuated TGF-β1–induced HKF activation and ECM production; overexpression of TXNDC5 was sufficient to trigger HKF activation and ECM production.

(A) Protein and (B) transcript expression levels of fibroblast activation marker (periostin) and ECM proteins (COL1A1, fibronectin, and CCN2) were increased in control (Scramble) HKFs following TGF-β1 (10 ng/mL) treatment. Knockdown of TXNDC5 attenuated the upregulation of these fibrogenic markers induced by TGF-β1 in HKFs (n = 5–10). (C) Overexpression of TXNDC5 was sufficient to induce upregulation of fibroblast activation marker (Periostin) and ECM proteins (COL1A1, fibronectin) in HKFs (n = 3–10). (D) Treatment of TGF-β1 (10 ng/mL) increased the cellular proliferation activity of HKFs, which was abrogated by TXNDC5 knockdown. (E) Overexpression of TXNDC5 increased the cellular proliferation activity of HKFs. In D and E, n = 10. Data are representative of 3 or more independent experimental replicates. For all panels, data are presented as mean ± SEM. The statistical significance of differences for 2 groups was determined by 2-sided t test and among 3 or more groups it was determined using 1-way ANOVA, followed by Sidak’s post hoc tests. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4. Deletion of Txndc5 attenuated renal fibrosis induced by UUO.

(A) Picrosirius red staining (top 2 panels) and Masson’s trichrome staining (bottom 2 panels) of kidney sections from WT and Txndc5^–/– mice 10 days after UUO. Bar graphs of the quantitative results of Picrosirius red and Masson’s trichrome staining areas are shown on the right (n = 5–10). Scale bar: 100 μm. (B) SHG images of the kidney sections from WT and Txndc5^–/– mice 10 days after UUO. The quantitative results of SHG-positive areas showed increased accumulation of fibrillar collagen (green) in WT but not in Txndc5^–/– mice kidneys following injury. For each of the kidney sections imaged for SHG, TPEF imaging was obtained to show the profile of the scanned tissue (red color in bottom panels) (n = 3). Scale bar: 50 μm. (C) Immunoblots showed protein expression levels of fibroblast activation marker (Periostin) and ECM (COL1A1) in the whole-kidney extract from WT and Txndc5^–/– mice 10 days after UUO (n = 3–6). Data are representative of 3 or more independent experimental replicates. For all panels, data are presented as mean ± SEM. The statistical significance of differences among 3 or more groups was determined using 1-way ANOVA, followed by Sidak’s post hoc tests. **P < 0.01, ***P < 0.001.

Figure 5. Deletion of Txndc5 ameliorated renal fibrosis induced by uIRI.

(A) Picrosirius red staining (top 2 panels) and Masson’s trichrome staining (bottom 2 panels) of kidney sections from WT and Txndc5^–/– mice 28 days after uIRI. Bar graphs of the quantitative results of Picrosirius red and Masson’s trichrome staining areas are shown on the right (n = 6). Scale bar: 100 μm. (B) SHG images of the kidney sections from WT and Txndc5^–/– mice 28 days after uIRI. The quantitative results of SHG-positive areas showed increased accumulation of fibrillar collagen (green) in WT but not in Txndc5^–/– mice kidneys following injury. For each of the kidney sections imaged for SHG, TPEF imaging was obtained to show the profile of the scanned tissue (red color in bottom panels) (n = 3). Scale bar: 50 μm. (C) Immunoblots showed protein expression levels of fibroblast activation marker (Periostin) and ECM (COL1A1) in the whole-kidney extract from WT and Txndc5^–/– mice 28 days after uIRI (n = 5–9). Data are representative of 3 or more independent experimental replicates. For all panels, data are presented as mean ± SEM. The statistical significance of differences among 3 or more groups was determined using 1-way ANOVA, followed by Sidak’s post hoc tests. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 6. TXNDC5-induced fibrogenic responses are mediated through TGFBR1.

(A) Immunoblots showed that treatment of TGF-β1 (10 ng/mL) induced upregulation of TGFBR1 and phosphorylation of SMAD3, whereas TGFBR2 was not affected. Knockdown of TXNDC5 abolished TGFBR1 upregulation and SMAD3 phosphorylation was induced by TGF-β1 in HKFs (n = 5–11). (B) Overexpression of TXNDC5 was sufficient to upregulate TGFBR1 and phospho-SMAD3 in HKFs (n = 5–6). (C) Knockdown of TGFBR1 abolished the upregulation of fibroblast activation markers and ECM proteins induced by TXNDC5 overexpression (n = 6–12). Data are representative of 3 or more independent experimental replicates. For all panels, data are presented as mean ± SEM. The statistical significance of differences for 2 groups was determined by 2-sided t test and among 3 or more groups it was determined using 1-way ANOVA, followed by Sidak’s post hoc tests. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 7. TXNDC5 enhances the protein folding and stability of TGFBR1.

(A) Cycloheximide chase assay showed that knockdown of TXNDC5 accelerated degradation of TGFBR1 in HKFs (n = 6). (B) Overexpression of TXNDC5 slowed down the degradation of TGFBR1 in HKFs. The 2 groups of samples were loaded and run on the same gel but not in neighboring lanes (n = 6). (C) Treatment of proteasome inhibitor MG132 restored the downregulation of TGFBR1 induced by TXNDC5 depletion in HKFs (n = 11). (D) Protein Co-IP experiments in HEK cells with ectopic expression of TXNDC5 and Myc-tagged TGFBR1 showed physical interaction between TXNDC5 and TGFBR1. (E) A dual fluorescence–labeled human TGFBR1 construct (CFP-TGFBR1-YFP) was used for FRET-based protein folding assay in HEK293 cells. Overexpression of TXNDC5 showed increased FRET efficiency compared with empty vector–transfected control cells, whereas AAA-mutant TXNDC5 transfection failed to increase the FRET signal (n = 10). Data are representative of 3 or more independent experimental replicates. For all panels, data are presented as mean ± SEM. The statistical significance of differences for 2 groups was determined by 2-sided t test and among 3 or more groups it was determined using 1-way ANOVA, followed by Sidak’s post hoc tests. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 8. TGF-β1 induces TXNDC5 expression through ER stress– and ATF6-dependent transcriptional regulation.

(A) ER stress components including BiP and activated ATF6 (ATF6-p50) were upregulated following the treatment of TGF-β1 (10 ng/mL) in HKFs (n = 3). (B) ER stress inhibitor 4-PBA blocked TXNDC5 transcript upregulation induced by TGF-β1 (10 ng/mL) (n = 6–8). (C) Quantitative real-time PCR showed effective ATF6 knockdown by shATF6 in HKFs (n = 6). (D) Depletion of ATF6 reversed TXNDC5 transcript upregulation induced by TGF-β1 (10 ng/mL) (n = 6). (E) In NIH-3T3 cells, TGF-β1 treatment increased the transcriptional activity of WT but not ATF6 binding–deleted mouse Txndc5 promotor (n = 24). Data are representative of 3 or more independent experimental replicates. For all panels, data are presented as mean ± SEM. The statistical significance of differences for 2 groups was determined by 2-sided t test and among 3 or more groups it was determined using 1-way ANOVA, followed by Sidak’s post hoc tests. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 9. Targeted deletion of Txndc5 in renal fibroblasts attenuated kidney fibrosis.

(A) Illustration of experimental design to induce Txndc5 deletion specifically in renal fibroblasts. (B) Picrosirius red staining of kidney sections from WT and Txndc5^cKO mice 10 days after UUO (n = 6–7). Scale bar: 50 μm. (C) Immunoblots to quantify fibroblast activation marker (POSTN) and ECM (COL1A1 and CCN2) proteins in whole-kidney lysates from Col1a2-Cre and Txndc5^cKO mice 10 days after UUO (n = 5–6). (D) SHG images of kidney sections from Col1a2-Cre and Txndc5^cKO mice 10 days after UUO. The quantitative results of SHG-positive areas showed accumulation of fibrillar collagen in Col1a2-Cre kidneys, which was ameliorated in Txndc5^cKO mice (n = 3). Scale bar: 50 μm. Data are representative of 3 or more independent experimental replicates. For all panels, data are presented as mean ± SEM. The statistical significance of differences among 3 or more groups was determined using 1-way ANOVA, followed by Sidak’s post hoc tests. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 10. Targeted deletion of Txndc5 in renal fibroblasts mitigated kidney fibrosis induced by uIRI.

(A) Picrosirius red staining of kidney sections from WT and Txndc5^cKO mice 28 days after uIRI (n = 7–11). Scale bar: 50 μm. (B) Immunoblots to quantify fibroblast activation marker (POSTN) and ECM (COL1A1 and CCN2) proteins in whole-kidney lysates from Col1a2-Cre and Txndc5^cKO mice 28 days after uIRI (n = 4–11). (C) SHG images of kidney sections from Col1a2-Cre and Txndc5^cKO mice 28 days after uIRI. The quantitative results of SHG-positive areas showed accumulation of fibrillar collagen in Col1a2-Cre kidneys, which was ameliorated in Txndc5^cKO mice (n = 3). Scale bar: 50 μm. Data are representative of 3 or more independent experimental replicates. For all panels, data are presented as mean ± SEM. The statistical significance of differences among 3 or more groups was determined using 1-way ANOVA, followed by Sidak’s post hoc tests. **P < 0.01, ***P < 0.001.

Figure 11. Induced deletion of Txndc5 in renal fibroblasts mitigated the progression of established kidney fibrosis.

(A) Illustration of experimental design to induce Txndc5 deletion specifically in renal fibroblasts in mouse kidneys with established fibrosis. (B) Picrosirius red staining of kidney sections from Col1a2-Cre and Txndc5^cKO mice. Ten days after UUO, Col1a2-Cre and Txndc5^cKO mice showed a similar extent of renal fibrosis prior to tamoxifen injection. Eleven days after tamoxifen treatment, the fibrotic areas more than doubled (increased from 5.1% to 10.6%) in Col1a2-Cre, but barely changed in Txndc5^cKO (changed from 5.3% to 6.9%) mouse kidneys (n = 5–6). Scale bar: 50 μm. (C) Protein expression levels of fibroblast activation marker (periostin), ECM (CCN2), and TGFBR1 in whole-kidney lysate from Col1a2-Cre and Txndc5^cKO mice (n = 4–6). (D) Schematic summary of the proposed profibrotic mechanisms by which TXNDC5 contributes to the pathogenesis of renal fibrosis. TF: transcription factor. Data are representative of 3 or more independent experimental replicates. For all panels, data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 by 2-sided t test.