Alternative splicing of uromodulin enhances mitochondrial metabolism for adaptation to stress in kidney epithelial cells

Abstract

In the kidney, cells of thick ascending limb of the loop of Henle (TAL) are resistant to ischemic injury, despite high energy demands. This adaptive metabolic response is not fully understood even though the integrity of TAL cells is essential for recovery from acute kidney injury (AKI). TAL cells uniquely express uromodulin, the most abundant protein secreted in healthy urine. Here, we demonstrate that alternative splicing generates a conserved intracellular isoform of uromodulin, which contributes to metabolic adaptation of TAL cells. This splice variant was induced by oxidative stress and was upregulated by AKI that is associated with recovery, but not by severe AKI and chronic kidney disease (CKD). This intracellular variant was targeted to the mitochondria, increased NAD+ and ATP levels, and protected TAL cells from hypoxic injury. Augmentation of this variant using antisense oligonucleotides after severe AKI improved the course of injury. These findings underscore an important role of condition-specific alternative splicing in adaptive energy metabolism to hypoxic stress. Enhancing this protective splice variant in TAL cells could become a therapeutic intervention for AKI.

Keywords: Cell biology; Hypoxia; Mitochondria; Nephrology; Protein traffic.

Figure 1. Identification of AS-UMOD.

(A) The upper panel shows the primary structure and domains of UMOD. The 4 EGF-like domains are represented by the Roman numerals I through IV. D10C, domain with conserved 10 cysteines; ZP, zona pellucida; IHP, internal hydrophobic patch; EHP, external hydrophobic patch. The lower panel shows the exon/intron structure of the UMOD gene from Refseq (NCBI database). (B and C) Sashimi plot visualizes differentially spliced exons of the UMOD transcript isolated from (B) human and (C) mouse kidneys. Each numeral on the semicircle represents the number of RNA-Seq reads. Reads indicating alternative splicing sites of exon 2 and exon 10 skipping were highlighted in blue and yellow, respectively. n = 3 for human and n = 4 for mouse kidneys. (D) Definition of abbreviation. (E) Percent-splice-in (PSI) value of AS-UMOD calculated from Nanopore long-read RNA-Seq data (n = 3 for human and n = 4 for mouse kidneys). (F) RT-PCR for AS-UMOD and C-UMOD from human kidney cDNA. (G) RT-PCR product of F was purified and subsequent Sanger sequencing confirmed the existence of AS-UMOD (exon 10 skipping UMOD) in human kidneys. (H) RT-PCR for AS-Umod and C-Umod from mouse kidney cDNA. (I) RT-PCR product of H was purified and subsequent Sanger sequencing confirmed the existence of AS-Umod (exon 10 skipping Umod) in mouse kidneys. Data are represented as mean ± SEM.

Figure 2. AKI induces AS-UMOD expression.

(A and B) Relative mRNA expression of AS-Umod and C-Umod normalized to Gapdh in IRI mice. WT mice underwent sham, mild IRI, or severe IRI surgery and were harvested 24 hours after the surgery. n = 9–10 per group. (C) Immunofluorescence of subcortical region of murine kidneys 24 hours after the surgery. n = 5 mice per group. Scale bars: 100 μm. (D) Apical membrane localization of UMOD and AS-UMOD, determined by the ratio of apical membrane: whole tubules mean signal intensity was quantified using ImageJ (NIH). n = 20 tubules from 5 mild IRI kidneys for each group. (E and F) Relative mRNA expression of AS-Umod (E) and C-Umod (F) normalized to Gapdh in LPS-induced AKI mice. 5 mg/kg LPS was injected via intraperitoneal injection and mice were harvested 24 hours after injection. n = 6 per group. (G and H) Relative mRNA expression of AS-Umod (G) and C-Umod (H) normalized to Gapdh in cisplatin-induced AKI mice. 20 mg/kg cisplatin was injected via intraperitoneal injection and mice were harvested 72 hours after injection. n = 6 per group. (I and J) Relative mRNA expression of AS-UMOD (I) and C-UMOD (J) normalized to NKCC2 in human kidney samples from the KPMP. n = 7–12 per group. (K) Relative mRNA expression of AS-Umod and C-Umod normalized to Gapdh in MKTAL cells treated with various concentrations of hydrogen peroxide (H₂O₂) for 6 hours. n = 4 per group. (L) Relative mRNA expression of AS-Umod, C-Umod, and Glut1 normalized to Hprt in hypoxia conditions. MKTAL cells were cultured in control (normoxia) or hypoxia conditions for 6 hours. n = 4 per group. Data were analyzed by unpaired t test (between 2 conditions, D–H, and L) or 1-way ANOVA with embedded comparisons between 2 individual groups (among multiple conditions, A, B, and I–K) and are represented as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Figure 3. AS-UMOD is a cytoprotective intracellular isoform of UMOD.

MDCK cells stably expressing C-UMOD or AS-UMOD were established by lentiviral transduction. (A and B) Immunoblotting of UMOD in MDCK cell lysate and medium, respectively. Coomassie staining was used as a loading control for medium. n = 4. (C) Immunofluorescence of C-UMOD and AS-UMOD in MDCK cells. n = 3. Scale bar: 10 μm. (D) Immunofluorescence of UMOD C148W, an ADTKD-causing mutant in MDCK cells. n = 3. Scale bar: 10 μm. (E) Relative mRNA expression of ER stress–related genes normalized to GAPDH expression. n = 3. (F) LDH assay in MDCK cells. MDCK cells were cultured in normoxia or hypoxia conditions for 6 hours. LDH concentration in the media was measured and normalized to total cell number. n = 3. Data were analyzed by unpaired t test (between 2 conditions, B) or 1-way ANOVA with embedded comparisons between 2 individual groups (among multiple conditions, E and F) and are represented as mean ± SEM. *P < 0.05; ***P < 0.001; ****P < 0.0001.

Figure 4. AS-UMOD enhances mitochondrial energy generation.

(A) Immunoblotting of MDCK cells expressing C-UMOD or AS-UMOD after subcellular fractionation. The same amount of protein was applied for each fraction. The ratio of mitochondrial/total (mitochondrial, cytosolic, and membrane) UMOD expression was quantified by densitometry analysis. n = 3. (B) Immunofluorescence of MDCK cells expressing C-UMOD or AS-UMOD. Colocalization analysis between UMOD and mitochondria (Mitotracker) in MDCK cells. Manders’ tM1 represents a fraction of UMOD overlapping with mitochondria. n = 30 cells per group from 3 independent experiments. Scale bar: 10 μm. (C) ATP/ADP ratio of mitochondria isolated from MDCK cells. n = 4. (D) Mitochondrial respiration measurement in MDCK cells expressing C-UMOD or AS-UMOD using Seahorse. OCR, oxygen consumption rate; FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone. n = 3. (E) Transmission electron microscopy in MDCK cells expressing C-UMOD or AS-UMOD. Scale bar: 1 μm. Mitochondrial number per 100 μm² cell area (excluding nucleus) was quantitated. n = 18 cells for each group from 2 independent experiments. (F) Relative mRNA expression of PGC1α and NRF1 normalized to GAPDH. n = 4. (G) Interactome map of C-UMOD and AS-UMOD in MDCK cells obtained from AP-MS analysis. (H) Coimmunoprecipitation with anti-UMOD antibody to validate the AP-MS analysis. Asterisk indicates lower band is the target band for SLC25A22. n = 2. (I) The ratio of mitochondrial/cytosolic glutamate levels in MDCK cells. n = 3. (J) NAD⁺ levels normalized to protein concentration. n = 3. (K) ADP/ATP carrier-mediated ATP export after ADP addition to the isolated mitochondria from MDCK cells. n = 3. Data were analyzed by unpaired t test (between 2 conditions, A, B, D–F, and I–K) or 1-way ANOVA with embedded comparisons between 2 individual groups (among multiple conditions, C) and are represented as mean ± SEM. *P < 0.05; **P < 0.01; ****P < 0.0001.

Figure 5. Generation of Umod exon10 heterozygous knockout (Exon10^+/–) MKTAL cell line.

(A) Two sgRNAs were designed to cut the intronic region around exon 10. (B and C) Genotyping PCR (B) and subsequent Sanger sequencing of the PCR product (C) confirmed successful heterozygous knockout of Umod exon 10. (D) Relative mRNA expression of AS-Umod and C-Umod normalized to Gapdh in WT (Exon10^+/+) and Umod exon10 heterozygous knockout (Exon10^+/–) MKTAL cells. n = 4. (E and F) Immunoblotting of UMOD in cell lysate (E) and medium (F). Red asterisk corresponds to AS-UMOD. Coomassie staining was used as a loading control for medium. n = 3. (G) NAD⁺ levels normalized to protein concentration. n = 4. (H) ATP levels normalized to protein concentration. n = 4. (I) ATP/ADP ratio of mitochondria isolated from Exon 10^+/+ and Exon 10^+/– MKTAL cells. n = 4. Data were analyzed by unpaired t test and are represented as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 6. Cotransduction of C-UMOD and AS-UMOD.

HA-C-UMOD and Myc-AS-UMOD were cotransduced to MDCK cells and their interaction was evaluated. (A) Coimmunoprecipitation with anti-HA antibody in MDCK cells expressing MYC-AS-UMOD only (lane 1) or an equal amount of HA-C-UMOD and Myc-AS-UMOD (lane 2). CANX was used as a positive control of coimmunoprecipitation. Red asterisk corresponds to CANX. n = 2. (B) Intracellular ATP levels normalized to protein concentration. n = 4. (C and D) Immunoblotting analysis of MDCK cells expressing HA-C-UMOD only (lane 1) or equal amount of HA-C-UMOD and Myc-AS-UMOD (lane 2). Densitometric analysis of HA-C-UMOD is presented. n = 3. (E) Secreted HA-C-UMOD normalized by intracellular HA-C-UMOD expression. n = 3. Data were analyzed by unpaired t test (between 2 conditions, C–E) or 1-way ANOVA with embedded comparisons between 2 individual groups (among multiple conditions, B) and are represented as mean ± SEM. *P < 0.05.

Figure 7. Identification of SSOs that induce AS-UMOD in MKTAL cells.

(A) Design of SSOs to induce AS-Umod expression. Numbers in parentheses indicate position from the first base of exon 10. (B) Relative mRNA expression of AS-Umod normalized to Gapdh in MKTAL cells transfected with 30nM SSOs for 24 hours. Lipofectamine alone and nontargeted SSO were used as negative controls. n = 3. (C) Relative mRNA expression of AS-Umod and C-Umod normalized to Gapdh in MKTAL cells transfected with various concentrations of Umod SSO for 48 hours. Umod SSO corresponds to SSO (–13). n = 3. (D–I) MKTAL cells were treated with 30 nM scrambled SSO or Umod SSO for 48 hours. (D and E) Immunoblotting of UMOD in cell lysate and medium, respectively. Red asterisk corresponds to AS-UMOD. Coomassie staining was used as a loading control for medium. n = 4. (F) Immunofluorescence of UMOD. n = 2. Scale bar: 10 μm. (G) The ratio of mitochondrial/cytosolic glutamate levels. n = 3. (H) NAD⁺ levels normalized to protein concentration. n = 3. (I) ATP levels normalized to protein concentration. n = 3. Data were analyzed by unpaired t test (between 2 conditions, E and G–I) or 1-way ANOVA with embedded comparisons between 2 individual groups (among multiple conditions, C) and are represented as mean ± SEM. *P < 0.05; **P < 0.01; ****P < 0.0001.

Figure 8. AS-UMOD induction protects TAL cells and ameliorates severe IRI.

WT mice underwent severe IRI and SSO treatment (25 mg/kg) and were harvested 72 hours after IRI. (A) Schematic of experimental design. (B) Relative mRNA expression of AS-Umod and C-Umod normalized to Gapdh. n = 10–11 per group. (C) Immunofluorescence of murine kidneys. White arrows indicate AS-UMOD, which is induced in the cytosol of TAL cells after Umod SSO treatment. n = 4 mice per group. Scale bars: 50 μm. (D) Serum creatinine and urea concentration. n = 14–16 per group. (E) PAS-stained kidney sections and quantification of injury. n = 9–11 per group. Scale bar: 500 μm. (F) Relative mRNA expression of injury-related genes normalized to Gapdh in the whole kidney. n = 10–11 per group. (G and H) Primary TAL cells were isolated by magnetic cell separation, and cells unbound to the beads were defined as non-TAL cells. (G) Relative mRNA expression of injury-related genes normalized to Gapdh in TAL and non-TAL cells. n = 7–8 per group. (H) ATP levels normalized to protein concentration in TAL and non-TAL cells. n = 7–8 per group. Data were analyzed by unpaired t test and are represented as mean ± SEM. *P < 0.05; ****P < 0.0001.

Figure 9. Graphical visualization of alternative splicing of UMOD.

C-UMOD is a GPI-anchored protein and is sorted to the plasma membrane. C-UMOD regulates the activities of membrane transporters and maintains extracellular homeostasis once secreted into the extracellular region. AKI induces alternative splicing of UMOD and generates AS-UMOD, a non-GPI anchored isoform. AS-UMOD showed preferential localization in the mitochondria compared with C-UMOD, facilitating mitochondrial energy generation as a metabolic adaptation to cellular injury. However, mitochondrial localization of AS-UMOD remains partial, and we cannot exclude the possibility that AS-UMOD in ER could also affect mitochondrial function. The mechanism by which a portion of AS-UMOD targets the mitochondria remains unknown. The schema was created in BioRender. Nanamatsu, A. (2025) https://BioRender.com/k97g401