TNIK depletion induces inflammation and apoptosis in injured renal proximal tubule epithelial cells

Abstract

In the aftermath of acute kidney injury (AKI), surviving proximal tubule epithelia repopulate injured tubules to promote repair. However, a portion of cells fail to repair [termed failed-repair proximal tubule cells (FR-PTCs)] and exert ongoing proinflammatory and profibrotic effects. To better understand the molecular drivers of the FR-PTC state, we reanalyzed a mouse ischemia-reperfusion injury single-nucleus RNA-sequencing (snRNA-seq) atlas to identify Traf2 and Nck interacting kinase (Tnik) to be exclusively expressed in FR-PTCs but not in healthy or acutely injured proximal tubules after AKI (2 and 6 wk) in mice. We confirmed expression of Tnik protein in injured mouse and human tissues by immunofluorescence. Then, to determine the functional role of Tnik in FR-PTCs, we depleted TNIK with siRNA in two human renal proximal tubule epithelial cell lines (primary and immortalized hRPTECs) and analyzed each by bulk RNA-sequencing. Pathway analysis revealed significant upregulation of inflammatory signaling pathways, whereas pathways associated with differentiated proximal tubules such as organic acid transport were significantly downregulated. TNIK gene knockdown drove reduced cell viability and increased apoptosis, including differentially expressed poly(ADP-ribose) polymerase (PARP) family members, cleaved PARP-1 fragments, and increased annexin V binding to phosphatidylserine. Together, these results indicate that Tnik upregulation in FR-PTCs acts in a compensatory fashion to suppress inflammation and promote proximal tubule epithelial cell survival after injury. Modulating TNIK activity may represent a prorepair therapeutic strategy after AKI.

NEW & NOTEWORTHY The molecular drivers of successful and failed repair in the proximal tubule after acute kidney injury (AKI) are incompletely understood. We identified Traf2 and Nck interacting kinase (Tnik) to be exclusively expressed in failed-repair proximal tubule cells after AKI. We tested the effect of siTNIK depletion in two proximal tubule cell lines followed by bulk RNA-sequencing analysis. Our results indicate that TNIK acts to suppress inflammatory signaling and apoptosis in injured renal proximal tubule epithelial cells to promote cell survival.

INTRODUCTION

Epithelial cells line the segmented nephron from podocytes in the glomerulus to principal cells in the collecting duct. Proximal tubule epithelial cells perform bulk reabsorption of sodium, glucose, potassium, and water among many other substances, making them both highly metabolically active and also susceptible to injury. Several acute insults such as ischemia can cause tubular injury leading to acute kidney injury (AKI), a clinical syndrome characterized by a rapid loss of kidney function (1). Risk factors for developing AKI include diabetes, hypertension, chronic kidney disease (CKD), pregnancy, and exposure to environmental toxins such as lead (2, 3). Nephron injury or total nephron loss can occur depending on the severity of damage. Entirely new nephrons cannot regenerate after severe damage, and there are currently no targeted therapeutics to treat AKI. However, lineage tracing studies in mice revealed that injured epithelial cells can successfully repair after acute injury, repopulating tubular epithelia and restoring nephron function (4, 5).

Previous single-nucleus RNA-sequencing (snRNA-seq) studies in our laboratory defined the transcriptional signature of a distinct population of proximal tubule epithelial cells that fail to repair after bilateral ischemia-reperfusion injury (IRI) that we have termed FR-PTCs (6). In other work, this cell population has been alternatively named adaptive PT, which consists of successful or maladaptive repair states, mapping to and correlating with FR-PTCs (7, 8). The transcriptional signature of FR-PTCs reflects a proinflammatory and profibrotic phenotype, which may contribute to unsuccessful repair, local tissue inflammation, fibrosis, and the AKI-to-CKD transition (6). Among the unique marker genes characterizing the FR-PTC state we identified the kinase Traf2 and Nck Interacting Kinase (Tnik). To date, there are no reports describing any role for TNIK in the kidney.

TNIK is a member of the sterile 20 (STE20) family of kinases, which consist of two major branches: p21-activated kinase (PAK) I and II and germinal center kinase (GCK) I–VI and VIII (9). TNIK itself belongs to the GCK IV family, which possess highly conserved NH₂-terminal kinase and COOH-terminal GCK homology domains and differ from PAKs, which have COOH-terminal kinase and NH₂-terminal p21 GTPase-binding domains (9). The intermediate domain of GCK IV family members like TNIK is less conserved, which may allow for unique protein binding interactions. For instance, TNIK interacts with both TRAF2 and NCK as well as the RAS family small GTP-binding protein RAP2A (10, 11). TNIK plays a role in regulating the cytoskeleton, as its overexpression in several cell lines leads to disruption of F-actin structure and inhibition of cell spreading (11). Furthermore, as an effector of RAP2A, TNIK not only has been implicated in regulating the actin cytoskeleton but has also been reported to be involved in brush border formation in intestinal epithelial cells (10, 12).

Additional reported roles of TNIK include activation of Wnt target genes through the formation of a transcriptional complex with TCF4 and β-catenin in proliferating intestinal crypts (13). In other contexts, TNIK regulates the proliferation and survival of cancerous cells, which has served as rationale to develop several small molecules to inhibit the kinase activity of TNIK (14, 15). Along with TNIK, other STE20 GCK family kinases such as MAP4K4 and MINK have also been implicated in regulating cellular stress and inflammatory signals upstream of JNK and NF-κB cascades, respectively, in various cell types (16, 17). TNIK has been shown to activate JNK in a dose-dependent manner in Phoenix-A cells but not NF-κB (10, 11). In primary B cells infected with Epstein–Barr tumor virus, however, TNIK silencing prevents the oncoprotein LMP1 (latent member protein 1) activation of NF-κB and JNK (17). Similarly, Tnik silencing in neurons blocks DLK (dual leucine zipper kinase) activation of stress-induced JNK signaling after nerve growth factor deprivation (16). Roles for TNIK in transforming growth factor (TGF)-β signaling and fibrosis have also been described. TNIK silencing in hepatic stellate cells leads to disruption of TGF-β-induced procollagen I and fibronectin expression, and TNIK is upregulated in human cirrhotic tissues and mouse models of liver fibrosis (18). TGF-β treatment has also been shown to induce nuclear translocation of TNIK (19).

In this study, we aimed to define the role of Tnik in FR-PTCs. We confirmed that TNIK protein is expressed after mouse and human AKI. Then we characterized two different proximal tubule epithelial cell lines that express TNIK to study its function in culture. We performed TNIK silencing followed by bulk RNA-sequencing analysis to understand molecular pathways and genes modulated by TNIK silencing. We found that TNIK silencing resulted in a proinflammatory and dedifferentiated cellular state that reduced PT cell viability and induced apoptosis. Targeting TNIK activity in proximal tubule epithelial cells may represent a novel therapeutic strategy to improve AKI outcomes.

MATERIALS AND METHODS

Surgery

As described in our previous report (6), mice underwent treatments and protocols approved by the Animal Care and Use Committee at Washington University in St. Louis. Briefly, 8- to 10-wk-old male mice were anesthetized with isoflurane and subcutaneously administered 1 mg/kg slow-release buprenorphine for analgesic support. Bilateral flank incisions were made to access the kidneys, and ischemia-reperfusion injury was induced with the placement of atraumatic microaneurysm clamps (Roboz; RS-5422) at the renal pedicle and their subsequent removal after 18.5 min. Control mice underwent the same surgical procedure without clamping of the renal pedicle. Throughout the procedure, body temperature was monitored and maintained at 36.5–37.5°C with a homeothermic monitoring system (Harvard Apparatus). Absorbable suture was used to close the peritoneal layer, and flank incisions were closed and secured with autoclip wound clips. Kidneys were harvested at 4 h, 12 h, 2 days, 14 days, and 6 wk after surgery and stored for downstream processing.

Cell Lines

Primary human renal proximal tubule epithelial cells (hRPTECs) were acquired from Lonza (CC-2553) at passage 2 and expanded once in REGM Renal Epithelial Cell Growth Medium BulletKit (Lonza CC-3190) until reaching 90% confluence. Primary hRPTEC stocks were then cryopreserved after expansion. For experimental assays, cells were thawed once, cultured, and passaged once directly into plates. Assays were not performed beyond the experimental passage. Human telomerase reverse transcriptase (hTERT)-immortalized RPTECs were acquired from ATCC (CRL-4031) and expanded in hTERT-immortalized RPTEC growth medium, which consists of DMEM-F-12 medium supplemented with the hTERT RPTEC growth kit (ACS-4007). HEK293T cells were maintained in DMEM (Thermo Fisher Scientific; 11-965-118) supplemented with 10% heat-inactivated fetal bovine serum (Sigma; F4135) and 1% penicillin-streptomycin (Thermo Fisher Scientific; 15140-122).

siRNA and Plasmid Transfections

siRNA transfection of both primary and immortalized RPTECs was facilitated with Lipofectamine RNAiMAX (Thermo Fisher Scientific; 13778075). The siRNA-lipid complex was prepared with Opti-MEM (Thermo Fisher Scientific; 31985-070) according to the manufacturer’s protocol and used to treat overnight cultures of primary or hTERT-immortalized hRPTECs seeded in six-well plates at 1 × 10⁵ or 4 × 10⁵ cells/well, respectively. A Countess II (Thermo Fisher Scientific) automated cell counter was used to count cells before seeding. Cells were incubated with 50 nM siGENOME nontargeting siRNA #5 (Horizon Discovery; D-001210-05-05) or 50 nM pooled siGENOME human TNIK siRNA (Horizon Discovery; M-004542-03-0020). After 24 h of incubation, the culture medium was replaced with fresh REGM or hTERT-immortalized RPTEC growth medium and cells were harvested for real-time qPCR, transcriptomic, or immunoblotting analysis at either 48 or 72 h after siRNA-lipid complex addition. To generate TNIK-overexpressing lysates for immunoblotting positive controls, a c-Flag-tagged human TNIK gene open reading frame (ORF) cDNA clone expression plasmid (Sino Biological; HG13036-CF) was transfected into HEK293T cells with Lipofectamine for 24 h.

Immunoblot Procedures and Analysis

Whole cells were lysed directly in six-well plates with 2× sodium dodecyl sulfate (SDS) sample buffer (0.125 M Tris·HCl pH 6.8, 4% SDS, 0.15 M DTT, 20% glycerol, 0.1% bromophenol blue). Lysate protein concentrations were measured on the FlexStation 3 multimode plate reader (Molecular Devices) with the Quant-IT Protein Assay Kit (Thermo Fisher Scientific; Q33210). Equal amounts of protein lysates were resolved on 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis handcasted gels and transferred under wet conditions at 4°C to Immobilon-P polyvinylidene difluoride (Millipore Sigma; IPVH00010) membranes. After transfer, membranes were dried, reactivated in methanol, rinsed in 1× TBS, and blocked in 3% BSA or 5% nonfat dry milk prepared with 0.05% TBS-Tween 20 or 0.1% TBS-Tween 20, respectively. After blocking, membranes were then probed overnight at 4°C with rotation using primary antibodies against rabbit polyclonal Tnik (GeneTex; GTX13141, 1:1,000), rabbit polyclonal poly(ADP-ribose) polymerase (Parp) (Cell Signaling Technology; 9542, 1:1,000), or mouse monoclonal γ-Tubulin (Sigma-Aldrich; T6557, 1:1,000). The following day membranes were washed three times in 1× 0.05% or 1× 0.1% TBS-Tween 20 and incubated with secondary rabbit (Cytiva; NA934, 1:2,000) or mouse (Cytiva; NA931, 1:4,000) horseradish peroxidase (HRP)-linked whole antibodies. Membranes were then washed three times with 0.05% TBS-Tween 20, rinsed once with 1× TBS, and then incubated in ECL Prime Western Blotting Detection Reagent (Cytiva; RPN2236). Finally, membranes were exposed with a ChemiDoc MP Imaging system (Bio-Rad). ImageJ2 was used for densitometry analysis, whereby band densities were plotted and the resulting peaks enclosed and quantified with the line and wand tools, respectively.

Real-Time RT-PCR

RNA was harvested directly from six-well cell culture plates and isolated with the RNeasy Mini Kit (Qiagen; 74104) according to the manufacturer’s protocol. After nanodrop (DeNovix DS-11+ Spectrophotometer) quantification, 500 ng of total RNA was used to synthesize cDNA with a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems; 4368813). iTaq Universal SYBR Green Supermix (Bio-Rad; 1725121) was then added to cDNA to perform quantitative real-time PCR with a real-time PCR detection system (Bio-Rad CFX Connect). Analyses of samples were performed in triplicate, and human GAPDH was used to normalize gene expression. Fold change over samples transfected with scramble siRNA (scr-siRNA) was calculated with the ΔΔC_T method (where C_T is threshold cycle). The following human primer sequence pairs were used: TNIK, 5′- AAGGTAACACGTTGAAAGAGGAG-3′ (forward) and 5′- AGTCAGCAAGACATTTTGCCC-3′ (reverse); vascular adhesion molecule-1 (VCAM-1), 5′-GGGAAGATGGTCGTGATC CTT-3′ (forward) and 5′-TCTGGGGTGGTCTCGATTTTA-3' (reverse); and GAPDH, 5′- GACAGTCAGCCGCATCTTCT-3′ (forward) and 5′- GCGCCCAAT ACGACCAAATC-3′ (reverse).

snRNA-Seq Analysis

The snRNA-seq dataset from our previous study (GSE139107) was reanalyzed. This dataset contains a series of measurements of gene expression in the kidneys of mice that were subjected to bilateral ischemia-reperfusion injury at 4 h, 12 h, 2 days, 2 wk, and 6 wk. To ensure consistency with our previous study, the entire analysis was performed on the complete Seurat object that was published previously. This allowed us to maintain the same clustering information. plot1cell R package (https://github.com/TheHumphreysLab/plot1cell) deposited in our GitHub repository was used to visualize the gene expression in proximal tubule cells from each time point.

Library Preparation, Sequencing, and Bulk RNA-Seq Analysis

RNA was harvested directly from six-well cell culture plates with the RNeasy Mini Kit (Qiagen; 74104) and submitted to the Genome Technology Access Center. The quality of the total RNA was assessed with the Agilent 4200 Tapestation. For library preparation, 500 ng to 1 μg of total RNA was used. To remove ribosomal RNA, the RNase-H method with RiboErase kits (Kapa Biosystems) was employed. The mRNA was then fragmented by heating it to 94°C for 8 min in reverse transcriptase buffer. cDNA was synthesized from the mRNA with the SuperScript III RT enzyme (Life Technologies) and random hexamers according to the manufacturer’s instructions. A second-strand reaction was performed to produce double-stranded cDNA. The cDNA was modified by adding an A base to the 3′ ends and ligating Illumina sequencing adapters to the ends. The ligated fragments were amplified for 15 cycles with primers with unique dual index tags. The fragments were sequenced on a NovaSeq-6000 (Illumina) using paired end reads of 150 bases in length. Basecalling and demultiplexing were performed with Illumina’s bcl2fastq software, with a maximum of one mismatch allowed in the indexing read during demultiplexing. The resulting fastq files were processed with an in-house pipeline available at https://github.com/HaojiaWu/Bulk_RNAseq_analysis_pipeline. In brief, RNA-seq reads were aligned to the GRCh38 genome with STAR version 2.7.3a. Gene counts were obtained by counting the number of uniquely aligned and unambiguous reads with Subread:featureCount version 2.0.3. The count matrices were then input into edgeR for data normalization and differential gene analysis. Pathway analysis was performed on the ToppGene suite (https://toppgene.cchmc.org) using the differentially expressed genes (DEGs) derived from pairwise comparison as input. ggplot2 R package was used for data visualization.

Human Tissue Procurement

This research complies with all relevant ethical regulations and has been approved by the Washington University Institutional Review Board. Kidney cortex biopsies were obtained either from patients undergoing partial or radical nephrectomy for renal mass or from kidneys rejected for transplantation.

Immunofluorescence Procedures and Analysis

Mouse kidney or human (54-yr-old female donor with serum Cr 3.5 at harvest; Fig. 1E) kidney cortex samples were fixed with 10% formalin overnight, embedded in paraffin, and cut at 5-µm thicknesses. Antigen retrieval in antigen unmasking solution (Vector Laboratories; H-3301-250) was performed before staining. The sections were blocked for 30 min with 1% bovine serum albumin in PBS at room temperature and incubated with primary antibodies against rabbit polyclonal Tnik (Sigma-Aldrich; HPA012128, 1:200) and goat polyclonal Kidney Injury Molecule-1 (Kim-1) (R&D; AF1817, 1:400) at 4°C overnight. Sections were incubated with Lotus tetragonolobus lectin (LTL) (Vector Laboratories; FL-1321-2, 1:400) and secondary antibodies conjugated with Cy3 (Jackson ImmunoResearch; 111-167-003, 1:400) or Cy5 (Jackson ImmunoResearch; 705-175-147, 1:400). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (Thermo Fisher Scientific; D1306). Sections were mounted in Prolong Gold antifade reagent (Thermo Fisher Scientific; P36930). Human primary or hTERT-immortalized RPTECs were seeded in four-well glass chamber slides (Millipore Sigma; PEZGS0416) at either 2 × 10⁴ or 1.2 × 10⁵ cells/chamber, respectively. Cells were grown for 72 h and then fixed in 4% paraformaldehyde (PFA). After fixation, cells were washed with 1× PBS and then permeabilized for 15 min at room temperature with permeabilization solution (0.1% Triton X-100 in 1× PBS). Cells were then blocked for 1 h at room temperature in blocking buffer (1% BSA in 1× PBS) and then incubated overnight at 4°C with rabbit polyclonal anti-Tnik (GeneTex; GTX13141, 1:100) or mouse monoclonal Vcam-1 (Thermo Fisher Scientific, Invitrogen; MA5-11447, 1:200). The next day cells were washed with wash buffer (0.05% Tween 20 in 1× PBS) and then incubated in secondary donkey anti-rabbit AlexaFluor 488 (Thermo Fisher Scientific; A-21206, 1:200) and donkey anti-mouse AlexaFluor 568 (Thermo Fisher Scientific; A-10037, 1:200) for 1 h at room temperature. Cells were then counterstained for 10 min with NucBlue (Thermo Fisher Scientific; R37605) and washed with wash buffer and 1× PBS. Chamber slides were then mounted with Prolong Gold antifade reagent (Thermo Fisher Scientific; P36930) and allowed to cure overnight. Imaging analysis was performed with a Nikon Eclipse Ti microscope using the widefield fluorescence laser scan confocal modality. ImageJ2 was used to analyze immunocytochemistry studies.

Figure 1.

Single-nucleus (sn) RNA-sequencing (snRNA-seq) detection of Traf2 and Nck interacting kinase (Tnik) exclusively in failed-repair proximal tubule (PT) cells (FR-PTCs). A: snRNA-seq detected upregulated Tnik (a serine-threonine kinase) expression at 14 days and 6 wk after acute kidney injury (AKI) in mice. UMAP, Uniform Manifold Approximation and Projection. B: dot plot displaying Tnik expression along with proinflammatory/profibrotic mediators in FR-PTCs after AKI. C: dot plot displaying increased expression of Tnik in FR-PTCs after unilateral ureteral obstruction (UUO) in mice (20). D: dot plot displaying increased expression of Tnik in injured PT; note that Tnik expression remains elevated 6 mo after AKI in mice (21). E: immunofluorescence studies confirm TNIK expression in dedifferentiated [Lotus tetragonolobus lectin (LTL) negative] and injured [kidney injury molecule-1 (Kim-1) positive] proximal tubules in both mouse and human AKI tissues. IRI, ischemia-reperfusion injury. GFP, green fluorescent protein (D). A and B: n = 3 mice per condition. E: n = 1 (mouse) and n =1 (human). A ×60 objective was used to acquire images in E.

Flow Cytometry

Primary hRPTECs were treated with 0.05% trypsin-EDTA at 37°C for 5 min and collected. After blocking with normal mouse serum (Thermo Fisher Scientific; 31881) for 10 min on ice, cells were stained with anti-human VCAM-1 antibody (Thermo Fisher Scientific; 13106982), followed by PE Streptavidin (BioLegend; 405203) in 1× HBSS (Thermo Fisher Scientific; 14065056) containing 1% bovine serum albumin (Sigma-Aldrich; 3116956001) and 0.035% NaHCO₃ (Sigma-Aldrich; S6014) on ice. Stained cells were analyzed with a FACS Canto II (BD Biosciences). Data analyses were performed with FlowJo software (BD Biosciences).

Cell Viability and Apoptosis Assays

The RealTime-Glo Cell Viability Assay (Promega; G9711) was used to measure cell viability in pooled siRNA targeting human TNIK (siTNIK)-depleted primary hRPTECs. Cells were transfected with siGENOME siTNIK or scr-siRNA (time = 0). Cell Viability detection reagents were prepared according to the manufacturer’s instructions and added at the same time as the siGENOME siRNA (time = 0). Luminescence was measured every 24 h for 72 h with a FlexStation 3 multimode plate reader (Molecular Devices). The RealTime-Glo Annexin V Apoptosis and Necrosis Assay (Promega; JA1011) was used to detect apoptosis. Cells were transfected with siGENOME siTNIK or scr-siRNA (time = 0). The apoptosis reagents were prepared according to the manufacturer’s instructions and added just before the first 24 h after siGENOME siRNA treatment. Luminescence and fluorescence were measured every 24 h for 120 h with a FlexStation 3 multimode plate reader (Molecular Devices).

Statistical Analysis

Densitometry analysis of immunoblots and quantification of immunocytochemistry data were performed with ImageJ2 (version 2.14.0/1.54f). Prism 10 for macOS [version 10.0.1(170)] was used to perform unpaired t tests, multiple t tests (corrected for multiple comparisons with the Holm–Šídák method), and ordinary one-way ANOVA (corrected for multiple comparisons with the Dunnett method). Statistical significance was defined as P < 0.05 and are reported in GraphPad style. Error bars represent standard deviation.

RESULTS

snRNA-Seq Reveals Exclusive Expression of Tnik in FR-PTCs

In this study, we confirmed induction of acute kidney injury following our protocol for bilateral ischemia-reperfusion injury (b-IRI) in mice (Supplemental Fig. S1). Previous snRNA-seq studies in our laboratory identified a distinct population of cells that fail to repair, FR-PTCs, after acute kidney injury in mice (6). We detected Tnik expression in FR-PTCs 14 days after b-IRI, which persisted for 6 wk, though at lower levels relative to day 14 (Fig. 1A). Blood urea nitrogen levels remained elevated relative to baseline at time points of Tnik expression (6). Histological alterations included a reduction of differentiated proximal tubules and expansion of the renal interstitium at 14 days and 6 wk after b-IRI, which was assessed by both immunofluorescence and periodic acid-Schiff staining (6). In contrast to healthy and acutely injured proximal tubule cells, FR-PTCs not only exclusively express Tnik but also express profibrotic and proinflammatory mediators such as Traf2, Tgfbr2, Col4a1, Pdgfb, Nfkb1, Ccl2, and the injury marker Cd44 (Fig. 1B).

Analysis of two independent single-cell time course studies also confirmed upregulation of Tnik in animal models of fibrosis and inflammation. In a unilateral ureteral obstruction (UUO)/renal fibrosis time course study analyzed by single-cell combinatorial indexing RNA-sequencing, Tnik expression was detected 10 days after UUO (Fig. 1C) (20). An AKI single-nucleus multiomics time course study detected Tnik expression in injured PT 4 wk after ischemia-reperfusion injury, which persisted for 6 mo (Fig. 1D) (21). Although examined at different experimental time points, these studies highlight that PT cells expressing Tnik persist several weeks to months after initial injury.

Using immunofluorescence, we confirmed Tnik to be coexpressed with injured (Kim-1 positive) and dedifferentiated (LTL negative) proximal tubules (Fig. 1E) in mouse kidneys 6 wk after b-IRI. TNIK also coexpressed with dedifferentiated proximal tubules in human AKI tissue (Fig. 1E). Additionally, snRNA-seq analysis confirmed that FR-PTCs coexpressed Tnik and vascular adhesion molecule-1 (Vcam-1), a well-described FR-PTC marker (Fig. 1B) (6, 22).

Transcriptomic Response of TNIK Silencing in hRPTECs

To better understand potential roles for FR-PTCs in AKI recovery and the AKI-to-CKD transition, we next surveyed proximal tubule cell lines for TNIK expression. We found that human primary RPTECs, when cultured on hard plastic, are heterogeneous, with some cells adopting an injury phenotype and expressing TNIK as well as the FR-PTC marker VCAM-1 (Fig. 2A). In primary hRPTECs, TNIK localized to the cytoplasm and VCAM-1 localized to the cell surface membrane. We also found that VCAM-1 coexpressed with TNIK in the cytoplasm of primary hRPTECs (Fig. 2B). We detected significantly more TNIK⁺ (33.49%) primary hRPTECs relative to VCAM-1⁺ (12.54%) and TNIK⁺-VCAM-1⁺ (9.42%) primary hRPTECs in culture (Fig. 2C). Differentiation markers including ASS1 (23) and PDZK1 (8) as well as hypoxia markers HIF1A (24, 25) and PDK1 (26) were also detected in primary hRPTECs by bulk RNA-seq (Supplemental Fig. S2, A and B). Similar to primary hRPTECs, we detected TNIK in hTERT-RPTECs by immunocytochemistry (Fig. 2E). As with primary hRPTECs, TNIK localized to the cytoplasm (Fig. 2E). In contrast to primary hRPTECs, we could not detect VCAM-1 in hTERT-RPTECs by bulk RNA-seq or immunocytochemistry (Fig. 2, D and E).

Figure 2.

Detection of Traf2 and Nck interacting kinase (TNIK) and vascular adhesion molecule-1 (VCAM-1) in primary human renal proximal tubule epithelial cells (hRPTECs). A: TNIK as well as VCAM-1 are detected by RNA-sequencing (RNA-seq) in hRPTECs. B, top: in primary hRPTECs, TNIK is localized to the cytoplasm, whereas VCAM-1 is localized to the cell membrane. B, bottom: VCAM-1 is also found localized to the cytoplasm along with TNIK expression in primary hRPTECs. C: quantification of TNIK⁺, VCAM-1⁺, and TNIK⁺/VCAM-1⁺ cells in primary hRPTEC cultures. D and E: TNIK is also expressed in human telomerase reverse transcriptase (hTERT)-RPTECs and localizes to the cytoplasm. VCAM-1 is undetectable by both RNA-seq and immunocytochemistry in hTERT-RPTECs. Immunocytochemistry data are reported as the mean percentage of the total cell count (4,567); ****P < 0.0001 vs. TNIK⁺. For C, n = 8 biological replicates (4 or 5 ×20 images were analyzed per replicate) and error bars represent 1 SD from the mean. For B and E, a ×60 objective was used to acquire images at top and a ×2 digital zoom was applied to the ×60 objective to acquire images at bottom. n = 2 biological replicates for A and D.

To investigate functional roles for Tnik, we next optimized an siRNA knockdown strategy. Primary hRPTECs were transfected with either a nontargeting siRNA (scramble siRNA, scr-siRNA) or with pooled siRNA targeting human TNIK (siTNIK) for 48 or 72 h. We observed a 79% and 97.5% knockdown of TNIK by qPCR and immunoblotting analysis at 72 h, respectively (Fig. 3, A and C). We additionally performed siRNA-mediated knockdown of TNIK in hTERT-RPTECs. At 72 h, TNIK expression was suppressed by 75% and 96.8% as confirmed by qPCR and immunoblotting analysis, respectively (Fig. 3, B and D). We then harvested RNA samples from TNIK-depleted primary hRPTECs (48 and 72 h) and hTERT-RPTECs (72 h) and performed bulk RNA-seq on n = 6 individual samples per condition. RNA-seq data were analyzed with our in-house script (see materials and methods). Clustering analysis in hRPTECs revealed two distinct major clusters, scr-siRNA and siTNIK (Fig. 4A). In response to 48 and 72 h of TNIK silencing in primary hRPTECs, 894 and 2,041 differentially expressed genes were detected, respectively (log fold change greater than 1 or less than −1, adjusted P value < 0.05), whereas 495 differentially expressed genes were identified in hTERT-RPTECs at the 72 h time point (log fold change greater than 1 or less than −1, adjusted P value < 0.05) (Fig. 4B). Five hundred seventeen shared genes were detected between 48-h and 72-h siTNIK primary hRPTECs, with 205 genes upregulated and 312 genes downregulated (Fig. 4C). Two hundred eighty shared genes between 72-h siTNIK primary hRPTECs and hTERT-RPTECs were detected, with 192 genes upregulated and 88 genes downregulated (Fig. 4D).

Figure 3.

siRNA-mediated silencing of Traf2 and Nck interacting kinase (TNIK) in human renal proximal tubule epithelial cells (hRPTECs). Primary and human telomerase reverse transcriptase (hTERT)-RPTECs were transfected with scramble siRNA (scr-siRNA) or pooled siRNA targeting human TNIK (siTNIK) for 72 h. TNIK silencing was confirmed by both qPCR (A and B) and immunoblotting (C and D) before RNA-sequencing (RNA-seq) analysis. qPCR data are expressed as the mean fold change relative to scr-siRNA; GAPDH was used as an internal control; **P = 0.0018 vs. scr-siRNA, ***P = 0.0003 vs. scr-siRNA. Immunoblotting data are expressed as the mean ratio of TNIK to γ-Tubulin; **P = 0.0037 vs. scr-siRNA, ****P < 0.0001 vs. scr-siRNA. For both assays, n = 3 biological replicates, and error bars represent 1 SD from the mean. L, ladder; PC, positive control.

Figure 4.

RNA-sequencing (RNA-seq) clustering and analysis of Traf2 and Nck interacting kinase (TNIK)-depleted human renal proximal tubule epithelial cells (hRPTECs). A: TNIK-depleted primary hRPTECs and human telomerase reverse transcriptase (hTERT)-RPTECs cluster in Uniform Manifold Approximation and Projection (UMAP) space. B–D: both hRPTEC cell lines respond to TNIK silencing by up- or downregulating genes that are shared across time and cell line. A portion of genes modulated by TNIK silencing are also unique to their respective time point and cell line. n = 6 biological replicates per each condition. Log fold change greater than 1 or less than −1, adjusted P < 0.05.

We first focused our analysis on shared DEGs between the 48 and 72 h time points in siTNIK-transfected primary hRPTECs. TNIK depletion resulted in an enrichment of pathways associated with inflammation and fibrosis, with upregulation of Gene Ontology (GO) terms such as the type I interferon signaling pathway, cytokine binding and receptor activity, extracellular matrix and extracellular matrix-associated proteins, as well as tumor necrosis factor receptor superfamily binding (Fig. 5A). TNIK depletion caused downregulation of pathways related to proximal tubule cellular function and maturation such as organic anion transport, organic acid metabolic process and transport, and basic amino acid transmembrane transporter activity (Fig. 5B).

Figure 5.

Traf2 and Nck Interacting Kinase (TNIK) silencing induces inflammation and suppresses differentiation. A: combined Gene Ontology (GO) term analysis of day 2 and 3 TNIK-depleted primary human renal proximal tubule epithelial cells (hRPTECs) shows upregulated pathways associated with inflammation such as type I interferon signaling. B: pathways associated with proximal tubule cell differentiation are downregulated, such as organic acid transport. C: heat map highlighting shared up- and downregulated genes between days 2 and 3 in TNIK-depleted primary hRPTECs. Note downregulation of HNF-4α, which is the master regulator of proximal tubule cell differentiation. siTNIK, pooled siRNA targeting human TNIK.

Specific shared enriched genes between 48- and 72-h siTNIK-transfected primary hRPTECs associated with the identified upregulated GO terms include FGF1, IFIT1, IFI6, IFI27, IL7R, TNFSF15, COL11A1, COL12A1, COL13A1, as well as NOG, whereas specific shared depleted genes associated with downregulated GO terms include SLC47A1, ASS1, SLC30A2, SLC3A1, HNF4A, and SLC38A3. When we examined DEGs that were shared between primary hRPTECs and hTERT-RPTECs at 72 h, we found similar genes and pathways. For example, like the 48- and 72-h siTNIK-transfected primary hRPTECs, 72-h hTERT-RPTECs and primary hRPTECs also share upregulation of inflammatory and fibrotic pathways such as cellular response to type I interferon as well as positive regulation of I-kB kinase/NF-kB signaling (Supplemental Fig. S3A) and downregulation of pathways related to renal proximal tubule epithelial cell differentiation and function such as basic amino acid transmembrane transporter activity (Supplemental Fig. S3B). Similar genes were shared as well between siTNIK-transfected 72-h primary and hTERT-RPTECs (Supplemental Fig. S3C).

Cytokine Treatment Downregulates TNIK Expression in Primary hRPTECs

Since we observed increases in pathways associated with inflammatory signaling and cytokine binding following TNIK silencing, we tested the effect of cytokine treatment on TNIK expression in primary hRPTECs. Primary hRPTECs were treated with combined doses of 20 ng/mL IFNγ and 20 ng/mL TNFα for 24 h (27). To assess the degree of injury in primary hRPTECs after treatment, we used qPCR and flow cytometry analysis to detect gene expression and protein expression, respectively, using the failed-repair marker VCAM-1. By qPCR, we observed that VCAM-1 gene expression levels significantly increased 2.68-fold in primary hRPTECs after combined IFNγ and TNFα treatment relative to vehicle-treated cells (Fig. 6A). Flow cytometry analysis also revealed significant increases in the level of surface VCAM-1 protein expression. In vehicle-treated cells, we detected that 28.9% of cells expressed VCAM-1 (Fig. 6B). This ratio was significantly increased to 68% after combined IFNγ and TNFα treatment (Fig. 6, B and C). We then utilized these samples to test for TNIK gene expression after vehicle or combined IFNγ and TNFα treatment for 24 h. We found that TNIK gene expression levels in combined IFNγ- and TNFα-treated primary hRPTECs significantly decreased 0.571-fold relative to vehicle-treated primary hRPTECs (Fig. 6D). These results highlight an inverse relationship between TNIK expression and inflammation whereby proinflammatory signals suppress TNIK expression while TNIK expression itself exerts anti-inflammatory effects.

Figure 6.

Cytokine treatment upregulates vascular adhesion molecule-1 (VCAM-1) and downregulates Traf2 and Nck interacting kinase (TNIK) expression. Primary human renal proximal tubule epithelial cells (hRPTECs) were treated with a combination of 20 ng/mL each of IFNγ and TNFα for 24 h. qPCR analysis was used to detect VCAM-1 gene expression (A); flow cytometry was then used to detect surface levels of VCAM-1 protein expression (B and C). qPCR analysis was used to detect TNIK gene expression (D). Unpaired t test, 2-tailed, was used to compute P values: vehicle vs. IFNγ + TNFα, **P = 0.0046 (VCAM-1), ****P <0.0001 (VCAM-1) **P = 0.0074 (TNIK). Error bars represent 1 SD from the mean; n = 3 biological replicates per each condition.

TNIK Depletion Induces Apoptosis and Reduces Cell Viability in Primary hRPTECs

Apoptosis is critical for proper tissue development and homeostasis and is involved in disease processes. In AKI, apoptosis has been proposed to play a role in recovery through clearance of damaged proximal tubule cells and profibrotic myofibroblasts (28–30). Evidence from recent studies in human sarcoma cell lines reveals that genetic silencing of TNIK not only suppresses cell growth but also induces apoptosis (31). These findings led us to evaluate whether TNIK depletion by siRNA in primary hRPTECs might regulate cell survival.

We employed a kinetic assay to monitor cell viability over 72 h and measured luminescence every 24 h. We found that relative to primary hRPTECs transfected with scr-siRNA, primary hRPTECs transfected with siTNIK showed decreased viability at each time point tested (Fig. 7A). Externalized phosphatidylserine (PS) and subsequent Annexin V binding are hallmarks of early-stage apoptosis (32). Using a kinetic assay, we also detected Annexin V binding to PS by increased luminescence signals between 48 and 72 h after siTNIK depletion in primary hRPTECs (Fig. 7B). Between 72 and 96 h, the luminescence signal plateaued, indicating a mature apoptotic phenotype, followed by probable necrosis after 96 h (33). Apoptotic cells stimulate PARP-1 activity, which is involved in DNA repair processes and cleaved by various apoptotic proteases such as caspase-3 and -7 (34–36). By bulk RNA-sequencing analysis, we found differential expression of PARP-1 at 48 and 72 after scr-siRNA versus siTNIK transfection (Fig. 7C). These findings motivated us to test protein lysates from siTNIK-depleted primary hRPTECs for cleaved PARP-1 expression. By immunoblotting, we detected an 89-kDa PARP-1 fragment, a key biochemical signature of cells undergoing apoptosis, after 48 and 72 h of siTNIK silencing (Fig. 7, D–F) (37). Since siTNIK depletion in primary hRPTECs reduces cell viability and induces apoptosis, these data suggest that Tnik normally serves as a prosurvival and antiapoptotic factor in FR-PTCs and injured primary hRPTECs.

Figure 7.

Traf2 and Nck interacting kinase (TNIK) depletion reduces cell viability and induces apoptosis in injured primary human renal proximal tubule epithelial cells (hRPTECs). A: cell viability was measured in real time via luminescence signals produced by cell viability metabolism (MT) substrate reduction into NanoLuc substrate for NanoLuc luciferase. Data represent mean luminescence values [relative light units (RLU)] after transfection with scramble siRNA (scr-siRNA) or pooled siRNA targeting human TNIK (siTNIK) for 24, 48, or 72 h. B: induction of apoptosis was monitored via luminescence signals correlating with Annexin V binding to externalized phosphatidylserine (PS). C: differential expression [bulk RNA-sequencing (RNA-seq), n = 6 biological replicates] of poly(ADP-ribose) polymerase (PARP)-1 in scr-siRNA- vs. siTNIK-transfected primary hRPTECs at 48 and 72 h. FC, fold change. D: immunoblot of lysates from primary hRPTECs transfected with scr-siRNA or siTNIK for 48 or 72 h. Immunoblots were probed for TNIK or PARP-1; note detection of 89-kDa cleaved PARP-1 fragment at 48 and 72 h after siTNIK transfection. E and F: 48-h (E) and 72-h (F) mean ratios of 89-kDa cleaved PARP-1 to γ-Tubulin. Luminescence assays represent n = 9 biological replicates. Unpaired multiple t test, corrected for multiple comparisons with the Holm–Šídák method, was used to compute P values: scr-siRNA vs. siTNIK, ****P < 0.0001. Immunoblot assays represent n = 3 biological replicates. Unpaired t test, 2-tailed, was used to compute P values: scr-siRNA vs. siTNIK, *P = 0.0322; **P = 0.0049. Error bars represent 1 SD from the mean. ns, not significant; PC, positive control.

DISCUSSION

Our study has three primary findings. First, the kinase Tnik is specifically upregulated in FR-PTCs after AKI. Second, TNIK exerts anti-inflammatory functions in injured PT cells. Third, TNIK protects injured PT cell viability and is antiapoptotic. We present preliminary evidence of TNIK expression in FR-PTC in injured human kidney. These results will need to be generalized in future work examining TNIK expression in a much larger human kidney biobank.

In concert with Vcam-1 expression in FR-PTCs, Tnik upregulation might further contribute to the aberrant phenotype and survival of injured PT cells through suppression of inflammatory and apoptotic signals. In other systems, TNIK has been reported to promote the proliferation and survival of cancerous cells. Global Tnik deletion in two murine models of colorectal cancer resulted in a reduction of cancerous lesions in both colon and small intestine tissues (14). Additionally, in an optic nerve crush injury murine model, deletion of GCK IV family members (Tnik, Map4k4, and Mink1) in synergy with Pten loss resulted in enhanced axon regeneration (38). These studies suggest that in the context of injury and disease Tnik deletion might prove beneficial. Additional studies will be needed to test the outcome of Tnik deletion specifically in FR-PTCs in animal models of AKI, whereby physiological dynamics unattainable in cell culture studies can be modeled.

Additional studies implicate Tnik in modulating proinflammatory and profibrotic signaling. TNIK binds TRAF2 (Tnfr-associated factor 2), which is a member of the TNF superfamily also upregulated in FR-PTCs (Fig. 1B) (11). TRAF proteins function as E3 ubiquitin ligases and adaptor proteins that bind to a distinct cytoplasmic motif on the TNFR and other cell surface receptors (39–41). The function of the protein-protein interaction between germinal center kinases like TNIK and TRAF2 is thought to transduce extracellular signals from inflammatory cytokines such as TNF to intracellular MAPK stress-activated protein kinases such as JNK (42, 43). Activated JNK, in turn, translocates to the nucleus to modulate transcription factors that control a variety of cellular processes including cell survival, differentiation, growth, and apoptosis (44, 45). TNIK has also been proposed to regulate the trafficking of profibrotic/matrix proteins because of its role in actin cytoskeleton control (11, 18). In an immunofluorescence-based siRNA screen, TNIK was found to regulate procollagen I trafficking, as TNIK silencing induced increased intracellular retention and decreased secretion of TGF-β-induced procollagen I in immortalized hepatic stellate cells (18). These findings in part suggest that TNIK may be involved in regulating the profibrotic effects of TGF-β via protein trafficking. Our findings here highlight distinct, anti-inflammatory signaling by TNIK in injured PT epithelia. Whether this is mediated by TRAF2 or other binding partners remains to be investigated. Furthermore, mechanisms of TNIK regulated cytoskeletal and TGF-β signaling in injured PT epithelia also warrant dedicated studies.

We propose a model whereby Tnik is upregulated in FR-PTCs to counter inflammatory signals in FR-PTCs induced by the postischemic environment. In this context, Tnik expression may represent a counterregulatory mechanism in FR-PTCs, which are otherwise proinflammatory, to promote their survival. An important unanswered question is the fate of FR-PTCs. Whether they might be recruited to the healthy repair pathway and then redifferentiate into healthy repaired PT or recruited to the maladaptive pathway remaining dedifferentiated and causing further tissue damage is unknown and will require lineage analysis to resolve.

Intriguingly, TNIK was found to be among a set of injury-repair response-associated transcripts whose expression correlated with reduced estimated glomerular filtration rate (eGFR) at initial biopsy in kidney allografts with AKI (46). This observation, taken together with findings in this study, provides possible mechanistic insight on the correlation of TNIK expression and renal function decline. Additionally, genome-wide association studies (GWAS) have identified the single-nucleotide polymorphism rs9839909 as a kidney disease risk locus, whose target gene has been annotated as TNIK (47). These observations further support the hypothesis that TNIK is involved in renal injury and repair. Further studies are needed to better understand the role of TNIK in renal health, disease, and regeneration.

DATA AVAILABILITY

All siRNA RNA-sequencing data for this study have been deposited in the National Center for Biotechnology Information’s Gene Expression Omnibus (GEO) under Accession No. GSE216015. Previously published single-nucleus RNA-sequencing data [used to generate Fig. 1 (6)] are deposited in GEO under Accession No. GSE139107.

SUPPLEMENTAL MATERIAL

Supplemental Figs. S1–S3: https://doi.org/10.6084/m9.figshare.24898290.v6.

GRANTS

This work was funded by National Institutes of Health Grants DK103740 and UC2DK126024 (to B.D.H.) and Grants T32DK007126, 1L60DK130173-01, and 1K99DK134882-01A1 (to S.T.J.B.) and by a grant from the Burroughs Wellcome Fund (to S.T.J.B.).

DISCLOSURES

B.D.H. is a consultant for Janssen Research & Development, LLC, Pfizer, and Chinook Therapeutics, held equity in Chinook Therapeutics, and received grant funding from Chinook Therapeutics, Pfizer, and Janssen Research & Development, LLC; all interests are unrelated to the present work. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

S.T.J.B. and B.D.H. conceived and designed research; S.T.J.B., Y.K., C.C., N.P.M., Y.Y., and Y.M. performed experiments; S.T.J.B., H.W., Y.K., C.C., Y.Y., Y.M., and B.D.H. analyzed data; S.T.J.B., H.W., Y.M., and B.D.H. interpreted results of experiments; S.T.J.B. prepared final manuscript figures; S.T.J.B. drafted manuscript; B.D.H. edited and revised manuscript; S.T.J.B., H.W., Y.K., C.C., N.P.M, Y.Y., Y.M., and B.D.H. approved final version of manuscript.