PAI-1 induction during kidney injury promotes fibrotic epithelial dysfunction via deregulation of klotho, p53, and TGF-β1-receptor signaling

Abstract

Renal fibrosis leads to chronic kidney disease, which affects over 15% of the U.S. population. PAI-1 is highly upregulated in the tubulointerstitial compartment in several common nephropathies and PAI-1 global ablation affords protection from fibrogenesis in mice. The precise contribution of renal tubular PAI-1 induction to disease progression, however, is unknown and surprisingly, appears to be independent of uPA inhibition. Human renal epithelial (HK-2) cells engineered to stably overexpress PAI-1 underwent dedifferentiation (E-cadherin loss, gain of vimentin), G2/M growth arrest (increased p-Histone3, p21), and robust induction of fibronectin, collagen-1, and CCN2. These cells are also susceptible to apoptosis (elevated cleaved caspase-3, annexin-V positivity) compared to vector controls, demonstrating a previously unknown role for PAI-1 in tubular dysfunction. Persistent PAI-1 expression results in a loss of klotho expression, p53 upregulation, and increases in TGF-βRI/II levels and SMAD3 phosphorylation. Ectopic restoration of klotho in PAI-1-transductants attenuated fibrogenesis and reversed the proliferative defects, implicating PAI-1 in klotho loss in renal disease. Genetic suppression of p53 reversed the PA1-1-driven maladaptive repair, moreover, confirming a pathogenic role for p53 upregulation in this context and uncovering a novel role for PAI-1 in promoting renal p53 signaling. TGF-βRI inhibition also attenuated PAI-1-initiated epithelial dysfunction, independent of TGF-β1 ligand synthesis. Thus, PAI-1 promotes tubular dysfunction via klotho reduction, p53 upregulation, and activation of the TGF-βRI-SMAD3 axis. Since klotho is an upstream regulator of both PAI-1-mediated p53 induction and SMAD3 signaling, targeting tubular PAI-1 expression may provide a novel, multi-level approach to the therapy of CKD.

1 |. INTRODUCTION

Progressive fibrosis and decline of organ function following renal injury are established contributors to chronic kidney disease (CKD), affecting over 15% of the United States population and posing a severe medical burden on health care systems worldwide.^1–4 The renal proximal tubular epithelium is vulnerable to injury in response to ischemia-reperfusion (IR), nephrotoxins, chemotherapeutic drugs, diabetes, hypertension, and ureteral obstruction, and is capable of regenerative (adaptive) repair the damage if the damage is only moderate. In repetitive or severe trauma, tubular epithelial cells undergo necrosis or apoptosis while the surviving epithelium undergoes dedifferentiation, growth arrest, and expression of fibrotic effectors (CCN2, TGF-β1) that promote fibroblast proliferation and ECM deposition via paracrine mechanisms.^5–8 Persistent maladaptive repair results in pathologic remodeling of the renal parenchymal architecture and predisposition to CKD, which eventually progresses to end stage renal disease (ESRD).^5–9 To date, no effective therapies are available to CKD patients,¹⁰ which necessitates the identification of novel mechanisms driving renal tubular dysfunction and fibrogenesis.

Plasminogen activator inhibitor-1 (PAI-1) is an established causative factor of CKD, as PAI-1 global ablation in mice affords protection from unilateral ureteral obstruction (UUO), diabetic nephropathy (DN), and chemically induced renal disorders.^11–14 PAI-1 is highly upregulated in the tubules, interstitium, and podocytes of the diseased kidney,^14,15 likely in response to the profibrotic cytokine TGF-β1 via cooperative signaling among SMAD and non-SMAD (Rac1/NOX, EGFR, Src, YAP/TAZ, ATM, and p53) pathways.^16–28 Indeed, PAI-1 deficiency attenuates interstitial fibrosis and glomerular disease in TGF-β1 overexpressing mice,²⁹ while conditional fibroblast-specific PAI-1 ablation (TNC-Cre/PAI-1^loxp/loxp) mitigates UUO-driven fibrogenesis.³⁰ These studies collectively confirm the profibrotic role of PAI-1 in progressive renal diseases of various etiologies.

PAI-1 is generally thought to be the principal inhibitor of urokinase-type plasminogen activator (uPA) activity, leading to increased extracellular matrix accumulation.³¹ Surprisingly, UUO-driven renal fibrogenesis in uPA^−/− mice is comparable to that of wild type mice,³² strongly suggesting that the PAI-1 contribution to kidney fibrosis occurs via uPA-independent mechanisms. These unexpected findings necessitate a re-evaluation of the precise causative role for PAI-1 in CKD progression, particularly in the tubular epithelium where PAI-1 is highly induced, regardless of injury etiology. Although PAI-1 expression is elevated during renal aging (another CKD risk factor), how this serine protease inhibitor interacts with other aging associated entities (such as klotho) during progressive renal disease requires clarification.

Global ablation of klotho predisposes mice to premature aging and development of renal diseases.³³ Klotho is predominantly expressed in the renal tubules and klotho levels are dramatically downregulated during kidney injury, which appears to be a contributing factor in the development of fibrosis. Renal tubular-specific loss of klotho expression also predisposes to CKD via hyperactivation of growth factor (eg, Wnt, IGF, TGF-β1, FGF) signaling and acquisition of a senescent-like secretory phenotype in tubular epithelial cells,^34–37 further implying a causative role of renal klotho repression in fibrosis. While the mechanisms of klotho deregulation are not fully understood, TGF-β1-induced epigenetic silencing and non-epigenetic pathways have emerged as major contributors to renal klotho silencing. Inhibition of miR-152 and miR-30a by TGF-β1 results in an upregulation in DNMT1/3a levels, leading to increased klotho promoter hypermethylation and subsequent gene silencing.³⁸ TGF-β1-mediated p53 activation also induces expression of miR-34a, which binds to the 3′ UTR of the klotho transcript, leading to suppression of klotho and renal fibrogenesis.³⁹ The definition of additional upstream controls on renal klotho downregulation, however, is necessary to fully explore the clinical utility of rescuing klotho expression as an anti-fibrotic strategy against kidney disease.

While loss of klotho is associated with increased PAI-1 expression during aging,⁴⁰ whether PAI-1 induction affects klotho levels in the context of renal fibrosis is currently not known. The specific pathologic consequences of robust and persistent PAI-1 upregulation in renal tubular epithelium, as is evident during several progressive kidney diseases, remain unclear. To address this issue, we tested the hypothesis that sustained PAI-1 expression in renal epithelial cells promotes epithelial dysfunction (as evidenced by dedifferentiation, cell cycle arrest and induction of fibrotic factors), via enhanced TGF-β1/p53 signaling and klotho loss, using a novel system in which human renal epithelial cells were engineered to constitutively express a PAI-1 construct under the control of a CMV promotor sequence.

2 |. MATERIALS AND METHODS

2.1 |. Creation of stable cell lines and reagents

Human proximal tubular epithelial HK-2 cells (CRL-2190-ATCC, Manassas, VA, USA) were grown in DMEM (1×) + GlutaMAX-I (10567-014, Gibco) supplemented with 5% fetal bovine serum (FBS; 16000-044, Gibco), 5 units/mL penicillin + 5μg/mL streptomycin (15140-122, Gibco). To generate stable PAI-1-expressing cultures, semiconfluent HK-2 cells were infected with lentiviruses bearing a cytomegalovirus (CMV) promoter–driven PAI-1 cDNA construct (CMV-PAI-1) (LPP-F0606-Lv105) or an empty vector (CMV-Con) (LPP-NEG-Lv105) (GeneCopoeia, Rockville, MD, USA) in 5 μg/mL Polybrene (sc-134220, Santa Cruz Biotechnology, Dallas, TX, USA) in 5% FBS/DMEM for 24 hours. The cells were allowed to recover for 24 hours before stable selection by incubating cells in complete medium (as described above) containing 5 μg/mL puromycin dihydrochloride (sc-108071, Santa Cruz Biotechnology, Dallas, TX, USA); media were changed every 3 days to maintain selection pressure. PAI-1 expression was confirmed by immunofluorescence imaging and immunoblot analysis. Due to the proliferative defects in the CMV-PAI-1 cultures, for each experiment CMV-PAI-1 dishes were initially seeded three times higher than CMV-Con dishes to achieve similar cell densities at the time of extraction for western blot analysis. To generate stable double-transgenic epithelial cell lines with p53 or TGF-β1/TGFβRI-silencing in the context of PAI-1 upregulation, low-density (40%) CMV-PAI-1 HK-2 cells were reinfected (for 1 day) with control (sc-108080) or p53 (sc-29435-v), TGF-β1 (sc-270322-v), and TGF-βRI (sc-40222-v) short hairpin RNA (shRNA) lentiviral constructs (Santa Cruz Biotechnology, Dallas, TX, USA) and maintained in complete medium containing puromycin (as described above). p53 and TGF-βRI depletion in PAI-1-overexpressing double-transgenic cultures were confirmed by western blot analysis and TGF-β1 depletion was confirmed by cytokine array analysis. To create double transgenic HK-2 cells with PAI-1 and klotho stable overexpression, semiconfluent CMV-PAI-1 cultures growing in serum-containing medium were infected with CMV-GFP-Klotho (LPP-Z9677-Lv122) or CMV-Control vector lentiviral particles (LPP-NEG-Lv105, GeneCopoeia, Rockville, MD, USA) in complete medium containing 5 μg/mL Polybrene for 24 hours. Stable double transductants were selected and klotho levels were confirmed by immunoblot analysis for GFP as well as klotho.

2.2 |. Western blot analysis

Cultured cells were lysed in Laemmli sample buffer (161-0737, Bio-Rad laboratories, Hercules, CA, USA) containing 5% 2-mercaptoethanol (Sigma) and boiled for 5 minutes. For electrophoresis, 30 μg of protein from each sample was loaded into Bio-Rad Mini-PROTEAN TGX 10% pre-cast gels and separated proteins were transferred to 0.2 μm nitrocellulose membranes and blocked in 5% milk in 0.05% Triton-X 100/PBS before cutting membranes at the correct molecular weight. Overnight incubation at 4°C utilized the following primary antibodies; rabbit anti-PAI-1 (1:3000) and rabbit anti-collagen-1 (1:5000) as previously described,¹⁵ rat anti-klotho (1:5000; Transgenic Inc.-KM2119), rabbit anti-SMAD3 (1:1000; Cell Signaling-9523), rabbit anti-p-Histone3 (1:1000; Cell Signaling-9701), rabbit anti-p21 (1:1000; Cell Signaling-2947), rabbit anti-caspase-9 (1:1000; Cell Signaling-9502), rabbit anti-c-caspase-3 (1:1000; Cell Signaling-9661), rabbit anti-Snail1 (1:1000; Cell Signaling-38798s), rabbit anti-α-SMA (1:1000; Abcam-ab32575), rabbit anti-phospho-SMAD3 (1:1000; Abcam-ab52903), rabbit anti-fibronectin (1:100,000; Abcam-ab2413), rabbit anti-LOXL2 (1:1000; Abcam-ab96233), rabbit anti-MMP-9 (1:1000; Abcam-ab38898), rabbit anti-MMP-2 (1:1000; Abcam-ab97779), rabbit anti-VEGF-C (1:1000; Abcam-ab9546), rabbit anti-endothelin-1 (1:1000; 18201-IBL America, Minneapolis, MN, USA), rabbit anti-vimentin (1:10,000; Santa Cruz-sc5565), rabbit anti-GAPDH (1:5000; Santa Cruz-sc25778), goat anti-CCN2 (1:500; Santa Cruz-sc14939), rabbit anti-TGF-βRI (1:1000; Santa Cruz-sc9048), rabbit anti-TGF-βRII (1:1000; Santa Cruz-sc220), mouse anti-p53 (1:1000; Santa Cruz-sc126), mouse anti-E-cadherin (1:1000; BD Biosciences-610181), and rabbit anti-p-p53 (1:1000; AF1043-R&D systems, Minneapolis, MN, USA). Membranes were washed three times in 0.05% Triton-X 100/PBS before incubation with appropriate HRP-conjugated secondary antibodies (goat anti-rabbit, 31460; goat anti-mouse, 31430; both from Thermo Fisher Scientific, Waltham, MA, USA; mouse anti-goat, sc-2354; Santa Cruz Biotechnology, Dallas, TX, USA) at a dilution of 1:5000 for 1 hour at room temperature. After three 15-minute washes in 0.05% Triton-X 100/PBS, immunoreactive proteins were visualized with ECL (Bio-Rad Clarity Western ECL Substrate, 170-5061, Bio-Rad Laboratories, Hercules, CA, USA). Membranes were not stripped and reprobed. Protein expression levels were quantified using ImageJ. Densitometric values of each band were acquired by color inversion and the area encompassing the largest band was used for the measurement of the mean intensity for all lanes of the same band using ImageJ software. When all densitometric values were recorded, an area outside the bands on the image was assessed as the negative background and further normalized to GAPDH, Lamin A/C or ERK2 as loading controls.

2.3 |. ELISA

CMV-Con and CMV-PAI-1 cells were grown to 50% confluency in triplicate 100 mm dishes, then incubated in serum-free media for 4 days. The serum-free conditioned media was collected and active TGF-β1 levels were determined by ELISA assay (ab100647; Abcam PLC, Cambridge, MA, USA) according to manufacturer’s recommendation. Briefly, protein standard samples and conditioned media from either CMV-Con or CMV-PAI-1 dishes were placed into wells of a 96 well plate precoated with capture antibodies for TGF-β1 and incubated overnight at 4°C with gentle rocking. Wells were then washed and incubated with a TGF-β1 biotinylated detection antibody for 1 hour at room temperature with gentle shaking, washed again prior to addition of HRP-streptavidin solution to each well, followed by more washes and a 30-minute incubation in the dark with a one-step substrate reagent to initiate a color change. Reaction stop solution was added and the absorbance of each well at 450 nm measured using a Biotek EL800 plate reader. Results were normalized to negative control wells and the TGF-β1 concentration in each sample was calculated using a standard curve.

2.4 |. Cytokine array

Total cell lysates with their compliment-conditioned media were analyzed for TGF-β1, TGF-β2, TGF-β3, or TIMP1 and TIMP2 using the human cytokine antibody array (ab133998-Abcam PLC, Cambridge, MA, USA). Using manufacturer’s instructions, CMV-Con, CMV-PAI-1 or CMV-PAI-1+Con shRNA CMV-PAI-1+TGF-β1 shRNA HK-2 cultures were grown to 80% confluency in complete medium, then serum-starved for 3 days, at which time confluency was reached. Conditioned media was harvested from each dish and cells were lysed using lysis buffer provided in the kit supplemented with (10 μM) MG132 (S2619; Selleckchem, Houston, TX, USA) as per manufacturer’s instruction. Precoated membranes were blocked using blocking buffer provided in the kit; 30 μg of protein from lysates was diluted in blocking buffer and undiluted conditioned media were incubated on the membranes overnight at 4°C, then washed 5× using kit wash buffer. Reconstituted biotinylated cytokine antibodies from the kit were then incubated with the membranes overnight at 4°C, followed by repeated washes and a final overnight incubation with HRP-Streptavidin-conjugated biotin detection antibodies supplied in the kit; membranes were washed and incubated with detection buffer for 2 minutes prior to film exposure. Densitometry was performed as described above for bands on both conditioned media and cell lysate embedded membranes using ImageJ.

2.5 |. Immunofluorescence

CMV-Con and CMV-PAI-1 cells plated in two-well chamber slides were allowed to grow for 3 days. Cells were fixed using 4% paraformaldehyde, permeabilized with 0.5% Triton X-100 in PBST (PBS + 0.05% tween-20) and incubated with antibodies to PAI-1¹⁵ or α-SMA overnight at 4°C, followed by a secondary Alexafluor 488 antibody (A-11034; Invitrogen, CA, USA) incubation for 1 hour. After washing, coverslips were mounted using Prolong anti-fade diamond mounting media containing DAPI (P36971; Invitrogen, CA, USA) for nuclear staining. Images were acquired at 20x magnification on an Olympus BX61 upright microscope with PCO.EDGE 4.2 scientific CMOS camera and operated by MetaMorph software (version 7.10.2, Molecular Device).

2.6 |. Cell-cycle analysis

CMV-Con or CMV-PAI-1 HK-2 cells were grown to 70–80% confluence in complete medium supplemented with puromycin, harvested with trypsin, incubated with soybean trypsin inhibitor (5 min), washed with PBS and fixed in 95% ethanol (1 hour). After two washes in PBS, cells were incubated with RNaseA (20 μL/mL) and propidium iodide (2.5 μg/mL) in PBS/Triton-X 100 for 2 hours in the dark. Cell cycle distributions were measured with a FACS Calibur flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA).

2.7 |. Cell growth analysis

CMV-PAI-1 and CMV-Con cells were plated (10 000 cells each) in 60 mm dishes with complete medium and allowed to grow for 5 days, fixed in 4% paraformaldehyde and stained with crystal violet. Five to 10 randomly chosen fields were used to determine cell number per field. All crystal violet images were acquired using an EVOS xl digital inverted microscope at 10x magnification.

2.8 |. Cell death analysis

Brightfield images of 4% paraformaldehyde-fixed and crystal violet-stained confluent monolayers before (day 0) or after stress by serum deprivation for 6 days (CMV-Con, CMV-PAI-1, CMV-PAI-1+Control-shRNA and CMV-PAI-1+p53-shRNA HK-2) and 8 days (Con-shRNA and PAI-1-shRNA HK-2) served to assess cell viability.

2.9 |. Annexin-V staining

CMV-PAI-1 and CMV-Con HK-2 cells were seeded into 150 mm culture dishes and allowed to propagate to 90% confluency in complete medium containing 5 μg/mL puromycin dihydrochloride. Cells were dissociated with 0.05% Trypsin-EDTA (1x; 25300-054, Gibco), processed, stained with FITC-conjugated Annexin V and propidium iodide (PI) as per the manufacturer’s instructions (ab14085, Abcam PLC, Cambridge, MA, USA), and 500 000 events/samples were collected using a BD LSR II flow cytometer with FACSDiva software and analyzed using FlowJo v10 (BD Biosciences, Franklin Lakes, NJ, USA). Four consecutive passages of cells were utilized for analysis. Briefly, dead cells and debris were excluded using forward scatter area (FSC-A) and side scatter area (SSC-A), followed by doublet exclusion using FSC-A and forward scatter height (FSC-H). Positive staining area for Annexin V and PI was established utilizing unstained and single-stained controls. Healthy cells were negative for both Annexin V and PI, while early apoptotic cells were classified as positive for Annexin V and negative for PI; frequencies are expressed as a percentage of live, single cells. Cell culture reagents were obtained from ThermoFisher Scientific (Waltham, MA, USA) unless otherwise specified.

2.10 |. Pharmacologic inhibition of TGF-β1 receptor activity

To investigate the involvement of TGF-β receptor-1 (TGF-βRI) in PAI-1 driven fibrotic responses, confluent CMV-Con cells were serum-starved for 3 hours before incubation with the TGF-βRI inhibitor SB431542 (10 μM) (S1067; Selleckchem, Houston, TX, USA) or an equal volume of DMSO (Sigma; D2650) for 1 hour prior to TGF-β1 (2 ng/mL) (240-B; R&D Systems, Minneapolis, MN, USA) stimulation for 24 hours prior to immunoblot analysis. CMV-PAI-1 cells were serum-starved for 3 hours and either stimulated with 10 μM SB431542 or equal volume of DMSO for 24 hours prior to extraction for western blotting. To confirm the functionality of the TGF-β1 neutralizing antibody (AB-101-NA; R&D Systems, Minneapolis, MN, USA), confluent and serum-deprived CMV-Con HK-2 cells were incubated with either 20 μg/mL TGF-β1 neutralizing antibodies or 20 μg/mL normal chicken IgY (AB-101-C; R&D Systems, Minneapolis, MN, USA) as the negative control, followed by stimulation with TGF-β1 for 24 hours; extracted lysates were subjected to western blot analysis. To investigate potential contribution of TGF-β1 ligand in genetic responses induced by PAI-1 overexpression, various concentrations of TGF-β1 neutralizing antibodies (20, 40, 60 μg/mL) as well as 60 μg/mL normal chicken IgY anti-sera were added to identically confluent serum-deprived CMV-PAI-1 cultures and extracts were prepared 24 hours later.

2.11 |. Statistical analysis

Two-tailed Student’s t-test analysis and ANOVA with post hoc correction was used to assess significant differences (ie, P > .05).

3 |. RESULTS

3.1 |. Sustained epithelial PAI-1 expression in HK-2 cells results in epithelial dedifferentiation, G2/M cell cycle arrest, fibrotic factor induction, and a predisposition to apoptosis

The pathological consequences of tubular PAI-1 induction during renal injury are currently unclear. To determine whether sustained PAI-1 expression promotes a renal epithelial maladaptive phenotype, we engineered human renal epithelial cells (HK-2) to stably express a CMV-driven PAI-1 expression construct (CMV-PAI-1) via lentiviral transduction (Figure 1A). PAI-1 upregulation in transgenic cells was confirmed by immunofluorescent imaging (Figure 1B) with over 98% PAI-1 positivity (Figure 1C; P < .001) and western blot analysis (>35-fold increase compared to vector transduced HK-2 control population) (CMV-Con) (Figure 1D,E; P < .01). Epithelial dedifferentiation (eg, acquisition of a fibroblastoid-like morphology) was evident in the CMV-PAI-1 population relative to CMV-Con cultures, which maintained the typical HK-2 renal epithelial cuboidal morphology (Figure 2A). Immunoblot analysis further confirmed a significant loss of the epithelial marker E-cadherin (>98% decrease) and robust upregulation of the mesenchymal markers vimentin (>28-fold) and α-smooth muscle actin (α-SMA) (>2.8-fold), which correlated with increased stress-fiber formation in the CMV-PAI-1 cultures relative to control vector transductants (Figure 2C). Collectively, these data suggest a novel role for PAI-1 renal epithelial dedifferentiation (ie, induction of a partial epithelial to mesenchymal transition) (Figure 2D–I; P < .001).

FIGURE 1.

Generation and validation of stable PAI-1 induction in renal epithelial cells. A, Schematic for generation of the PAI-1 stable transductants. HK2s cells are infected with either a control or PAI-1 vector driven by the CMV promoter, followed by stable selection with puromycin. PAI-1 expression in transgenic populations are confirmed in (B) by immunofluorescence imaging, showing PAI-1 [in green] and nuclear staining with DAPI [in blue]. 20x magnification. Scale bar = 165 μm. The highlighted region in the CMV-Con panel in (B) illustrates that these cells do express PAI-1 but at considerably lower levels relative to CMV-PAI-1 cultures. C, Percentage of PAI-1 positivity from 5 individual fields of CMV-Con and CMV-PAI-1 immunofluorescent staining from (B); data represented as (DAPI⁺/PAI-1⁺), n = 3, ***P < .001. D, Western blot analysis of PAI-1 expression between CMV-Con and CMV-PAI-1 cells in biological triplicate experiments with high and low exposure, with GAPDH serving as a loading control. n = 10, **P < .01. E, Histogram depicting PAI-1 levels from western blots in (D), data represented as (mean ± SD)

FIGURE 2.

PAI-1 overexpression in kidney epithelial cells results in dedifferentiation and G2/M growth arrest. A, Crystal violet staining of equally seeded (10 000 cells per dish) CMV-Con and CMV-PAI-1 cultures grown for 5 days. B, Quantification of cell number per field from (A). Three fields per replicate were counted for three separate studies (n = 3). Scale bar = 400 μm, P < .05. C, Immunofluorescent imaging of CMV-Con and CMV-PAI-1 cultures seeded equivalently, showing α-SMA [in red] and nuclear staining with DAPI [in blue]. 20x magnification. Scale bar = 165 μm. D-I, Western blot analysis of CMV-Con and CMV-PAI-1 cells confirming loss of expression of E-Cadherin (D, E; P < .001), upregulation of vimentin (D, F; P < .001), α-SMA (D, G; P < .01), p21 (D, H; P < .001) and p-H3 (D, I; P < .05). Histograms (E-I) depict the relative expression of the indicated markers for biological triplicate cultures for 10 independent studies (n = 10). J-K, Flow analysis of propidium iodide-stained CMV-Con and CMV-PAI-1 cultures demonstrating cell-scattering properties in (J) and the percentage of cells in the G1, S and G2/M phase of the cell cycle in (K). n = 3, *P < .05, ***P < .001

A recent study demonstrated that renal epithelial dedifferentiation induced by Snail1 or Twist expression in renal tubular epithelium leads to G2/M cell cycle arrest, which promotes fibrogenesis and the transition from acute renal injury to CKD.^6,41 Cell count analysis of equally seeded cultures of both transgenic CMV-PAI-1 and CMV-Con cells revealed major differences in growth characteristics, as evident by a >83% lower CMV-PAI-1 cell density relative to CMV-Con cells after 5 days (Figure 2B; P < .05), correlating with upregulation of the cell cycle arrest gene p21 (>55-fold) in PAI-1-transductants compared to vector controls (Figure 2D,H; P < .001). Flow analysis of propidium iodide-stained cultures confirmed that sustained PAI-1 expression leads to an increase in the percentage of cells undergoing G2/M growth inhibition with accompanying reductions in the fraction of cells residing in G1 phase (Figure 2J,K), as well as an increase in the levels of the G2/M arrest marker p-Histone H3 (p-H3^ser10) (>2.6-fold) (Figure 2D,I; P < .05).

CMV-PAI-1 cultures are more prone to cell death (induced by serum deprivation) than CMV-Con cells (initially set at a similar confluence). Consequent to monolayer lifting and cell death evident in CMV-PAI-1 cells, crystal violet staining shows that CMV-Con cultures retained a cellular monolayer while the CMV-PAI-1 dishes had very few attached cells (Figure 3A). Annexin-V positivity also confirmed the initiation of cell death signaling (early apoptosis) in CMV-PAI-1 cultures (>5-fold increase) relative to control transductants during growing conditions, despite no apoptotic stimuli (Figure 3B,C; P < .001). This increase in Annexin-V staining also correlated with higher levels of caspase-9 (>3-fold) (Figure 3D,E; P < .001) and cleaved caspase-3 (>56-fold) (Figure 3D,F; P < .05) in the CMV-PAI-1 population relative to controls. Conversely, suppression of PAI-1 expression via lentiviral shRNA transduction revealed that Con-shRNA expressors are more vulnerable to cell death (as evident by monolayer detachment from culture plates) than PAI-1-shRNA cultures (which were initially set at a similar confluency to Con-shRNA cells prior to serum withdrawal) (Figure 3G), further establishing a role for PAI-1 in promoting epithelial cell death. Therefore, PAI-1 sustained induction may contribute to tubular injury and nephron loss, two features associated with worsening of renal function.

FIGURE 3.

Overexpression of PAI-1 in renal epithelial cells results in a predisposition to apoptosis. A, Phase contrast images of confluent CMV-Con and CMV-PAI-1 HK2 cultures at day 0 and crystal violet images of the cultures taken 6 days after serum deprivation to assess cell monolayer detachment. n = 3, Scale bar = 400 μm B, Histogram showing the number of cells positive for Annexin-V staining between CMV-Con (Black) and CMV-PAI-1 (red) HK2s. C, Representative graphs showing the frequency of healthy cells (Annexin−/PI−) and early apoptotic cells (Annexin+/PI−). D, Western blot analysis showing biological triplicate experiments of CMV-Con and CMV-PAI-1 cultures for basal expression levels of pro-caspase-9 (D, E; P < .001) and cleaved caspase-3 (D, F; P < .05), with GAPDH as the loading control. G, Confluent Con-shRNA and PAI-1 shRNA HK2s subjected to phase contrast microscopy before serum deprivation. Cells are fixed and stained with crystal violet on day 8 to assess for monolayer detachment. n = 3, Scale bar = 400 μm. Data in (E) and (F) are presented as (mean ± SD). n = 3, *P < .05, **P < .01, ***P < .001

CMV-PAI-1 cells also upregulate fibronectin (>3.5-fold) (Figure 4A,B; P < .01), collagen-1 (>5.4-fold) (Figure 4A,C; P < .01) and CCN2 (>10-fold) expression (Figure 4A,D; P < .001) relative to controls. Additionally, PAI-1 sustained expression results in upregulation of the fibrotic molecules LOXL2 (>63-fold) (Figure 4E,F; P < .01), MMP-9 (>3.7-fold) (Figure 4E,G; P < .05), the transcriptional factor Snail1 (>3-fold) (Figure 4E,H; P < .01), TIMP1 (>4-fold) (Figure 4I; P < .05), and TIMP2 (>9.8-fold) (Figure 4J; P < .05), while other markers such as MMP-2, endothelin-1 and VEGF-C remained unchanged (Figure 4E), suggesting that PAI-1 tubular induction promotes a unique fibrotic program. These data collectively suggest that PAI-1 is a novel regulator of renal tubular dedifferentiation, G2/M cell cycle inhibition, apoptosis, and fibrogenesis.

FIGURE 4.

Sustained PAI-1 expression leads to fibrotic factor induction. Immunoblot analysis of biological triplicate experiments of CMV-Con and CMV-PAI-1 HK2 cell extracts showing relative expression levels of fibronectin (A, B; P < .01), collagen-1 (A, C; P < .01), and CCN2 (A, D; P < .001) for ten separate studies, n = 10. E-H, LOXL2 (E, F; P < .01), MMP-9 (E, G; P < .05), Snail1 (E, H; P < .01), MMP-2, Endothelin and VEGF-C (E; n.s.) expression levels between cultures assessed for 3 separate studies. I-J, Cytokine protein array analysis of CMV-Con and CMV-PAI-1 cell lysate extracts for TIMP1 (I; P < .05) and TIMP2 (J; P < .05) expression. n = 3. *P < .05, **P < .01, ***P < .001. Histograms (B-D and F-J) depict the relative expression (mean ± SD) of the indicated markers. ns, not significant

3.2 |. PAI-1-mediated p53 upregulation is critical for renal fibrotic effects and epithelial cell apoptosis

p53 is a critical regulator epithelial dysfunction, as tubular-specific ablation or pharmacological inhibition of p53 mitigates AKI progression to CKD.^{6,15,18,19,42,43} Therefore, we tested the hypothesis that PAI-1-driven tubular dysfunction is orchestrated via p53-dependent mechanisms. CMV-PAI-1 cells increased total p53 protein levels (>10-fold) (Figure 5A,B; P < .001) and p53 phosphorylation at site Ser15 (>12-fold) (Figure 5A,C; P < .05) relative to vector controls. Stable knockdown of p53 (>91% reduction) (Figure 5D,E; p< 0.001) in CMV-PAI-1 cells (termed CMV-PAI-1+p53-shRNA cultures) resulted in downregulated expression of p21 (>68% reduction) (Figure 5D,F; p< 0.01), p-H3 (>61% reduction) (Figure 5D,G; p< 0.01), as well as fibronectin (>50% reduction) (Figure 5H,I; p< 0.01) and collagen-1 (>64% reduction) (Figure 5H,J; p< 0.01) relative to CMV-PAI-1 cells stably infected with control shRNA (termed CMV-PAI-1+Con-shRNA). Therefore, p53 induction downstream of PAI-1 is critical for sustaining both the fibrotic phenotype and epithelial cell cycle arrest. Furthermore, p53 suppression in the context of PAI-1 sustained expression is associated with protection from apoptosis compared to similarly confluent CMV-PAI-1+Con-shRNA cells, which underwent monolayer detachment following serum deprivation, further implicating a role for p53 in PAI-1-mediated epithelial cell death (Figure 5K). Our previous studies established that p53 is also a non-SMAD transcription factor critical for TGF-β1-mediated renal fibrotic responses via SMAD3 cooperation.^18,19

FIGURE 5.

PAI-1 induced p53 upregulation is critical for the subsequent maladaptive response. A, Western blot assessments for total and p-p53 protein levels between CMV-Con and CMV-PAI-1 populations. B-C, Histograms depicting the relative expression of p53 levels (mean ± SD) for three independent studies, n = 3. D-G, Lysates of CMV-PAI-1+Con-shRNA and CMV-PAI-1+p53-shRNA double transductants are immunoblotted for p53 (D, E; P < .001), p21 (D, F; P < .01), p-H3 (D, G; P < .01), fibronectin (H, I; P < .01), collagen-1 (H, J; P < .01). Histograms in (E-G) and (I-J) depict the relative expression (mean ± SD) for indicated proteins from the immunoblots in (D) and (H), shown as biological triplicates for three independent studies (n = 3). K, Confluent monolayers of CMV-PAI-1+Con-shRNA and CMV-PAI-1+p53-shRNA HK2 cultures are serum-starved for 6 days. Phase contrast and crystal violet images are taken on day 0 and day 6 to assess cell monolayer detachment. Scale bar = 400 μm. L, Western blot analysis for FAK and p-ERK1/2 protein levels between CMV-Con and CMV-PAI-1 cultures, with ERK2 serving as a loading control, (n = 3). *P < .05, **P < .01, ***P < .001

Expression of FAK (a focal adhesion protein and a non-SMAD component), on the other hand, remains relatively even between CMV-Con and CMV-PAI-1 cells. These observations, however, do not rule out the potential involvement of FAK in PAI-1-driven genetic responses. We also noted modest increases in p-ERK2 levels while p-ERK1 levels are decreased in transgenic PAI-1 expressing cells (Figure 5L). Ongoing studies will fully define the involvement of non-SMAD cascades downstream of PAI-1 upregulation.

3.3 |. Klotho reduction downstream of PAI-1 is causatively linked to tubular dysfunction and is upstream of p53 induction

Loss of klotho expression and elevated PAI-1 levels are linked to accelerated aging and kidney disease, although the precise relationship between these entities is not well defined.⁴⁰ Renal tubular-specific depletion of klotho is associated with predisposition to renal fibrogenesis, implicating epithelial klotho loss in renal disease progression.⁴⁴ Administration of recombinant klotho protein or expression plasmid via intraperitoneal injection attenuates UUO-driven renal fibrosis, demonstrating the therapeutic efficacy of restoring klotho against progressive CKD.^34,36 Upstream controls on klotho repression, however, are not well understood; we therefore tested for the potential existence of a negative feedback loop, in which stable epithelial PAI-1 induction promotes klotho repression. We find klotho expression is significantly attenuated (>98% decrease) in PAI-1 stable transductants relative to CMV-Con cultures (Figure 6A,C; P < .01), suggesting that PAI-1 is a novel regulator of klotho levels in renal epithelial pathology. Moreover, genetic rescue of klotho in PAI-1 overexpressing cells via lentiviral transduction of a CMV-driven, wild type full length klotho expression construct linked to a GFP tag (CMV-PAI-1+CMV-Klotho-GFP) mitigates fibrotic factor CCN2 upregulation (>76% decrease) (Figure 6D,F; P < .001) and p21 induction (>79% decrease) (Figure 6D,G; P < .01) in contrast to CMV-PAI-1 cultures stably transduced with a control vector (CMV-PAI-1+CMV-Con). Surprisingly, klotho re-expression in CMV-PAI-1 cells also dramatically downregulates p53 levels (>84% reduction) (Figure 6D,H; P < .001) compared to the p53 levels in CMV-PAI-1+CMV-Con double-transductants, identifying a previously unknown upstream pathogenic role for klotho in control of renal p53 expression.

FIGURE 6.

Klotho downregulation consequent to PAI-1 induction contributes to epithelial dysfunction. A-C, Immunoblot analysis of CMV-Con and CMV-PAI-1 cell lysates for PAI-1 (A, B; P < .001) and klotho (A, C; P < .01) expression. Histograms in (B-C) represent the relative expression of the indicated proteins as (mean ± SD) for three independent studies (n = 3). D-I, Protein extracts of CMV-PAI-1+CMV-Con and CMV-PAI-1+CMV-Klotho double transgenic cultures are immunoblotted for klotho (D, E; P < .001), CCN2 (D, F; P < .001), p21 (D, G; P < .01), p53 (D, H; P < .001), p-SMAD3 (D, I; P < .01). GAPDH serves as loading control. Histograms in (E-I) depict the relative levels (mean ± SD) of the indicated proteins for three separate experiments (n = 3). **P < .01, ***P < .001

3.4 |. PAI-1-dependent fibrotic reprogramming is also orchestrated via ligand-independent TGF-β1-receptor-1 activation following klotho repression

Ectopic klotho expression in PAI-1-transduced cells is also associated with a substantial decrease SMAD3 phosphorylation (>71% reduction) (Figure 6D,I; P < .01) relative to levels observed in double transgenic CMV-PAI-1+CMV-Con populations. Recombinant klotho is known to interact with TGF-β1 receptor-2 (TGF-βRII), inhibiting receptor and SMAD3 activation downstream of TGF-β1.³⁷ These findings necessitate a clarification of the causative role that TGF-β1 signaling cascades play in the PAI-1-induced epithelial dysfunction. In fact, TGF-β1 receptor 1 and 2 protein levels (TGF-βRI, TGF-βRII) (>1.8 and 3.7-fold, respectively) (Figure 7A,B; P < .05 and Figure 7A,C; P < .001) as well as SMAD3 phosphorylation (>5-fold) (Figure 7A,D; P < .05) are significantly increased in CMV-PAI-1 cultures relative to vector controls. Pharmacological inhibition of TGF-βRI with 10μM SB431542, as anticipated, virtually eliminated TGF-β1-mediated induction of SMAD3 phosphorylation (>90% reduction) (Figure 7E,F; P < .05), fibronectin (>97% reduction) (Figure 7E,G; P < .01) as well as PAI-1 (>92% reduction) expression (Figure 7E,H; P < .001), confirming the functionality of the drug. SB431542 incubation also repressed SMAD3 activation (>82% reduction) (Figure 7I,J; P < .01), downregulated fibronectin (>62% reduction) (Figue 7I,K; P < .01) and vimentin (>64% reduction) levels (Figure 7I,L; P < .05) in PAI-1 overexpressing cells, suggestive of TGF-β1-receptor-1 involvement in the PAI-1-driven fibrotic reprogramming. Additionally, shRNA knockdown of TGF-βRI in CMV-PAI-1 cultures (>87% reduction) (Figure 7M,N; P < .05), indeed, leads to a >60% reduction in p-SMAD3 levels (Figure 7M,O; P < .05), a >52% decrease in fibronectin (Figure 7M,P; P < .05) and a >61% decrease in p21 expression (Figure 7M,Q; P < .01), conclusively demonstrating a role for TGF-βRI in the PAI-1-driven pathogenic responses.

FIGURE 7.

TGF-β1 receptor involvement in the PAI-1-driven fibrotic response. A-D, Western blot assessment of CMV-Con and CMV-PAI-1 lysate extracts for TGF-βRI (A, B; P < .05), TGF-βRII (A, C; P < .001), and p-SMAD3 (A, D; P < .05) protein levels for three independent studies. E-H, Drug control experiment for SB431542. CMV-Con cells are treated with either DMSO or the TGF-β1 receptor-1 inhibitor SB431542 (10 μM) followed by either no stimulation or stimulation with 2 ng/mL TGF-β1 for 24 hours prior to western blot analysis for protein levels of p-SMAD3 (E, F; P < .05), fibronectin (E, G; P < .01), and PAI-1 (E, H; P < .001). Histograms in (F-H) represent the relative expression of the indicated protein as (mean ± SD), n = 3. I-L, CMV-Con and CMV-PAI-1 cultures are treated with DMSO or 10 μM SB431542 for 24 hours, followed by western blot analysis for p-SMAD3 (I, J; P < .01), fibronectin (I, K; P < .01) and vimentin (I, L; P < .05). Histograms in (J-L) represent the relative levels (mean ± SD) of the indicated protein. M-Q, Western blot analysis of CMV-PAI-1+Con shRNA and CMV-PAI-1+TGF-βRI shRNA HK2 lysate extracts for TGF-βRI (M, N; P < .05), p-SMAD3 (M, O; P < .05), fibronectin (M, P; P < .05), and p21 (M, Q; P < .01). N-Q, represent the relative expression of the indicated protein as (mean ± SD), n = 3. *P < .05, **P < .01, ***P < .001, n.s., not significant

Based on these findings, it was critical to investigate the potential requirement for TGF-β1 ligand in PAI-1-driven genetic responses. ELISA studies indicate that the difference between active TGF-β1 ligand levels between CMV-Con and CMV-PAI-1 cultures (maintained under serum-deprived conditions) is less than 12 pg/mL (153 vs 164 pg/mL), which barely makes this difference statistically significant (Figure 8A; P < .05). Furthermore, we determined TGF-β1, TGF-β2, and TGF-β3 ligand expression levels in the conditioned media by protein cytokine array analysis in a second approach, and report that there are, in fact, no appreciable differences in secreted TGF-β1, 2, and 3 ligand levels between CMV-Con and CMV-PAI-1 cultures (Figure 8B–D). Expression of fibronectin and collagen-1 levels are significantly higher in CMV-PAI-1 transductants relative to CMV-Con cultures stimulated with TGF-β1 ligand (2 ng/mL), suggesting that the PAI-1-driven fibrotic response is more robust than one induced by TGF-β1 (Figure 8E). Incubation of PAI-1 overexpressing cells with a TGF-β1 neutralizing antibody had no effect on the PAI-1-mediated SMAD3 activation and fibrotic factor expression (fibronectin, vimentin, p53, and p21 induction) at all three increasing doses tested (20, 40, and 60 μg/mL) relative to control antisera (IgY chicken control) stimulated CMV-PAI-1 cultures (Figure 8G). The TGF-β1 neutralizing antibody, as expected, completely blocked TGF-β1-stimulated SMAD3 phosphorylation as well as fibronectin induction and repression of E-cadherin expression in CMV-Con cultures even at the lowest dose tested (20 μg/mL) (Figure 8F), confirming functionality of the antibody. To definitively rule out TGF-β1 ligand involvement in PAI-1 driven tubular dysfunction, TGF-β1 was silenced using lentiviral shRNA approaches in CMV-PAI-1 cultures. Lysate levels of p-SMAD3, fibronectin, collagen-1, vimentin, and p53 in CMV-PAI-1+Con shRNA double transductants are similar to levels in CMV-PAI-1+TGF-β1 shRNA expressing cells (Figure 8H). Therefore, the fibrotic effects induced by PAI-1 overexpression occur via TGF-β1 ligand independent, yet TGF-β1-receptor-dependent mechanisms. Collectively, our studies identify novel signaling cooperation among p53, klotho, and TGF-βRI/SMAD3 in promoting renal epithelial dysfunction induced by sustained PAI-1 expression. Furthermore, klotho repression appears to be critical for both increases in p53 levels and SMAD3 phosphorylation in the context of PAI-1-driven maladaptive fibrotic phenotype.

FIGURE 8.

Dysfunction driven by PAI-1 overexpression is independent of TGF-β1 ligand synthesis or release. A, ELISA analysis for active TGF-β1 ligand concentrations in the conditioned media isolated from serum-starved CMV-Con and CMV-PAI-1 cultures. n = 3. B-D, Cytokine protein array analysis of CMV-Con and CMV-PAI-1 conditioned media for active TGF-β1 (B), TGF-β2 (C), and TGF-β3 (D). Graphs depict the relative levels of secreted ligands (mean ± SD), n = 3. E, Immunoblot comparison of fibrotic responses in the cellular lysates of TGF-β1 stimulated or unstimulated CMV-Con and untreated CMV-PAI-1 culture extracted in parallel. F, CMV-Con HK2 cells pretreated with 20 μg/mL of TGF-β1 neutralizing antibody or 20 μg/mL IgY control antisera are stimulated with 2 ng/mL TGF-β1. Cells are harvested after 24 hours and expression of p-SMAD3, fibronectin and E-cadherin are analyzed by western blot. n = 3. G, Equally seeded CMV-PAI-1 HK2s are treated with various concentrations of TGF-β1 neutralizing antibody (0, 20, 40, 60 μg/mL) or 60 μg/mL IgY control antisera for 24 hours prior to western blot analysis of extracts for p-SMAD3, fibronectin, collagen-1, vimentin, p53 and p21, with GAPDH is serving as a loading marker. n = 3, *P < .05, **P < .01, ***P < .001, n.s., not significant. H, Western blot analysis of cell lysate extracts from CMV-PAI-1+Con shRNA and CMV-PAI-1+TGF-β1 shRNA HK2 cells for the indicated fibrotic markers. Cytokine protein array analysis of whole cell lysates (I) and conditioned media (J) used to validate TGF-β1 knockdown

4 |. DISCUSSION

The present findings indicate that persistent renal tubular PAI-1 expression promotes epithelial dedifferentiation and plasticity (eg, loss of E-cadherin expression and induction of vimentin), G2/M cell cycle arrest and apoptosis, upregulation of profibrotic cytokines (CCN2), and ECM deposition (eg, fibronectin and collagen-1), identifying a previously unknown role for PAI-1 in renal tubular dysfunction. G2/M proliferative arrest of renal epithelial cells in response to kidney injury is a major contributor to tubular dysfunction, pathological epithelial-fibroblast communication, and CKD progression.⁶ Epithelial plasticity, evident during renal trauma, precedes G2/M arrest and subsequent development of interstitial fibrosis.^6.41 PAI-1 is known to be highly upregulated in kidney tubules, and this study identifies that renal epithelial PAI-1 is a novel regulator of dedifferentiation, cell cycle inhibition and fibrotic maladaptive repair, which is orchestrated via klotho-, p53-and TGF-βRI-dependent mechanisms. PAI-1 deficient mice exhibit less fibrosis,^11–14 while those engineered to overexpress PAI-1 display reduced E-cadherin levels as well as increased ECM accumulation and inflammation compared to wild type mice identically subjected to UUO.⁴⁵ These in vivo observations linking sustained PAI-1 upregulation to renal fibrogenesis are consistent with our in vitro findings that tubular-specific PAI-1 induction promotes fibrotic tubular dysfunction mimicking, to a degree, the in vivo pathology. It is important to note, however, there are limitations to our approach. HK-2 cell culture may not fully mirror the phenotype and/or morphology of the tubular epithelium in vivo. Future work employing a PAI-1 deficient mouse model, therefore, will address the consequences of PAI-1 expression on tubular dysfunction within the context of UUO-initiated renal fibrosis.

The fibrotic effects driven by PAI-1 expression are mediated, in part, by TGF-βRI-dependent but ligand-independent mechanisms, as TGF-β1-neutralizing antibodies or shRNA-mediated TGF-β1 knockdown had virtually no effect on PAI-1-directed tubular dysfunction, while TGF-β1 receptor chemical blockade or genetic depletion attenuated the induction of fibrotic effectors and growth arrest genes. Consistent with these findings, TGF-βRI/II expression and SMAD3 phosphorylation are significantly increased in PAI-1-transduced epithelial cultures, suggesting that sustained PAI-1 expression intrinsically reprograms the cellular TGF-β1/SMAD3 machinery to facilitate profibrotic signaling without impacting TGF-β1 ligand synthesis or release (Figure 9).

FIGURE 9.

Model. PAI-1 upregulation leads to downregulation of klotho, upregulation of p53, and induction of TGF-β1 receptor signaling independent of the TGF-β1 ligand, resulting in expression and secretion of fibrotic markers, downregulation of E-cadherin and upregulation of vimentin, leading to dedifferentiation, and upregulation of p21, p-H3, causing G2/M cell cycle arrest and a propensity to cell death, collectively establishing a role for PAI-1 in tubular epithelial dysfunction. Klotho regulates both p53 and SMAD3 signaling, promoting PAI-1-mediated tubular maladaptive responses

Our demonstration that klotho is a critical downstream factor in PAI-1-driven tubular epithelial dysfunction represents another novel aspect of the current study. Klotho levels were downregulated in the context of sustained PAI-1 synthesis and re-expression of klotho mitigated fibrotic factor expression as well as SMAD3 signaling. Furthermore, klotho can bind directly to TGF-βRII to act as a negative regulator, preventing SMAD3 activation in renal epithelial cells.³⁷ Loss of klotho expression upon sustained PAI-1 signaling could, therefore, reverse klotho-directed TGF-β1 receptor inhibition, initiating SMAD3-dependent gene transcription. Renal klotho expression is, in fact, repressed in several pathologies, which accelerates tubular dysfunction. This study introduces PAI-1 as a novel klotho repressor. Since administration of soluble klotho protein mitigates fibrogenesis following UUO,^34,37 restoration of klotho levels by targeting PAI-1 or potential downstream intermediaries may have therapeutic utility in the management of renal disease.

PAI-1 also appears to upregulate renal p53 since p53 protein levels are increased by sustained renal PAI-1 expression and p53 stable suppression in PAI-1-transductants attenuates induction of fibrotic factors, reverses proliferative defects, and reduces susceptibility to cell death compared to CMV-PAI-1+Con shRNA double transgenic cells. Recent findings clarified the relationship among klotho, p53, cellular senescence and apoptosis in the kidney. Reduction of klotho, as is evident during ischemia-reperfusion (IR) injury in rats, correlates with increased TUNEL staining. Klotho rescue, conversely, is associated with reduced IR-mediated tubular damage.⁴⁶ Indoxyl sulfate-mediated reduction of renal klotho expression in Dahl-hypertensive rats resulted in increased p53 and p21 expression and β-gal staining, highlighting a potential, albeit complex inter-relationship in progressive kidney disease.⁴⁷ Repression of klotho in human lung fibro-blasts similarly increases p21 and p53 levels while p53 suppression attenuated premature senescence induced by klotho loss.⁴⁸ Our data indicate that klotho is an upstream regulator of p53 in the renal tubular epithelium, as ectopic klotho re-expression downregulated PAI-1-mediated p53 induction and tubular dysfunction.

Activation or induction of renal tubular p53 has pathophysiologic consequences, as kidney tubular-specific depletion of p53 affords protection from acute kidney injury and subsequent progression to CKD in response to ischemic and obstructive renal injury.^42,43,49 p53 cooperates with SMAD3 in orchestrating TGF-β1-induced fibrotic responses, including increased PAI-1 transcription (via binding of p53/SMAD3 complexes to the PAI-1 promoter) in renal epithelial cells.^18,19 Conversely, sustained PAI-1 synthesis markedly upregulates p53, leading to tubular dysfunction; highlighting bi-directional pathological cross-talk between p53 and PAI-1, which perpetuates renal tubular injury and fibrosis. It has become apparent, moreover, that PAI-1 expression is subject to multi-level controls that impact several facets of the tissue fibrotic program. Indeed, PAI-1 induction, as a result of renal epithelial cell depletion of PTEN and PPM1A, contributes to cell cycle arrest and tubular dysfunction since PAI-1 gene silencing reversed growth inhibition and attenuated the fibrotic response orchestrated by PTEN and PTEN loss.^50,51 Therefore, These studies collectively establish a role for PAI-1 upregulation in pathological cell cycle control, orchestrated by context-dependent p53-and SMAD3-dependent mechanisms. PAI-1-induced tubular apoptosis via a p53-dependent mechanism may well have profound implications for CKD progression since epithelial cell death is a major contributor to nephron loss and fibrogenesis.

In summary, sustained PAI-1 expression contributes to epithelial dedifferentiation, G2/M proliferative arrest, fibrogenesis and apoptosis, identifying a novel function for PAI-1 in renal tubular dysfunction in addition to its effects on extracellular matrix accumulation and renal scarring. Inhibition of tubular PAI-1 expression and/or function may provide a novel therapeutic approach to attenuate fibrotic renal maladaptive repair and limit injury to the tubular architecture. Since tubular PAI-1 expression is prominent in various nephropathies, inhibition of PAI-1 upregulation may have broader translational implications to limit renal tubular dysfunction.

Abbreviations:

α-SMA

alpha-smooth muscle actin

AAN

aristolochic acid nephropathy

AKI

acute kidney injury

ATM

ataxia-telangiectasia mutated kinase

CCN2

cellular communication network factor 2

CKD

chronic kidney disease

CMV

cytomegalovirus

DMEM

Dulbecco’s modified Eagle’s medium

DN

diabetic nephropathy

ECM

extracellular matrix

EGFR

epidermal growth factor receptor

ESRD

end stage renal disease

FBS

fetal bovine serum

FGF

fibroblast growth factor

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

HK-2

human kidney-2

IGF

insulin-like growth factor 1

IR

ischemia reperfusion

NOX

NADPH oxidase

PAI-1

plasminogen activator inhibitor-1

PPM1A

protein phosphatase Mg²⁺/Mn²⁺ dependent-1A

PTEN

phosphatase and tensin homolog

SMAD

mothers against decapentaplegic

TAZ

transcriptional coactivator with PDZ-binding motif

TGF-β1

transforming growth factor-β1

TNC

tenascin C

uPA

urokinase-type plasminogen activator

UUO

unilateral ureteral obstruction

YAP

yes-associated protein