Preischemic targeting of HIF prolyl hydroxylation inhibits fibrosis associated with acute kidney injury

Abstract

Acute kidney injury (AKI) due to ischemia is an important contributor to the progression of chronic kidney disease (CKD). Key mediators of cellular adaptation to hypoxia are oxygen-sensitive hypoxia-inducible factors (HIF), which are regulated by prolyl-4-hydroxylase domain (PHD)-containing dioxygenases. While activation of HIF protects from ischemic cell death, HIF has been shown to promote fibrosis in experimental models of CKD. The impact of HIF activation on AKI-induced fibrosis has not been defined. Here, we investigated the role of pharmacologic HIF activation in AKI-associated fibrosis and inflammation. We found that pharmacologic inhibition of HIF prolyl hydroxylation before AKI ameliorated fibrosis and prevented anemia, while inhibition of HIF prolyl hydroxylation in the early recovery phase of AKI did not affect short- or long-term clinical outcome. Therefore, preischemic targeting of the PHD/HIF pathway represents an effective therapeutic strategy for the prevention of CKD resulting from AKI, and it warrants further investigation in clinical trials.

ischemia-reperfusion injury (IRI) in the kidney is a major cause of acute kidney injury (AKI). Despite advances in understanding the pathophysiology of AKI and important improvements in clinical care, the overall prognosis of patients with AKI remains poor and is associated with a mortality rate of 40–80% in the intensive care setting (31). Similarly, preventive strategies have been unsuccessful as the incidence of AKI is increasing and will nearly double over the next decade as the population ages (1). In addition to challenges of acute clinical management, AKI is increasingly recognized as an important contributor to end-stage renal disease (ESRD) (1, 8, 16, 20); ∼6% of patients with AKI progress to ESRD within 2 yr of diagnosis (29). Similarly, animal studies examining long-term outcomes of AKI have detected irreversible functional and structural changes, including tubulointerstitial fibrosis and capillary rarefaction. These facts highlight the urgent need for novel therapeutic approaches that aim at preventing and/or reversing the pathophysiologic sequelae of AKI.

Hypoxic preconditioning, by which short exposure to hypoxia induces resistance to subsequent ischemic injury, has received much attention as a potential novel therapeutic strategy in the prevention of AKI (12). Key mediators of cellular adaptation to oxygen deprivation are hypoxia-inducible factor (HIF)-1 and -2, heterodimeric basic helix-loop-helix transcription factors, which regulate cellular energy metabolism, angiogenesis, erythropoiesis, apoptosis, and cell proliferation. The activity of HIF is controlled by oxygen-, iron-, and ascorbate-dependent dioxygenases, also known as prolyl-4-hydroxylase domain-containing proteins 1–3 (PHD1–3), which use 2-oxoglutarate as substrate for the hydroxylation of specific proline residues in HIF-α. This permits binding to the pVHL-E3 ubiquitin ligase complex, which results in proteasomal degradation of HIF-α (13, 17, 28). While HIFs have been implicated in mediating the cytoprotective effects of ischemic preconditioning, they promote fibrosis in experimental models of chronic kidney disease (CKD) (12, 13). Their role in the development of IRI-associated fibrosis, however, is unclear.

Here, we establish that pharmacologic inhibition of PHDs before AKI ameliorates fibrosis and anemia, features of transition from acute to chronic injury. Furthermore, we show that PHD inhibition following reperfusion injury does not result in adverse short- or long-term effects. Therefore, this therapeutic approach warrants further examination for its potential use in the prevention of CKD.

METHODS

Mice.

Eight-week-old male C57BL/6J mice were purchased from the Jackson Laboratory. For pharmacological HIF activation, prolyl-4-hydroxylase inhibitor GSK1002083A (Glaxo-Smith-Kline, Collegeville, PA) was dissolved in 1% methylcellulose and administered by oral gavage at a dose of 60 mg/kg. Mice subjected to renal IRI were treated with GSK1002083A before IRI or during the early postischemic period. Kidneys were harvested for histological, gene, and protein expression analyses on days 2 or 21 post-IRI. All procedures involving mice were performed in accordance with National Institutes of Health guidelines for the use and care of live animals and were approved by the Institutional Animal Care and Use Committee of Vanderbilt University.

IRI.

Following anesthesia with xylazine (10 mg/kg ip) and ketamine (90–120 mg/kg ip), kidneys were exposed through an abdominal midline incision and both renal pedicles were occluded with microaneurysm clamps (Fine Science Tools, Foster City, CA) within 3 min of each other. The abdomen was temporarily closed with sutures and body temperature was monitored by rectal probe and controlled with a heating pad at 37°C (Fine Science Tools). For experimental consistency, only mice with uniform color change in both kidneys were used for analysis. After 20 min, the clamps were removed and reperfusion of the kidneys was confirmed visually. For mice that underwent unilateral renal IRI, only the right kidney was clamped and the left kidney was used as internal control.

Measurement of renal function.

Blood urea nitrogen (BUN) levels were measured using the QuantiChrom Urea Assay Kit (BioAssay Systems, Hayward, CA). Glomerular filtration rate (GFR) was determined based on plasma FITC-inulin clearance following a single bolus injection (24). Correlation of GFR with BUN levels was confirmed on day 21 post-IRI (R² = 0.71, P < 0.05).

RNA analysis.

Total RNA was isolated with TRIzol reagent (Invitrogen, Carlsbad, CA), followed by clean-up with the RNeasy Mini Kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. For real-time PCR analysis, RNA was reverse transcribed and subjected to PCR amplification either using SYBR Green PCR Master Mix or TaqMan Universal PCR Master Mix on an ABI 7300 platform (Applied Biosystems, Foster City, CA). mRNA expression levels were assessed with the relative standard curve method according to the manufacturer's instructions (Applied Biosystems). Primer sequences for collagen (Col)1α1, Col18α1, lysyl oxidase-like 2 (Loxl-2), plasminogen activator inhibitor-1 (Pai-1), phosphoglycerate kinase 1 (Pgk1), erythropoietin (Epo), glucose transporter 1 (Glut1), and lactate dehydrogenase a (Ldha) have been described previously (6, 14, 26). Primer sequences used to amplify kidney injury molecule-1 (Kim-1): forward 5′-AAACCAGAGATTCCCACACG-3′, reverse 5′-GTCGTGGGTCTTCCTGTAGC-3′; chemokine (cc motif) ligand-2 (Ccl2): forward 5′-CCCAATGAGTAGGCTGGAGA-3′, reverse 5′-TCTGGACCCATTCCTTCTTG-3′; F4/80: forward 5′-CTTTGGCTATGGGCTTCCAGT-3′, reverse 5′-GCAAGGAGGACAGAGTTTATCGTG-3′; tumor necrosis factor-α (Tnf-α): forward 5′-GCTGAGCTCAAACCCTGGTA-3′, reverse 5′-CGGACTCCGCAAAGTCTAAG-3′; transforming growth factor-β1 (Tgf-β1): forward 5′-TGGCGAGCCTTAGTTTGGA-3′, reverse 5′-TCGACATGGAGCTGGTGAAA-3′. 18S ribosomal RNA was used for normalization.

Protein analysis.

Whole kidney lysates were prepared by homogenization in ice-cold RIPA buffer in the presence of protease inhibitors (Roche Diagnostics, Indianapolis, IN). Protein (20 μg) was analyzed by SDS-PAGE (Novex; Invitrogen) and transferred to a nitrocellulose membrane (Novex; Invitrogen). α-Smooth muscle actin (SMA) was detected with a polyclonal rabbit anti-actin, N-terminal antibody (Sigma, St. Louis, MO) followed by incubation with horseradish peroxidase-conjugated secondary antibody. Immunoreactive bands were identified using ECL-plus (Thermo Scientific, Waltham, MA) reagent according to the manufacturer's instructions.

Morphologic analysis.

For histological sections, kidneys were fixed with 10% buffered formalin. Macrophage infiltration was assessed with a monoclonal rat F4/80 antibody (Abcam, Cambridge, MA). Extracellular matrix (ECM) accumulation was detected by Sirius red staining (0.1% fast green FCF and 0.1% direct red 80 in saturated picric acid) on kidney tissue sections. Ten random visual fields were analyzed per kidney section, and percentage of Sirius red-positive area or number of F4/80⁺ cells was determined with ImageJ software. For each mouse, average values from both kidneys were included in the analysis.

Statistical analysis.

Data reported represent mean values ± SE. Statistical analyses were performed with Prism 5.0b (GraphPad Software, La Jolla, CA) using the unpaired Student's t-test and the Mann-Whitney U-test. P values of <0.05 were considered statistically significant.

RESULTS

Pharmacological targeting of PHDs with structural analogs of 2-oxoglurate activates HIF signaling and protects from acute ischemic injuries in multiple tissues, including the kidney (4, 12). To examine whether preischemic PHD inhibition protected from the long-term sequelae of AKI, we designed a preconditioning protocol that consisted of oral administration of two prolyl-4-hydroxylase inhibitor (PHI) doses before renal IRI (Fig. 1A). GSK1002083A, a structural analog of 2-oxoglutarate, stabilizes both HIF-1 and HIF-2, as previously reported (18), and induces HIF-regulated genes in the kidney, including Glut1, Pgk1, Ldha, and Epo (Fig. 1B).

Fig. 1.

Experimental protocol of preischemic prolyl-4-hydroxylase domain (PHD) inhibition in renal ischemia-reperfusion injury (IRI). A: overview of the experimental protocol. Mice were either treated with prolyl-4-hydroxylase inhibitor (PHI) or vehicle alone at 48 and 6 h before bilateral renal pedicle clamping. B: effect of preischemic PHD inhibition on hypoxia-inducible factor (HIF) target gene expression. Shown are renal mRNA levels of glucose transporter 1 (Glut1), phosphoglycerate kinase 1 (Pgk1), lactate dehydrogenase a (Ldha), and erythropoietin (Epo) measured 6 h after the second dose. Bars represent means ± SE. *P < 0.05. **P < 0.01.

For our studies, we used a model of ischemic AKI that is characterized by persistent impairment of renal function, the development of fibrosis, inflammation, and anemia following injury. This is in line with findings from other investigators using similar models (35). We focused our analysis on day 21 post-IRI, a time point characterized by a ∼50% reduction in GFR in injured mice compared with control (113 ± 34 vs. 220 ± 34 μl/min, n = 3). Preconditioning with GSK1002083A afforded an impressive improvement in renal function at both early and late time points resulting in BUN levels of 57 ± 9 vs. 110 ± 5 mg/dl at 24 h and 29 ± 2 vs. 53 ± 10 mg/dl at day 21 post-IRI compared with vehicle-treated controls (P < 0.01; Fig. 2, A and B).

Fig. 2.

Preischemic PHD inhibition preserves kidney function following renal IRI. A–B: 10-day time course of blood urea nitrogen (BUN) following IRI, and BUN values on day 21 post-IRI (n = 6–7). Error bars represent SE. **P < 0.01. ***P < 0.001. ****P < 0.0001.

We next assessed whether PHI preconditioning affected fibrosis, a late consequence of AKI, which reflects an imbalance between injury and subsequent tissue repair. We assessed interstitial fibrosis by Sirius red staining and found that PHI-pretreated animals accumulated less collagen in cortex and medulla compared with controls (64 and 52% reduction, respectively, P < 0.01; Fig. 3A). Consistent with these findings is a substantially reduced expression of α-SMA, a myofibroblast marker that associates with increased ECM production (Fig. 3B), and the reduced expression of Kim-1, a renal injury marker (25-fold reduction, P = 0.009; Fig. 3C). In line with ameliorated fibrosis were decreased transcript levels of Col1α1, Col18α1, Loxl-2, Tgf-β1, and Pai-1 (Fig. 3C).

Fig. 3.

Preischemic PHD inhibition prevents fibrosis and inflammation associated with renal IRI. A: representative images of Sirius red-stained kidneys (cortex, top; medulla, bottom) 21 days post-IRI in PHI- or vehicle-treated mice. Bottom: quantification of fibrotic area in cortex or medulla (n = 6–7). B: immunoblot analysis of α-smooth muscle actin (SMA) levels in whole kidney protein extracts from PHI- or vehicle-pretreated mice at day 21 post-IRI. Ponceau S staining is shown to demonstrate equal loading. C: RT-PCR analysis for kidney injury molecule-1 (Kim-1), collagen (Col)1α1, Col18α1, lysyl oxidase-like 2 (Loxl-2), and transforming growth factor-β1 (Tgf-β1) in PHI- or vehicle-pretreated kidneys (n = 5–7) at day 21. D: F4/80 immunostaining in kidneys from PHI- or vehicle-treated mice at day 21. Graph shows the number of F4/80⁺ cells/high-power field (hpf) in cortex or medulla (n = 6–7). E–F: expression levels of F4/80, Tnf-α, and Ccl2 (n = 5–7), and hematocrit (Hct) values 21 days after IRI in PHI- or vehicle-treated mice (n = 6–7). Scale bars = 50 μm. Error bars represent SE. *P < 0.05. **P < 0.01.

Since CKD is associated with inflammation, we next examined whether preconditioning with PHI modulated inflammation. Immunohistochemical and RNA expression analysis of macrophage marker F4/80 revealed a significant decrease in renal inflammation at day 21, both in cortex and medulla (cortex: 56 ± 7 vs. 118 ± 16 cells/HPF in control mice, medulla: 92 ± 12 vs. 219 ± 29 cells/HPF in control mice, P < 0.01; ∼10-fold reduction in F4/80 mRNA levels, P = 0.007; Fig. 3, D and E). This decrease in macrophage numbers was furthermore associated with a reduction in F4/80, TNF-α, and Ccl2 (P < 0.01; Fig. 3E). Taken together, our data demonstrate that preischemic PHD inhibition preserved kidney function and prevented the development of fibrosis and inflammation.

As a result of AKI-associated renal fibrosis, mice developed anemia. Hematocrits (Hct) were significantly reduced in vehicle-treated but not in PHI-treated mice, which had normal Hcts (32 ± 1.6 vs. 47 ± 6%, P < 0.0001; Fig. 3F). Renal Epo mRNA levels were not different between groups (data not shown), indicating that extensive kidney fibrosis in control mice impaired their ability to adequately respond to anemia with increased Epo production (2).

In a subgroup analysis we identified a small number of PHI-pretreated animals that had significant impairment in renal function in the immediate post-IRI period. Mice in this subgroup were characterized by a rise in BUN at the 24-h time point, but not at days 10 and 21 compared with other PHI-pretreated mice (24-h BUN; 79 ± 9 vs. 28 ± 5 mg/dl, n = 3–6, P < 0.001; Fig. 4). A difference in collagen deposition, macrophage accumulation, and Loxl-2 or Tgf-β1 expression levels between those two groups was not found at day 21 (data not shown). When PHI-pretreated animals with increased BUN were compared with vehicle-treated animals with similar BUN elevation (24-h time point), the PHI-pretreated group developed less fibrosis at day 21. This finding could indicate that PHI-mediated protection from fibrosis may not only be a reflection of ameliorated initial acute injury and reduced cell death, but also involve renal repair mechanisms during the recovery period from IRI.

Fig. 4.

Subgroup analysis of preconditioned PHI-treated mice. BUN time course following bilateral renal IRI in a subgroup of PHI-treated (BUN high) or vehicle-treated mice with comparable BUN elevation (n = 3–4), and PHI-treated mice with low BUN (BUN low, n = 4–6) at day 1 post-IRI. Error bars represent SE. *P < 0.05. **P < 0.01.

Since epithelial HIF-1 has been shown to promote fibrosis in several experimental models of CKD (14, 19), we sought to determine whether the timing of PHI administration affected AKI-associated fibrosis. We therefore administered GSK1002083A to mice with comparable impairment of renal function (BUN levels) on days 2 and 4 post-IRI, when extensive tubular necrosis was present (Fig. 5, A and B, and data not shown). In contrast to preischemic treatment, post-IRI PHD inhibition did not confer protection (Fig. 5B). At day 21 post-IRI, BUN and GFR levels were not different from controls (Fig. 6, A and B). Furthermore, post-IRI treatment did not affect the degree of ECM accumulation, Kim-1, Col1α1, Col18α1, Loxl-2, Tgf-β1, F4/80, Tnf-α, and Ccl2 expression levels, or macrophage numbers (Fig. 6, C and D, and data not shown). These findings are consistent with the development of anemia in both groups (Hct; 30 ± 3 vs. 36 ± 1% in control animals; Fig. 6E). Together, our data suggest that PHD inhibition during the early recovery phase of AKI does not impair kidney function or promote fibrosis.

Fig. 5.

Postischemic PHD inhibition does not improve kidney function. A: outline of the experimental protocol used to investigate the effects of postischemic PHD inhibition on renal IRI. Mice with similar degree of renal injury as determined by BUN at day 2 following injury were treated with PHI or vehicle at days 2 and 4 post-IRI. For assessment of renal function, BUN was measured on days 1–5 and 10 post-IRI. B: 10-day time course of BUN measured in PHI- or vehicle-treated mice with PHI on days 2 and 4 post-IRI. Serum from 5–8 individual mice was sampled per time point.

Fig. 6.

Postischemic PHD inhibition does not impact acute kidney injury (AKI)-associated fibrosis, inflammation, and anemia. Functional and morphologic analysis of kidneys from PHI- or vehicle-treated mice at day 21 post-IRI. PHI was administered on days 2 and 4 post-IRI. A-B: BUN (n = 5) and glomerular filtration rate (GFR) values (n = 3). C: representative Sirius red-stained kidney sections and quantification of fibrotic areas in cortex or medulla (n = 5). D: F4/80 immunohistochemistry: representative images of cortex or medulla from PHI- or vehicle-treated mice; bottom graph shows quantification of macrophage numbers/hpf in cortex or medulla (n = 5). E: Hct levels in PHI- or vehicle-treated mice (n = 5). Scale bars = 50 μm. Error bars represent SE. ns, Not statistically significant.

Since the lack of renoprotection from postischemic PHD inhibition was in sharp contrast to preischemic PHD inhibition, we examined whether and to what degree postischemic PHI administration activated HIF signaling. For this, we analyzed the expression of HIF target genes in contralateral and injured kidneys from mice treated with GSK1002083A on day 2 postunilateral IRI, control kidneys from nonoperated mice treated with GSK1002083A or vehicle, and injured kidneys from vehicle-treated mice that underwent bilateral renal pedicle clamping (day 2 post-IRI). While a blunted HIF response was found with regard to Epo induction (6-fold reduction in PHI-treated IRI kidneys compared with noninjured contralateral kidneys, n = 4, P < 0.01; Fig. 7A), Pgk1, a HIF-1 target gene involved in glycolysis, was not induced by PHD inhibition in postischemic kidneys compared with noninjured kidneys (Fig. 7B). Significant differences were not found for Ldha and Glut1 (data not shown). Since renal Epo is exclusively regulated by HIF-2 and not by HIF-1 (18), our findings could indicate that PHD inactivation in the postischemic kidney may lead to differential activation of HIF signaling.

Fig. 7.

Submaximal HIF activation following postischemic PHD inhibition. Real-time PCR analysis of Epo (A) and Pgk1 (B) mRNA levels in kidneys from nonoperated PHI- and vehicle-treated mice (control kidneys, one dose, n = 3), animals treated with vehicle at day 2 postbilateral renal IRI (bIR; n = 3), and contralateral (CTL) and injured kidneys from mice subjected to unilateral ischemia treated with one dose of PHI at day 2 post-IRI (uIR; n = 4). Real-time PCR analysis was performed 6 h after PHI administration. Error bars represent SE. *P < 0.05. **P < 0.01. ***P < 0.001.

DISCUSSION

In the present study, we investigated whether targeting of the PHD/HIF pathway would impact the development of AKI-associated CKD. We found that preischemic PHD inhibition preserved kidney function and prevented the development of fibrosis, inflammation, and anemia following renal IRI, while inhibition of HIF prolyl hydroxylation during the recovery phase of AKI did not did not have adverse effects.

The molecular mechanisms by which PHD inhibition ameliorates AKI-associated fibrosis remain unclear. Although diminished fibrogenesis may be simply a consequence of improved cell survival during IRI, subgroup analysis of PHI-treated mice that had a significant rise in BUN raises the possibility that renoprotective effects of preischemic PHD inhibition extend beyond the immediate injury phase. It is plausible that certain Hif-regulated molecules enhance tissue repair during recovery from AKI. For example, Epo, which is induced in our model, has been shown to accelerate repair in experimental models of both, AKI and CKD (5). Another example is stromal cell-derived factor-1, a HIF-1-induced chemokine, that enhances recruitment of progenitor cells to regenerating tissues (7). To add complexity to the potential mechanisms that underlie the beneficial effects of pharmacologic PHD inhibition is the role of the PHD/HIF axis in the regulation of inflammation and innate immunity (12). Our finding of decreased inflammation in kidneys of PHI-pretreated animals is in agreement with pharmacologic and genetic studies, demonstrating that loss of PHD function suppresses inflammation in models of colitis and acute lung injury (27, 30, 32). While HIF has been shown to regulate myeloid cell and lymphocyte function (9, 21), HIF-independent effects have been proposed to operate through modulation of IκB kinase β activity, which regulates NF-κB signaling (10).

Our data highlight the importance of timing of HIF activation in the context of renal ischemic injury. In contrast to preconditioning with PHI, no clinical benefit was detected when PHI was administered following IRI, which may be explained by suboptimal induction of certain HIF target genes. While tissue acidosis and reactive oxygen species may have blunted transcriptional HIF responses in our model (11, 34), absence of renoprotection was also reported in a recent study, where PHI was administered at the onset of reperfusion (33). In contrast, in a model of cerebral ischemia, PHD inhibitor dimethyloxaloylglycine increased HIF-regulated gene transcription and reduced injury when administered 30 or 60 min post-IRI (23). Although tissue-specific effects may explain the discrepancy between this study and our findings, regulation of the HIF axis under ischemic conditions remains poorly understood and requires further investigation. In addition to timing, the duration of PHD inhibition/HIF activation appears to play a critical role in determining disease outcome. Mice with cardiomyocyte-specific Phd2 deletion, for example, had significantly improved cardiac function 3 wk after myocardial IRI (15), whereas sustained Phd2 ablation or tissue-specific HIF-1α stabilization resulted in cardiomyopathy (22). While we previously showed that epithelial cell-derived HIF-1 functions as a profibrotic transcription factor in experimental models of CKD (12, 13), transient global pre- or postischemic HIF stabilization did not produce adverse effects. We believe that this may be due to the short duration of HIF activation, but it may also indicate that HIF has cell type-specific functions with regard to fibrosis development (Kobayashi and Haase; Kapitsinou and Haase, unpublished observations).

Our findings have immediate implications for renal transplantation, where a relationship between delayed graft function and long-term outcomes is well-established (25). It is reasonable to speculate that preischemic PHI treatment of donor kidneys may benefit long-term graft survival. Although functional improvement was not reported in experimental models when PHI-pretreated donor kidneys were transplanted into isogenic rats, 10-day graft function did however improve in allografts (3). The lack of clinical benefit in the isogenic setting may be related to experimental design, as removal of the native kidney in this model may have accelerated renal repair as suggested by a recent study (35), thereby masking the protective effects of PHI therapy.

In summary, our study provides proof of principle that pharmacological targeting of PHDs represents an effective therapy in the prevention of the long-term sequelae of ischemic AKI and does not lead to adverse clinical outcomes when applied outside a tight therapeutic window. Our findings support the need for clinical studies that examine the use of PHD inhibition in controlled settings, for example, where patients with preexisting renal disease are at risk to develop procedure-related AKI that increases their likelihood to progress to ESRD.

GRANTS

This work was supported by the Krick-Brooks chair in Nephrology (V. H. Haase), the Vanderbilt Department of Medicine, National Institutes of Health Grants R01-DK081646 (V. H. Haase), R01-DK37097 (C. E. Swan), and a National Scientist Development Grant from the American Heart Association (P. P. Kapitsinou).

DISCLOSURES

K.J.D. and C.L.E. are employees of GlaxoSmithKline and own stocks; V.H.H. serves on the Scientific Advisory Board of Akebia Therapeutics, a company that develops prolyl-4-hydroxylase inhibitors for the treatment of anemia.

AUTHOR CONTRIBUTIONS

Author contributions: P.P.K. and V.H.H. conception and design of research; P.P.K., J.J., M.M., and C.E.S. performed experiments; P.P.K., J.J., M.M., C.E.S., and V.H.H. analyzed data; P.P.K., J.J., M.M., C.E.S., and V.H.H. interpreted results of experiments; P.P.K. prepared figures; P.P.K. and V.H.H. drafted manuscript; P.P.K. and V.H.H. edited and revised manuscript; P.P.K., J.J., M.M., C.E.S., K.J.D., C.L.E.-M., and V.H.H. approved final version of manuscript.