If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address reprint requests to Carsten Willam, M.D., Department of Nephrology and Hypertension, Friedrich-Alexander-University Erlangen-Nuremberg, Ulmenweg 18, 91054 Erlangen, Germany
The role of proximal versus distal tubular injury in the pathogenesis of acute kidney injury (AKI) is debatable. Inhibition of prolyl hydroxylases that regulate the degradation of hypoxia-inducible transcription factors (HIFs) is a promising therapeutic approach to optimize energy preservation under hypoxia and has successfully been applied to protect kidney structure and function in AKI models. Presently used prolyl hydroxylase inhibitors are lipophilic 2-oxoglutarate analogues (2OGAs) that are widely taken up in cells of most organs. Given the selective expression of organic anion transporters (OATs) in renal proximal tubular cells, we hypothesized that hydrophilic 2OGAs can specifically target proximal tubular cells. We found that cellular hydrophilic 2OGAs uptake depended on OATs and largely confined to the kidney, where it resulted in activation of HIF target genes only in proximal tubular cells. When applied in ischemia-reperfusion experiments, systemically active 2OGA preserved kidney structure and function, but OAT1-transported 2OGA was not protective, suggesting that HIF stabilization in distal tubular rather than proximal tubular cells and/or nontubular cells mediates protective effects. This study provides proof of concept for selective drug targeting of proximal tubular cells on the basis of specific transporters, gives insights into the role of different nephron segments in AKI pathophysiology, and may offer options for long-term HIF stabilization in proximal tubules without confounding effects of erythropoietin induction in peritubular cells and unwarranted extrarenal effects.
Acute kidney injury (AKI) is common, particularly occurs in hospitalized patients, and is associated with high rates of morbidity and mortality. Ischemia due to a reduction of renal blood flow below the limits of autoregulation is presumably the most important underlying pathophysiological factor. Although historically termed acute tubular necrosis, tubular injury mainly affects proximal tubules (pars recta) and thick ascending limbs (TALs) in the outer stripe of the outer renal medulla, where oxygen tension is uniquely low.
Both tubular segments respond differently to hypoxia because of their variable metabolic dependency on oxygen and the maintenance of energy stores depending on transport activity.
Hypoxia invokes diverse cellular responses, mediated to a large extent by hypoxia-inducible transcription factors (HIFs), such as glycolysis, angiogenesis, erythropoiesis, cell survival, and cell proliferation.
The oxygen-regulated HIF-α subunits are continuously synthesized but rapidly degraded by the proteasome in the presence of oxygen. Degradation depends on hydroxylation of conserved proline residues in the HIF-α chain by prolyl hydroxylase domain (PHD) enzymes,
When oxygen is absent in the cell and thus hydroxylation and degradation do not occur, HIF accumulates, dimerizes with its β-subunit, and induces its target genes. Two HIF-α subunits, three PHD isoforms, and the factor inhibiting HIF, which controls transcriptional activity by hydroxylating a C-terminal asparagine of HIF, are expressed in the kidney in a cell-specific manner, predominantly in distal tubular cells and interstitial cells.
Because PHDs use oxoglutarate as a cosubstrate, 2-oxoglutarate analogues (2OGAs) have been used to inhibit their enzymatic activity and stabilize HIF in the presence of oxygen.
Preconditional treatment with systemic 2OGAs has been shown to preserve kidney structure and function in models of acute ischemic and toxic kidney injury as well as kidney transplantation.
The protective effect of prolyl-hydroxylase inhibition against renal ischaemia requires application prior to ischaemia but is superior to EPO treatment.
On the basis of their capacity to induce erythropoietin, 2OGAs are also being evaluated for treatment of anemia in patients with chronic kidney disease.
The nephroprotective effects of systemic HIF stabilization are predominantly attributed to preserve proximal tubular cell structure and function. However, recent experiments showed that HIF stabilization in the TAL by segment-specific deletion of Vhl protected against ischemic AKI.
In a complimentary approach, we aimed to analyze the effects of stabilizing HIF in proximal renal tubules with the use of a pharmacologic strategy. 2OGAs used so far in animal experiments and in humans shared lipophilic properties, allowing them to freely cross cell membranes, which results in HIF stabilization in multiple different organs and cell types. Because oxoglutarate is an organic acid and organic anion transporters (OAT) are highly expressed in the kidney, we hypothesized that uptake and functional effects of hydrophilic 2OGAs could be limited to renal cells that bear these transporters. Transporters possibly involved in 2OGA handling are the sodium dicarboxylate transporter 3 (NADC3) and the transporters for organic anions 1 and 3 (OAT1, OAT3). These OATs are located in the basolateral membrane of proximal tubule cells and function as organic anion dicarboxylate exchangers.
Although OATs have already been described to be involved in drug/drug interaction and nephrotoxicity of drugs, they have to our knowledge not been used to explore treatment or prevention of renal injury. Here, we provide proof of concept for a selective drug targeting of proximal tubular cells in the kidney resulting in segment-specific cellular stabilization of the HIF proteins and activation of their downstream regulatory pathways and analyzed the role of the proximal tubule in HIF-mediated nephron protection against AKI.
Materials and Methods
Reagents
Cell culture reagents were obtained from Invitrogen (Karlsruhe, Germany), PAA Laboratories (Coelbe, Germany), and Biochrom (Berlin, Germany). Chemicals were used in analytical grade and purchased from Sigma-Aldrich (Taufkirchen, Germany), unless indicated otherwise.
Oxoglutarate Analogues
Pyridine-2,4-dicarboxylate (PDCA) was purchased from abcr (Karlsruhe, Germany), dimethyloxalylglycine (DMOG) was purchased from Cayman Chemicals (Ann Arbor, MI), and 2,2′-dipyridyl (DP) was purchased from ICN (Costa Mesa, CA).
N-Oxalylglycine (OG) was synthesized as monohydrate according to the literature,
and its purity was controlled by 1H and 13C NMR data as well as by elemental analysis:1H NMR (DMSO-d6, 300 MHz, 25°C): δ = 3.29 (d, 3JH,H = 6.2 Hz, 2 H), 8.99 (t, 3JH,H = 5.9 Hz, 1 H, -NH-) ppm; 13C NMR (D2O, 75.5 MHz, 25°C): δ = 41.1 (CH2), 160.4 (C = O), 161.8 (CO2H), 172.4 (CO2H) ppm; elemental analysis calculated for C4H5NO5 × H2O (165.10 g/mol): calc C 29.10, H 4.27, N 8.48; found C 29.38, H 4.32, N 8.06%.
2-(1-Chloro-4-hydroxyisoquinoline-3-carboxamido) acetic acid (ICA) was obtained in a six-step synthesis, starting from phthalic anhydride and methyl isocyanoacetate, according to a patent procedure
Weidmann K, Baringhaus K-H, Tschank G, Werner U: Substituted Isoquinolin-3-carboxyamides, their preparation and medical use. European Patent EP 1538160
Unusual mass spectrometric dissociation pathway of protonated isoquinoline-3-carboxamides due to multiple reversible water adduct formation in the gas phase.
Synthesis of heterocyclic-compounds using isocyano compounds 3. A facile synthesis of 1-oxo-1,2-dihydroisoquinoline-3-carboxylate and 2-pyridone-6-carboxylate derivatives.
Differential interaction of dicarboxylates with human sodium-dicarboxylate cotransporter 3 (NaDC3) and organic anion transporters 1 and 3 (OAT1 and OAT3).
Cells were selected by 10 μg/mL hygromycine and grown in high-glucose Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 1% penicillin/streptomycin, and 5 mg/L blasticidine. Control cells were transfected with the vector alone.
Human primary renal tubular epithelial cells were isolated from the kidney cortex of tumor nephrectomy specimens as previously described.
and ICA (40 mg/kg) were dissolved in 2.5% dimethyl sulfoxide, 0.5 M Tris buffer, OG (400 mg/kg), and PDCA (400 mg/kg) were dissolved in 0.5 M Tris buffer and injected intraperitoneally (injection volume, 200 μL).
Protein Extraction and Immunoblotting
Preparation of cell lysates and immunoblotting was performed as described earlier.
Proteins (100 μg) were separated on 10% polyacrylamide gels, transferred onto polyvinylidene difluoride membranes (Bio-Rad, Munich, Germany), and incubated with antibodies against HIF-1α (1:2000; Cayman Chemical), OAT1 (1:2000; Alpha Diagnostic, San Antonio, TX), and β-actin (1:5000; clone AC-74, Sigma-Aldrich). Subsequently, blots were exposed to horseradish peroxidase-conjugated secondary antibodies (Dako, Hamburg, Germany), followed by chemiluminescent detection (Amersham ECL-Plus; GE Healthcare, Buckinghamshire, UK). Densitometric analysis of protein bands was performed with ImageJ software version 1.46 (NIH, Bethesda, MD).
RNA Preparation and Real-Time PCR
Total RNA from cell culture experiments or animal tissue was extracted with peqGOLD Trifast reagent according to the manufacturer's protocol (Peqlab, Erlangen, Germany). cDNA was synthesized from 1 μg of total RNA and amplified in SYBR Green/Rox Master Mix (Fermentas, Leon-Rot, Germany) on an ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA). All reactions were performed in duplicates. Relative expression values were related to 18S ribosomal RNA with the use of the comparative ΔΔCt method. Primer sequences are given in Table 1.
Table 1Primer for PCR to Detect Human and Mouse mRNA Expression
Paraffin sections (2 to 4 μm) were dewaxed in xylene and rehydrated in a series of ethanol washes. The following primary antibodies were used for immunodetection: monoclonal mouse anti-human HIF-1α (1:10,000; α67; Novus Biologicals, Littleton, CO), polyclonal rabbit anti-mouse HIF-2α (1:10,000; PM9),
Biotinylated secondary anti-rabbit and anti-goat antibodies (Dako) were used in a dilution of 1:1000. Anti-HIF-1α, HIF-2α, and OAT1 antibodies were detected by catalyzed signal amplification system (Dako). The sections were counterstained with H&E, where appropriate.
Animal Experiments
All animal experiments were approved by local government authorities and conformed to the Guide for the Care and Use of Laboratory Animals as published by the U.S. National Institutes of Health. Male C57BL/6N mice, weighing 20 to 25 g, were purchased from Charles River (Sulzfeld, Germany).
Renal Ischemia-Reperfusion Injury
Renal ischemia-reperfusion injury (IRI) was induced by bilateral clamping of the renal pedicles for 25 minutes as previously described.
Mice were sacrificed 3 days after reperfusion under deep isoflurane anesthesia. Kidney tissues were divided to be either snap-frozen in liquid nitrogen or fixed by immersion in 4% neutral-buffered paraformaldehyde for paraffin embedding. Blood samples were obtained by inferior vena cava puncture. Plasma creatinine and urea were measured on a Cobas Integra 800 autoanalyzer (Roche Diagnostics, Mannheim, Germany).
Morphologic Analysis and Histologic Scoring
H&E-stained kidney sections were analyzed for morphologic criteria of acute tubular damage at a magnification of ×200, and results were graded with a semiquantitative score from 0 to 4 as defined before.
Histopathology was evaluated per section in 10 randomly selected, nonoverlapping fields by two nephropathologists independently in a blinded manner.
Statistical Analyses
All data represent means ± SEMs. Results were compared with two-tailed, unpaired t-test GraphPad Prism software version 5.04 (GraphPad Software Inc., San Diego, CA). Significance level was defined as P < 0.05.
Results
OAT1 Expression Enables Cellular Uptake of OG and HIF Stabilization in Isolated Tubular Cells
To test the hypothesis that proximal renal epithelial cells can be selectively targeted by hydrophilic 2OGAs, we first used the hydrophilic compound OG and its lipophilic derivative DMOG as prototypic 2OGAs. Both compounds were previously shown to inhibit the HIF PHDs in cell free systems, and DMOG was also effective in cellular systems and in vivo.
In human proximal tubular epithelial cells HKC-8 hypoxia and the lipophilic substances DP and DMOG, but not the hydrophilic OG, stabilized HIF-1α protein by Western blot analysis (Figure 1, A and B). Consistently mRNA levels of typical HIF target genes GLUT1, PDK1, VEGFA, and PHD3 were significantly induced only by hypoxia, DP, and DMOG, whereas OG-treated HKC-8 cells did not show any clear up-regulation (Figure 1C). Among the organic anion transporters only the OAT1 contributed to the uptake of OG across the basolateral cell membrane (Y. Hagos et al, manuscript accepted for publication). Therefore, we analyzed OAT1 protein and mRNA expression in HKC-8 cells, but OAT1 was not detectable, presumably as a result of dedifferentiation processes under tissue culture conditions (Figure 1, A and D). Furthermore, the organic anion transporters OAT1, OAT2, OAT3, and NADC3 were not found to be expressed in all renal tubular cell lines tested as determined by RT-PCR (Figure 1D). Concordantly, OG did also not stabilize HIF-1α protein in these cell lines.
Figure 1Dimethyloxalylglycine (DMOG) but not N-oxalylglycine (OG) induced HIF-1α accumulation in cultured renal epithelial cells. A: HKC-8 cells were exposed to DMOG and OG with the indicated concentrations for 6 hours. Cells subjected to 1% O2 or 100 μmol/L 2,2′-dipyridyl (DP) were used as positive controls for HIF-1α activation. Cellular HIF-1α, OAT1, and β-actin were determined by Western blot analysis. One representative blot of five independent experiments is shown. B: Densitometric analysis of HIF-1α protein signals of five independent experiments. Data were normalized to protein signals of β-actin which were used as loading control and expressed as arbitrary densitometric units. C: HKC-8 cells were incubated with DMOG 1 mmol/L and OG 500 μmol/L, respectively, for 6 hours. Relative mRNA levels of the HIF target genes GLUT1, PDK1, VEGFA, and PHD3 in relation to normoxic control cells were measured by real-time PCR (n = 5). D: Expression of the human organic anion transporters OAT1, OAT2, OAT3, and NADC3 in total kidney and renal epithelial cell lines (HKC-8, HEK293, IHKE-1, PTEC, Caki-2, VHL-deficient renal carcinoma cells RCC-VHL, and VHL-reconstituted RCC+VHL) was analyzed by PCR. β-Actin confirmed equal RNA loading. Means ± SEMs are shown. *P < 0.05 compared with control.
Therefore, we used HEK293 cells stably expressing human OAT1 for further in vitro analyses. In these cells OG resulted in a potent and dose-dependent stabilization of HIF-1α, indicating an OAT1-mediated transport of hydrophilic 2OGAs (Figure 2, A and C). This was consistent with the observation that uptake of para-aminohippurate, the prototypical substrate of OAT1, was inhibited by increasing concentrations of glutarate and 2-oxoglutarate.
Differential interaction of dicarboxylates with human sodium-dicarboxylate cotransporter 3 (NaDC3) and organic anion transporters 1 and 3 (OAT1 and OAT3).
Additional application of probenecid, a pharmacologic inhibitor of OAT1 transport, ablated OG-induced HIF-1α protein signals in OAT1-expressing HEK293 cells, further supporting the finding of OAT1-dependent transport of OG into these cells (Figure 2B). HIF-1α stabilization coincided with a significant up-regulation of the HIF target genes GLUT1, PDK1, VEGFA, and PHD3 quantified by real-time PCR. OAT1 mRNA itself was not induced by hypoxia or chemical HIF stabilizers (Figure 2D).
Figure 2OAT1 mediated transport of N-oxalylglycine (OG) into renal epithelial cells. A: HEK293 cells were stably transfected with an OAT1 expression construct (+OAT1; right) or the empty vector (−OAT1; left) and exposed to increasing concentrations of OG for 6 hours. HIF-1α, OAT1, and β-actin protein were detected by immunoblot. Data shown are representative for three independent experiments. B: OAT1-expressing HEK293 cells were preincubated with 500 μmol/L probenecid (Prob) as an inhibitor of OAT1 transport before stimulation with 500 μmol/L OG. Protein signals of HIF-1α, OAT1, and β-actin were analyzed by Western blotting. C: Densitometric analysis of HIF-1α protein levels in vehicle (n = 3; left panel) and OAT1-transfected HEK293 cells (n = 3; right panel). Signal intensities were normalized to that of β-actin and were expressed as arbitrary densitometric units. D: OAT1-expressing HEK293 cells were incubated with dimethyloxalylglycine (DMOG) 1 mmol/L or OG 500 μmol/L for 6 hours. GLUT1, PDK1, VEGFA, PHD3, and OAT1 relative mRNA levels were determined by real-time PCR (n = 5). Means ± SEMs are shown. *P < 0.05 compared with control.
HIF Stabilization by OG Is Restricted Primarily to Proximal Tubules in Mice
Next, we analyzed whether OG was taken up in vivo in the mouse kidney and whether transport depended on OAT1 expression. Therefore, we injected i.p. DMOG and OG into mice and evaluated renal HIF-1α signals with the use of immunohistochemistry. One to 3 hours after DMOG injection marked HIF-1α signals were seen in tubular epithelial cells of the renal cortex and outer medulla (Figure 3A) and also for HIF-2α in interstitial cells (Figure 3D) and glomerular cells, as described previously.
In contrast, 1 hour after OG injection expression of HIF-1α in the kidney was mainly restricted to tubular cells in the cortex (Figure 3, B and C), whereas no HIF-2α signals were detectable (Figure 3E). Staining of tubular cells for HIF-1α predominantly colocalized with OAT1 (Figure 3B) and NaPi-IIa staining (Figure 3C). However, a minority of TAL and collecting duct cells, verified by immunostaining for Tamm-Horsfall protein and ENaC β-subunit, were HIF-1α positive too, potentially implying additional mechanisms for OG uptake in the mouse. After DMOG injection HIF-1α signals were still detectable 6 to 24 hours later. In comparison only a few positive cells were detectable 2 hours after OG injection, indicating a significantly shorter lasting effect.
Figure 3HIF accumulation in mice kidney after N-oxalylglycine (OG) and dimethyloxalylglycine (DMOG) injection. Consecutive kidney sections were immunostained for HIF-1α (A, B, and C, top row), OAT1 (A and B, bottom row), and sodium phosphate cotransporter 2a (NaPi-IIa), respectively (C, bottom row). Mice had been intraperitoneally injected once with DMOG 3 hours (A and D) or OG 1 hour (B, C, and E) before kidney removal. Original magnification: ×400 (A–C); interference-contrast microscopy (D and E).
With the use of RT-PCR we indentified Oat1, Oat2, Oat3, and NaDC3 in the brain and kidney, in addition Oat2 in the liver and Oat3 in the intestine of C57BL/6N mice (Figure 4A).
Figure 4OAT expression and HIF-1α accumulation in different organs. A: Expression of the organic anion transporter genes Oat1, Oat2, Oat3, and NaDC3 mRNA in different organs of C57BL/6N mice was analyzed by RT-PCR. B: Immunohistochemical localization of HIF-1α in cerebral cortex, small intestine, liver (original magnification, ×200), lung, and heart (original magnification, ×400) of mice injected with dimethyloxalylglycine (DMOG; 3 hours; left column) and N-oxalylglycine (OG; 1 hour; right column), respectively.
To check for extrarenal HIF activation after OG administration in vivo, we performed immunohistochemistry for HIF-1α in different mouse organs 3 hours after DMOG and 1 hour after OG injection. In animals injected with DMOG an ubiquitous HIF stabilization in most organs tested was seen, whereas OG administration did not result in significant HIF-1α signals in the brain, liver, lung, and heart. Only some intestinal epithelial cells were HIF-1α positive (Figure 4B), potentially indicating an OAT1-independent uptake mechanism of OG in the mouse.
Use of PDCA and ICA as HIF Stabilizers
Although the prototypic 2OGAs OG and DMOG enabled us to stabilize HIF in different cell populations, their inhibitory potential and selectivity for HIF PHDs was comparatively low.
We therefore aimed to use another two 2OGAs with predominant lipophilic or hydrophilic properties, respectively, to target HIF PHDs in vivo. Comparing potential substances, which incorporate the oxoglutarate backbone and proved to have HIF stabilizing effects in vitro, we chose two compounds that are presently known to have high binding affinity to PHD2, which is the functionally most important PHD isoform: PDCA, a hydrophilic 2OGA (Figure 5A), and the lipophilic 2OGA ICA (Figure 5B). Affinities of both compounds to PHD2 are 10- to 20-fold higher than those of OG or DMOG with a half maximal inhibitory concentration (IC50) of 1.91 for PDCA and 0.07 μmol/L for ICA, compared with 18.5 μmol/L for OG and DMOG.
Figure 5Pyridine-2,4-dicarboxylate (PDCA) and 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetic acid (ICA) induce HIF-1α in mice. Chemical formulae for PDCA (A) and ICA (B). C: Western blot analysis for HIF-1α in HKC-8 cells after exposure to increasing concentrations of PDCA and ICA for 6 hours. Data shown are representative for five independent experiments. D: Immunoblot analysis for HIF-1α and OAT1 in HEK293 cells with inducible expression of OAT1. PDCA 500 μmol/L stabilizes HIF-1α only in OAT1-expressing cells (+OAT1). Lanes were run on the same gel but were not contiguous. Cell lysates from three independent experiments were pooled and analyzed. Immunohistochemistry for HIF-1α (E) and sodium phosphate cotransporter 2a (NaPi-IIa; F) in kidney of mice 1 hour after injection of PDCA. Immunohistochemistry for HIF-1α (G) and HIF-2α (H) in mouse kidney 3 hours after injection of ICA. Original magnification, ×400 (E–H); interference-contrast microscopy (H).
Both substances in concentrations up to 1000 μmol/L were not toxic in lactate dehydrogenase release assays in isolated primary tubular epithelial cells. In parallel to results obtained with DMOG and OG (Figure 1A), immunoblotting of HKC-8 cell lysates showed induction of HIF-1α in ICA- but not in PDCA-treated cells (Figure 5C). After PDCA exposure HIF-1α accumulated only in stably OAT1-transfected HEK293 cells (Figure 5D), as shown for OG in Figure 2A. When we injected PDCA into mice, the HIF-1α expression patterns were comparable with results obtained with OG administration. HIF-1α signals were mainly found in the cortex (Figure 5E) and primarily within proximal tubules, as verified by immunostaining for NaPi-IIa (Figure 5F). Similar to OG, 2 hours after injection of PDCA HIF-1α signals were mostly not detectable any more. In contrast, 1 hour after injection of ICA we detected HIF-1α signals in nuclei of tubular epithelial cells of virtually all nephron segments (Figure 5G). Immunohistochemistry staining varied across different nephron segments, with particularly strong HIF-1α signals in collecting ducts, moderate intensity in TALs and distal convoluted tubules, and less intensity in proximal tubules, as verified by immunostaining of consecutive kidney sections with antibodies against β-ENaC, Tamm-Horsfall protein, sodium chloride cotransporter, and NaPi-IIa, respectively. HIF-1α signals after ICA injection increased over a period of 3 hours, were constantly expressed for 12 hours, and were still markedly detectable in the inner medulla 24 hours after i.p. injection. ICA, but not PDCA, also stabilized HIF-2α in interstitial fibroblasts (Figure 5H) and glomerular cells (Figure 5H). We concluded that in analogy to OG, PDCA induced HIF-1α predominantly in proximal renal tubules, whereas ICA induced HIF-1α in all nephron segments with preponderance of the distal nephron and, in addition, HIF-2α in interstitial and glomerular cells.
Differential HIF Targeted Gene Expression in Kidneys of PDCA- and ICA-Treated Mice
We next tested for effects of the 2OGAs PDCA and ICA on target gene expression 6 hours after i.p. injection in mice (Figure 6, A and B). Because the HIF-stabilizing effect of PDCA was shorter than that of ICA, PDCA was injected three times (6, 4, and 2 hours before kidney removal), whereas ICA was applied only once. Both ICA and PDCA markedly induced Glut1, Pdk1, and Phd3 mRNA in whole kidney extracts. In contrast Epo, which is regulated by HIF-2α in peritubular fibroblasts,
Hypoxia-inducible factor-2alpha-expressing interstitial fibroblasts are the only renal cells that express erythropoietin under hypoxia-inducible factor stabilization.
Differentiating the functional role of hypoxia-inducible factor (HIF)-1alpha and HIF-2alpha (EPAS-1) by the use of RNA interference: erythropoietin is a HIF-2alpha target gene in Hep3B and Kelly cells.
was only up-regulated by ICA and not by PDCA, consistent with the observation that only ICA resulted in detectable HIF-2α protein signals in interstitial fibroblasts (Figure 5H). Then, we directly compared the mRNA level of target genes in time-course experiments. At 6 hours after injection both PDCA and ICA markedly up-regulated Phd3 and Vegfa to a similar extent. In line with the time course of HIF-1α stabilization, Phd3 and Vegfa were still significantly induced 24 hours and slightly elevated 72 hours after ICA administration, whereas PDCA had no detectable effect at 24 and 72 hours after injection (Figure 6C). Consistent with the data obtained in HEK293+OAT1 cells for DMOG and OG (Figure 2D), Oat1 was not induced at any time point tested by either ICA or PDCA in mouse kidneys (Figure 6C).
Figure 6Induction of HIF target genes in kidneys after pyridine-2,4-dicarboxylate (PDCA) and 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetic acid (ICA) injection. mRNA expression level of Glut1, Pdk1, Phd3, and Epo in total kidney lysates were determined by real-time PCR 6 hours after injection of (A) PDCA (n = 6) and (B) ICA (n = 6) in comparison with vehicle-injected control mice (n = 4 each). C: Kidney Phd3, Vegfa, and Oat1 mRNA expression 6, 24, and 72 hours after administration of PDCA, ICA, or vehicle (n = 3 each per group and time point). Means ± SEMs are shown. *P < 0.05 compared with control.
On the basis of the above data we next aimed to evaluate whether selective HIF stabilization in proximal tubular cells was sufficient to induce protection against IRI. We applied PDCA and ICA before clamping both renal pedicles for 25 minutes. Three days later we assessed kidney function, histomorphology, and tissue expression of biomarkers of renal injury. Treatment with ICA, given once 6 hours before IRI, significantly reduced plasma creatinine and urea levels in comparison with vehicle treatment (0.22 ± 0.03 versus 0.39 ± 0.07 mg/dL; P = 0.03; and 117.6 ± 11.0 versus 189.4 ± 23.6 mg/dL; P = 0.009, respectively; Figure 7A). Correspondingly, semiquantitative histologic evaluation showed markedly better preserved renal structure in treated animals (1.69 ± 0.05 versus 2.26 ± 0.10; P < 0.0001; Figure 7B). ICA treatment also significantly reduced kidney tissue expression of Ngal mRNA, which reflects distal tubular injury.
Kidney injury molecule-1 (KIM-1), a putative epithelial cell adhesion molecule containing a novel immunoglobulin domain, is up-regulated in renal cells after injury.
also tended to be lower, but the difference did not reach statistical significance (Figure 7C). In contrast, treatment with PDCA 6, 4, and 2 hours before IRI had no effects compared with vehicle on plasma creatinine (0.47 ± 0.13 versus 0.45 ± 0.07 mg/dL) and urea (197.3 ± 34.7 versus 204.7 ± 22.9 mg/dL) 3 days after ischemia (Figure 7A). In addition, histomorphologic scores (Figure 7B) and biomarker mRNA induction (Figure 7C) were not different, overall indicating no structural or functional preservation of kidneys treated with PDCA.
Figure 7Assessment of kidney function and structure 3 days after ischemia-reperfusion injury in mice treated with 2-oxoglutarate analogue (2OGA). A: Plasma creatinine and urea in mice preconditioned with pyridine-2,4-dicarboxylate (PDCA; n = 16) or 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetic acid (ICA; n = 12) in comparison with corresponding vehicle-injected (n = 12 for PDCA, and n = 10 for ICA) and sham-operated mice (n = 4 each). B: Semiquantitative histologic scoring of acute tubular damage in sham-operated (n = 6 each), PDCA- (n = 30) or ICA-pretreated (n = 22), and vehicle-injected kidneys (n = 20 each for PDCA and ICA). C: mRNA expression of the renal biomarkers Ngal and Kim1 in postischemic kidneys measured by real-time PCR in PDCA- or ICA-pretreated (n = 6 each) in comparison with vehicle-injected (n = 6 each) and sham-operated (n = 3 each) kidneys. Means ± SEMs are shown. *P < 0.05 compared with control. ATN, acute tubular necrosis.
This study indicates that hydrophilic 2OGAs, such as OG and PDCA, resulted in HIF-1α expression that was confined to the kidney and particularly to proximal tubular cells. The HIF-1α expression pattern largely overlapped with renal tubular expression of OAT1 and together with in vitro results suggested that these compounds are transferred into the cell via the OAT1. With the use of ICA, a 2OGA with higher affinity to PHD2 as compared with DMOG, we were able to confirm that 2OGAs with lipophilic properties resulted in efficient stabilization of HIF-1α and HIF-2α in a wide variety of organs and cell types. As shown previously for systemic hypoxia and other PHD inhibitors, HIF expression was not uniform across different cell types, implying important additional mechanisms of regulation.
To test the functional significance of selective HIF stabilization in proximal tubular cells, we used an ischemia-reperfusion model of AKI. Although structural damage predominantly affects proximal tubular cells after IRI,
With the use of ICA pretreatment we were able to reapprove previous studies showing that administration of lipophilic PHD inhibitors before an ischemic insult protected the kidney and ameliorated structure and function.
The protective effect of prolyl-hydroxylase inhibition against renal ischaemia requires application prior to ischaemia but is superior to EPO treatment.
Surprisingly, PDCA was ineffective in this respect and resulted in no detectable difference in renal structure or function. Despite the short-lasting HIF-1α-stabilizing effect of PDCA, the mRNA levels of the HIF target genes Phd3 and Vegfa were comparable with ICA at 6 hours after injection. Although we observed a more persisting Phd3 and Vegfa induction after ICA administration after 24 hours, this was unlikely to explain why the protective effect was only observed with ICA, because ICA administration after an ischemic insult was ineffective.
The protective effect of prolyl-hydroxylase inhibition against renal ischaemia requires application prior to ischaemia but is superior to EPO treatment.
Therefore, differences in AKI protection between the two different compounds were less likely because of differences in target gene kinetics but rather were related to the type of cells in which HIF was induced. This implies that distal tubular or nontubular cells, such as peritubular endothelial cells or interstitial fibroblasts, in which only ICA but not PDCA stabilized HIF-2α, presumably mediated the protective effects on HIF stabilization. A number of previous observations may be relevant in context of these findings. i) Kojima et al
found that HIF-2α in the renal endothelium plays a protective role in ischemic AKI. ii) HIF-2α mainly regulates erythropoietin (EPO) synthesis in peritubular cortical fibroblasts.
Hypoxia-inducible factor-2alpha-expressing interstitial fibroblasts are the only renal cells that express erythropoietin under hypoxia-inducible factor stabilization.
Hydrophilic 2OGAs, in contrast to their lipophilic counterparts, were apparently unable to enter these cells, resulting in a lack of HIF-2α stabilization and Epo induction. EPO has previously been reported to convey protection against IRI,
The protective effect of prolyl-hydroxylase inhibition against renal ischaemia requires application prior to ischaemia but is superior to EPO treatment.
iii) At least for the HIF-mediated nephroprotection the proximal renal tubule does not seem to be the primary target. HIF induction in the distal tubule and in particular in the TAL may be essential for the protective effect instead. We recently reported that selective HIF stabilization in TALs induced by site-specific Vhl knockout mediated protection against AKI.
iv) Eventually, protective extrarenal signals could also have been generated by HIF expression outside the kidneys in response to ICA treatment and could theoretically convey remote preconditioning. Genetic models of cell-specific HIF stabilization will probably provide the best tools to further explore the importance of different renal cell types for HIF-dependent protection.
The results of the present study may also offer a novel approach to assess the role of HIF in chronic progression of renal injury. Although hypoxia is widely considered as an important pathomechanism in kidney disease,
its precise role is difficult to evaluate and effects of HIF remain controversial. Although some investigators reported evidence for protective effects of HIF induction,
Importantly, pharmacologic in vivo testing of long-term effects of HIF stabilization has so far been difficult because lipophilic 2OGAs induce erythrocytosis that is associated with morbidity (eg, thrombosis) and mortality. The selective targeting of tubular cells with PDCA or other hydrophilic 2OGAs will circumvent this obstacle.
Finally, hydrophilic 2OGAs may also be re-evaluated as inhibitors of collagen prolyl hydroxylases. Inhibition of the collagen-4 prolyl hydroxylase by 2OGAs was successfully applied in several studies to reduce development of fibrosis as a hallmark of progressive organ dysfunction. Thus, 2OGAs have been applied to ameliorate hepatic cirrhosis,
Inhibition of collagen synthesis with prolyl 4-hydroxylase inhibitor improves left ventricular function and alters the pattern of left ventricular dilatation after myocardial infarction.
Although some of the compounds used in these studies were subsequently also found to stabilize HIF, selectivity for HIF prolyl hydroxylases over collagen hydroxylases is currently attempted for compounds developed for management of long-term anemia.
Conversely, compounds with preferential inhibition of collagen prolyl hydroxlases could potentially be generated and targeted to tubular cells to modulate synthesis of tubular collagen.
In summary, this study proves feasibility to specifically target diverse tubular segments with 2OGAs and to stabilize HIF or to block other 2-oxoglutarate-dependent enzymes such as the collagen-4 prolylhydroxylase selectively in proximal tubular epithelial cells. Regional HIF stabilization offers a new tool to investigate renal pathophysiology and is likely to be valuable for exploring the long-term effects of HIF induction in the kidney without inducing erythrocytosis and with minimal effects on other organs.
Acknowledgments
We thank Andrea Dengler, Hans Fees, Miriam Reutelshöfer, and Brigitte Rogge for expert technical assistance, Dr. Lorraine C. Racusen (Baltimore, MD) for providing the HKC-8 cell line, and Drs. Patrick H. Maxwell (London, UK), David H. Ellison (Portland, OR), Heini Murer (Zurich, Switzerland), and Christoph Korbmacher (Erlangen, Germany) for providing us with antibodies for HIF-2α, sodium chloride cotransporter, NaPi-IIa, and β-ENaC, respectively.
References
Brezis M.
Rosen S.
Hypoxia of the renal medulla–its implications for disease.
The protective effect of prolyl-hydroxylase inhibition against renal ischaemia requires application prior to ischaemia but is superior to EPO treatment.
Weidmann K, Baringhaus K-H, Tschank G, Werner U: Substituted Isoquinolin-3-carboxyamides, their preparation and medical use. European Patent EP 1538160
Unusual mass spectrometric dissociation pathway of protonated isoquinoline-3-carboxamides due to multiple reversible water adduct formation in the gas phase.
Differential interaction of dicarboxylates with human sodium-dicarboxylate cotransporter 3 (NaDC3) and organic anion transporters 1 and 3 (OAT1 and OAT3).
Hypoxia-inducible factor-2alpha-expressing interstitial fibroblasts are the only renal cells that express erythropoietin under hypoxia-inducible factor stabilization.
Differentiating the functional role of hypoxia-inducible factor (HIF)-1alpha and HIF-2alpha (EPAS-1) by the use of RNA interference: erythropoietin is a HIF-2alpha target gene in Hep3B and Kelly cells.
Kidney injury molecule-1 (KIM-1), a putative epithelial cell adhesion molecule containing a novel immunoglobulin domain, is up-regulated in renal cells after injury.
Inhibition of collagen synthesis with prolyl 4-hydroxylase inhibitor improves left ventricular function and alters the pattern of left ventricular dilatation after myocardial infarction.
Supported by a grant from the Deutsche Forschungsgemeinschaft (SFB 423, TP A14 to K.U.E. and C.W.) and the University Erlangen-Nuremberg (Emerging Field Initiative: Medicinal Redox Inorganic Chemistry to N.B.).
00575-5&id=gr1.jpg){kind=link}
00575-5&id=gr2.jpg){kind=link}
00575-5&id=gr3.jpg){kind=link}
00575-5&id=gr4.jpg){kind=link}
00575-5&id=gr5.jpg){kind=link}
00575-5&id=gr6.jpg){kind=link}
00575-5&id=gr7.jpg){kind=link}