We examined whether antagonism of the CXCR4 receptor ameliorates the loss of renal function following ischemia-reperfusion. CXCR4 is ubiquitously expressed on leukocytes, known mediators of renal injury, and on bone marrow hematopoietic stem cells (HSCs). Plerixafor (AMD3100, Mozobil) is a small-molecule CXCR4 antagonist that mobilizes HSCs into the peripheral blood and also modulates the immune response in in vivo rodent models of asthma and rheumatoid arthritis. Treatment with plerixafor before and after ischemic clamping ameliorated kidney injury in a rat model of bilateral renal ischemia-reperfusion. Serum creatinine and blood urea nitrogen were significantly reduced 24 h after reperfusion, as were tissue injury and cell death. Plerixafor prevented the renal increase in the proinflammatory chemokines CXCL1 and CXCL5 and the cytokine IL-6. Flow cytometry of kidney homogenates confirmed the presence of significantly fewer leukocytes with plerixafor treatment; additionally, myeloperoxidase activity was reduced. AMD3465, a monocyclam analog of plerixafor, was similarly renoprotective. Four weeks postreperfusion, long-term effects included diminished fibrosis, inflammation, and ongoing renal injury. The mechanism by which CXCR4 inhibition ameliorates AKI is due to modulation of leukocyte infiltration and expression of proinflammatory chemokines/cytokines, rather than a HSC-mediated effect. The data suggest that CXCR4 antagonism with plerixafor may be a potential option to prevent AKI.
- renal ischemia-reperfusion
- immune modulation
acute kidney injury (AKI) is a major unmet medical need impacting >700,000 patients annually. It is a significant healthcare burden (12), and no pharmacological agents exist for prevention or treatment (27). Mortality is high, ranging from 50 to 80% and has not changed in over four decades. Recent data strikingly demonstrate that ∼70% of elderly patients with AKI are at risk of progressing to chronic kidney disease (CKD) (25). Alternatively and in addition, CKD is a risk factor for AKI (20, 23).
Inflammation is a significant pathophysiological component of AKI. In response to reperfusion injury and nephrotoxicity, cytokines and chemokines are elaborated, creating a proinflammatory microenvironment. The CXC chemokine, CXCL12, also known as stromal-derived factor 1 (SDF-1), increases in expression (59), and has been demonstrated in vitro to be chemotactic for a number of leukocyte populations, including neutrophils, monocytes, and lymphocytes (4, 49, 58), key mediators of renal injury (for review, see Refs. 6 and 30). CXCL12 is the sole ligand for the G protein-coupled receptor CXCR4 (4, 49), which is ubiquitously expressed by leukocytes (48). CXCR4 contributes to the chemotaxis, activation, and homeostasis of these cells (48). It is also expressed on hematopoietic stem cells (HSCs) within the bone marrow. Here, CXCR4 interactions with SDF-1 result in stem cell retention and maturation within the bone marrow niche (1, 48). CXCR4 is also expressed on endothelial progenitor cells (EPCs) (28). Both HSCs and EPCs have been demonstrated to be renoprotective in preclinical rodent models of AKI (19, 35, 38, 41, 51).
Plerixafor (AMD3100, Mozobil) is a small-molecule CXCR4 antagonist (17, 21) approved in the United States, European Union, and a number of other countries for use in combination with granulocyte colony-stimulating factor (G-CSF) to mobilize HSCs to the peripheral blood for collection and subsequent autologous transplantation in patients with multiple myeloma and non-Hodgkin's lymphoma (15). It disrupts CXCR4/CXCL12 interactions, mobilizing HSCs from the bone marrow into the systemic circulation.
Beneficial effects of plerixafor have also been reported in preclinical rodent models of ischemia-reperfusion injury (IRI) and inflammation. In a mouse model of myocardial infarction, a bolus administration of plerixafor enhances marrow progenitor cell recruitment to sites of neovascularization, resulting in preservation of myocardial function and integrity (28). Similarly, in a mouse model of hindlimb ischemia, plerixafor with or without G-CSF increases mobilization of bone marrow mononuclear cells to enhance angiogenesis (11, 26). Antagonism of CXCR4 also modulates the immune system. Here, plerixafor inhibits in vivo autoimmune joint inflammation presumably through interference with the migration of Mac-1+CXCR4+ leukocytes to the arthritic joint (14, 46). It also attenuates allergic lung inflammation, decreasing eosinophilia and lymphocyte accumulation, with an associated shift of cytokines and chemokines from a Th2 toward a Th1 response (45). Plerixafor also attenuates inflammatory cell recruitment after myocardial infarct (28), significantly lowering mRNA expression of CD68, a monocyte/macrophage marker, in the ischemic border zone of the heart.
By antagonizing the CXCR4/CXCL12 pathway, we hypothesized that plerixafor could improve the outcome from AKI via either stem cell-dependent and/or -independent (anti-inflammatory) mechanisms. Using a rodent model of renal IRI, we found that when administered at the time of ischemic clamping, plerixafor attenuates the renal failure that develops within 24 h postreperfusion. Secondary effects include a reduction in renal fibrosis, inflammation, and putative biomarkers of renal injury.
MATERIALS AND METHODS
All studies involving animals were approved by the Genzyme Institutional Animal Care and Use Committee (IACUC).
Pharmacokinetic Profiling of Plerixafor
Two studies were conducted. In the first, four male Wistar rats were dosed subcutaneously (sc) with [14C]plerixafor at a single dose of 12.1 mg/kg with blood sampling via cannulation of the femoral artery at 0.5, 1, 2, 4, 7, and 24 h postdosing. In the second study, two groups of four male Wistar rats were dosed either with a single dose of 1 mg/kg, or with a dose of 1 mg/kg daily for 7 days with blood sampling at 0.5, 1, 2, 4, 7, 24, 31, 48, 55, 72, and 168 h postdosing. Plerixafor concentration was determined in plasma by liquid chromatography-reverse isotope dilution (LC-RID). A standard concentration of cold plerixafor was added to the plasma sample, and samples were extracted with 1% trifluoracetic acid in methanol and reconstituted in water. The samples were run on a RP18 column, and fractions corresponding to plerixafor were analyzed by liquid scintillation counting. Pharmacokinetic analysis was performed using WinNonlin V.4.01 (Pharsight, Mountain View, CA).
White Blood Cell Mobilization
After various times and doses of plerixafor, whole blood was collected by retroorbital bleeding and analyzed for circulating white blood cells (WBCs) using a Sysmex XT-V Automated Hematology Analyzer (Sysmex, Kobe, Japan).
Rat Model of Bilateral Renal IRI
Male Sprague-Dawley rats (190–220 g, Taconic Laboratories, Wilmington, MA) were acclimated at the animal facility at Genzyme for 3–5 days. Animals were anesthetized (50 mg/kg ketamine/6 mg/kg xylazine im), injected intraperitoneally with sterile saline (8 ml at 37°C), prepped for aseptic surgery, and placed on heated (37°C) rodent operating tables (Harvard Apparatus, Holliston, MA). Bilateral flank incisions were made, renal pedicles were cleaned of adherent connective tissue, and the renal artery and vein were clamped with nontraumatic micro-aneurysm clamps (Roboz Surgical Instruments, Gaithersburg, MD). Peritoneal and skin incisions were temporarily closed, and animals were placed into an Intensive Care Unit (Harvard Apparatus) at 37°C; body temperature was monitored rectally. After 45 min of renal ischemia, clamps were removed, and kidneys were visually inspected for reperfusion of blood. Animals were injected with buprenorphine (0.06 mg/kg) 3.5 h after reperfusion with additional pain medication every 12 h for the first 48 h after surgery, unless euthanized at the 24-h time point.
Animals were euthanized (Euthasol, Virbac, Fort Worth, TX) and perfused with HBSS containing protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO). Kidneys were removed and dissected free of adherent connective tissue and cut from the cortex to the hilus into five equal pieces. Samples were then flash frozen in liquid nitrogen for molecular and biochemical analyses. For morphology, the perfusate was switched to a solution of 2% paraformaldehyde-75 mM dl-lysine-10 mM sodium periodate. Kidneys were removed, cut as described above, and fixed kidney pieces were embedded in paraffin for staining of sections with hematoxylin and eosin (H&E) and Masson's trichrome. Alternatively, kidney pieces were processed for frozen sectioning by embedding in Optimal Cutting Temperature (OCT) compound (Sakura Finetek, Torrance, CA) followed by rapid freezing in liquid nitrogen and storing at −80°C as described previously (65). In some instances, unfixed kidneys were immediately embedded in OCT and frozen.
Plerixafor (AMD3100, Mozobil; Genzyme, Cambridge, MA) was injected sc at various doses and regimens (see results) at a volume of 3 ml/kg. Parallel experiments were also conducted with AMD3465, a structurally related analog with comparable pharmacokinetic and pharmacodynamic properties (Genzyme). Doses were selected based on toxicity assessment in mouse and converted to those for rat using allometric scaling factor (data not shown). For studies with G-CSF [neupogen (Filgrastim), Amgen, Thousand Oaks, CA], animals were injected sc for 5 days at 100 μg/kg.
Measurement of Renal Function
Blood samples were drawn immediately before euthanasia from the retroorbital sinus into 3.6-mg EDTA-coated tubes (BD Biosciences, Franklin Lakes, NJ). Plasma was evaluated for creatinine and urea nitrogen using an Integra 400 Automated Analyzer (Roche Diagnostics, Indianapolis, IN) or a Beckman Synchron CX5 Bioanalyzer (Beckman Coulter, Brea, CA).
Paraffin-embedded 5-μm kidney sections stained with H&E were evaluated for injury on a scale of 0–4, with 0 representing no injury; 1, <25% injury; 2, 25–50% injury; 3, 50–75% injury; and 4, 75–100% injury. Four different fields per section of the outer medulla were blindly scored at ×10 magnification.
Paraffin sections (5 μm) stained with Masson trichrome were evaluated for fibrosis in the corticomedullary region. Fibrosis was blindly scored at ×4 magnification on a scale of 0–4, with 0 representing no fibrosis; 1, <25% fibrosis; 2, 25–50% fibrosis; 3, 50–75% fibrosis; and 4, 75–100% fibrosis.
For immunostaining frozen sections (5 μm) cut on a Leica CM1850 cryostat (Leica Microsystems, Buffalo Grove, IL), the protocol of Zuk et al. (64) was used with some modifications. These included postfixation in 3.2% paraformaldehyde in PBS and detection of antigen:antibody complexes with FITC- or Cy3-conjugated secondary antibody (goat anti-rabbit IgG, donkey anti-mouse IgG, or goat anti-mouse IgG) cross adsorbed against various species (Jackson ImmunoResearch, West Grove, PA) for 1 h at room temperature (RT). Sections were mounted in Vectashield Hard Set Mounting Medium with or without 4,6-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA), viewed on a Nikon Eclipse 80i fluorescent microscope (Melville, NY) and photographed with Spot Advanced software (Diagnostic Instruments, Sterling Heights, MI).
Primary antibodies included rabbit anti-rat IgG (1:100), rabbit anti-rat albumin (1:100; Abcam, Cambridge MA), mouse monoclonal anti-rat CD68 (1:50; clone ED1; Millipore, Billerica, MA), and monoclonal anti-human/mouse CXCL12/SDF-1 (clone 79018, R&D System, Minneapolis, MN) used at 5 μg/ml on unfixed kidneys permeabilized with acetone (−20°C, 8 min). Controls for immunohistochemistry were as previously described (64, 65).
To analyze progammed cell death (44), DNA double-strand breaks were detected by transferase-mediated dUTP nick-end labeling (TUNEL) staining with an ApopTag Fluorescein Direct in situ detection kit (Chemicon, Temecula, CA) following the manufacturer's protocol.
Quantification of immunoreactivity was carried out by photographing four to five fields of the outer medulla per section per animal followed by Metamorph imaging (Metamorph Imaging Series Software, Molecular Devices, Dovington, PA). For IgG and albumin, images were recorded at ×40, whereas for TUNEL staining and CD68, ×20 was used.
Total RNA was extracted with an RNeasy Maxi Kit (Qiagen, Valencia, CA). Snap-frozen tissues were lysed for 2 min at RT in RLT Lysis Buffer containing β-mercaptoethanol using zirconium beads (Biospec, Bartlesville, OK) at a 1:5 ratio in a bead beater (Biospec). The lysate was then purified using an RNeasy column, dissolved in RNAse-free water, and RNA quality and concentration were determined. Total RNA (2 μg) was reverse-transcribed to cDNA by Clontech Sprint RT Reagent (Clontech, Mountain View, CA) using random hexamer primers with the following cycling conditions: 42°C for 60 min, 70°C for 15 min, 4°C end. Quantitative (q) PCR was done using 4 μg of newly synthesized cDNA, TaqMan RT-PCR Mastermix (Applied Biosystems, Foster City, CA), and FAM/TAMRA-labeled primers/probes following a standard RT-PCR. Premade probes and primers specific for rat (Applied Biosystems) were as follows: KIM1 (Havcr1) NM_173149.1, Ngal (Lcn2) NM_130741.1, Il6 NM_012589.1, Cxcl5 NM_022214.1, CXCl1 NM_008176.3, MCP1/Ccl2 NM_031530.1, Il10 NM_012854.2, Cxcl12 NM_053647.1, Col3a1 NM_032085.1, Tgfb1 NM_021578.2, Fn1 NM_019143.2, Col4a1 NM_00113509.1, Il18 NM_019165.1, Il1b NM_031512.2, and Tnfa NM_012675.3.
Flow Cytometry of Kidney Homogenates
Freshly perfused rat kidneys were minced and incubated with 20 μg/ml collagenase type IA (Sigma-Aldrich) in ice-cold Dulbecco's PBS with 2 mM EDTA for 15 min at 37°C. The digested kidney suspension was filtered (100- and 40-μm Falcon cell strainers, BD Biosciences) and centrifuged (1,200 rpm for 10 min). The cell pellet was washed (PBS containing 1% BSA) and lysed [1× lysis buffer (BD Biosciences) for 5 min, RT]. After further washing, the number of cells was counted (ViCell XR, Beckman Coulter). Then, 1 × 106 cells/reaction were blocked (5 min, RT) with nonspecific Fc binding anti-mouse CD16/32 (clone 2.4G2, BD Biosciences) and reacted with primary antibody (2 μg/100 μl flow cytometry reaction) for 20 min at 4°C in the dark. After washing, samples were stained with Live/Dead Fixable Cell Stain (Invitrogen, Carlsbad, CA) for 30 min at RT in the dark to allow gating only of viable cell populations. To determine absolute cell counts, Count Bright Absolute Counting Beads (Invitrogen) were added to each reaction per the manufacturer's protocol. Flow cytometric analysis was performed using a FACS LSRII flow cytometer (BD Biosciences) and analyzed using FlowJo software (TreeStar, Ashland, OR).
Mouse anti-rat antibodies (BD Biosciences) included CD45 PE (clone OX-1), CD11b FITC (clone WT.5), CD4 APC (clone OX-35), CD8a FITC (Clone OX-8), CD45RA PE (clone OX-33), Granulocyte FITC (HIS 48), and isotype-matched controls.
Colony-Forming Unit Assay
One milliliter of peripheral blood was lysed in 1× lysis buffer (BD Biosciences) for 15 min at RT in the dark followed by washing remaining WBCs three times with PBS containing 2% FBS. Viable WBCs were counted (ViCell XR) and mixed with Methocult GF methylcellulose medium (StemCell Technologies, Vancouver, British Columbia) containing 100 U/ml penicillin and 100 μg/ml streptomycin to give a final density of 1 × 105 viable WBCs/1.5 ml of medium. The Methocult-cell mixture was dispensed into Ultra Low Cell Adherence Surface six-well plates (Corning, Lowell, MA), and cultures were incubated at 37°C, 5% CO2, and 95% humidity for 10–12 days. Three wells were plated per animal. Colony-forming unit-granulocyte/macrophages (CFU-GM) were counted using a Ziess Axiovert 200 inverted microscope (Carl Zeiss Microscopy, Thornwood, NY) at ×2.5. The number of CFU-GM was determined by dividing the number of viable cells per milliliter by the number of plated cells per well and multiplying by the number of counted CFUs/well. Positive controls included freshly harvested cells from bone marrow and peripheral blood of G-CSF-treated normal rats.
To measure polymorphonuclear leukocyte infiltration, snap-frozen kidneys were homogenized on ice [50 mM phosphate buffer (pH 6.0)] and centrifuged (14,000 rpm for 20 min at 4°C) (29). The pellet was resuspended (0.5% HDTM-A), sonicated on ice (5 s each, 3 times), followed by four freeze-thaws. The supernatant was mixed with reaction buffer (50 mM phosphate buffer, 0.167 mg/ml O-oianisodine hydrochloride and 0.0005% H2O2), and absorbance was measured at 470 nm using a SPECTRAmax Plus Microplate Reader (Molecular Devices, Sunnyvale, CA). MPO activity was calculated (7) and normalized to kidney weight.
Results are expressed as means ± SE. One way ANOVA followed by post hoc analysis with the Neuman-Keuls multiple comparison test was used to determine statistical significance (P ≤ 0.05) using GraphPad Prism 6 (GraphPad Software, San Diego CA).
Pharmacokinetic Properties of Plerixafor in Rats
Plerixafor is rapidly absorbed following sc administration, with 100% bioavailability based on a comparison of pharmacokinetics following intravenous administration (data not shown). It is rapidly cleared from plasma in an apparent biphasic fashion with an elimination half-life (t½) of 0.9 and 1.16 h at 1 and 12.1 mg/kg, respectively, after a single sc injection (Fig. 1A). The plasma concentration at 1 mg/kg (single sc dose) is only detectable up to 7 h postdosing and is below the quantification limit (13 pmol/ml) at later time points. Due to the limited number of available data points beyond 7 h, the calculation of pharmacokinetic parameters is based upon the first elimination phase (0–7 h). The maximum concentration (Cmax) and area under the curve (AUC0–7) are similar following either single or repeat administration of 1 mg/kg, indicating there is no accumulation of plerixafor in plasma with repeat dosing. Dose-normalized Cmax values (Cmax/D) and AUC values (AUC0–7/D) for 1 and 12.1 mg/kg single doses indicate dose-proportional pharmacokinetics. Calculated pharmacokinetic parameters are shown in Table 1.
Plerixafor Transiently Mobilizes WBCs
In agreement with studies in mouse (8), dog (9), and human (40), plerixafor induces a transient leukocytosis in rats (Fig. 1, B–E). Total WBCs (Fig. 1B) become elevated within 1 h postinjection, peaking at ∼2–4 h. There is no marked difference in peak mobilization at doses of 1 or 5 mg/kg. However, values return to baseline sooner for the 1 mg/kg dose (7 h), remaining elevated at the higher 5 mg/kg dose. There is a lesser rise in peripheral WBCs at 0.1 mg/kg, becoming equivalent to the 1 mg/kg dose when a second injection (0.1 mg/kg) is administered 3 h later. WBC populations which are elevated include lymphocytes (Fig. 1C), neutrophils (Fig. 1D), and monocytes (Fig. 1E). Kinetics of mobilization mirror the rise in total WBCs, as previously reported (8, 40).
Plerixafor Mobilizes Hematopoietic Progenitor Cells
Leukocytosis associated with CXCR4 blockade by plerixafor has been used as an indicator of the mobilization of hematopoietic stem and progenitor cells (HPCs) (5). In the rat, HPCs are measured in the peripheral blood 3 h after sc administration. The number of mobilized HPCs is similar with treatment of plerixafor and AMD3465, a monocyclam analog of plerixafor with comparable pharmacokinetic and pharmacodynamic properties (Fig. 2A) (5). Values equivalent to plerixafor and AMD3465 are also measured when rats are treated with G-CSF (Fig. 2A). The CFU/ml is about five to seven times greater than in vehicle-treated normal animals.
Elevation in peripheral HPCs is also measured following renal ischemia-reperfusion in drug-treated animals (Fig. 2B), but not in corresponding vehicle-treated controls. Based on the blood volume of a rat (64 ml/kg) (37), the number of HPCs mobilized in a 200-g rat on a background of renal IRI is ∼1.09 ± 0.24 × 104 for plerixafor, 1.26 ± 0.48 × 104 for AMD3465, and 0.26 ± 0.1 × 104 for vehicle.
Plerixafor Ameliorates Loss of Renal Function in AKI
The effect of plerixafor on renal function in a rat bilateral model of renal IRI was evaluated next by measuring plasma creatinine and blood urea nitrogen (BUN). In this model, animals are in renal failure within 24 h postreperfusion. Various doses and regimens were explored (Fig. 3A). No improvement in plasma creatinine or BUN is measured after a single dose of 5 mg/kg plerixafor, administered 15 min before renal ischemia or at a low dose of 0.01 mg/kg given 15 min before ischemia with a second dose 3 h later. However, a significant improvement in renal function (P < 0.001) is measured with a single dose administered at 1 mg/kg, 15 min before ischemia as well as with two doses of 0.1 mg/kg given 15 min before ischemia with a second dose 3 h later (Fig. 3B). Plasma creatinine is 1.72 ± 0.10 mg/dl in the vehicle-treated group compared with 0.94 ± 0.06 and 1.11 ± 0.06 at 1 mg/kg (single dose) and 0.1 mg/kg (2 doses), respectively (P < 0.001). BUN is also significantly lower (compare vehicle at 82.62 ± 4.23 vs 49.75 ± 3.70 mg/dl at 1 mg/kg and 55.21 ± 4.30 mg/dl at two doses of 0.1 mg/kg, P < 0.001). This improvement in renal function continues to be observed at later time points as plasma creatinine and BUN return to baseline [in a separate time course study, compare plasma creatinine of vehicle vs. 1 mg/kg plerixafor at 24 h (2.11 ± 0.20 vs. 1.21 + 0.13, P < 0.001), 48 h (1.70 ± 0.31 vs. 0.79 ± 0.06, P < 0.01), and 96 h (0.55 ± 0.04 vs. 0.44 ± 0.03, P < 0.05; vehicle, n = 15; drug, n = 6–8)].
Epithelial and Endothelial Injury Are Diminished with Plerixafor
To confirm that plerixafor attenuates the loss in renal function in this model, H&E-stained slides were scored for histopathological damage of kidney tubules (Fig. 4, A and B). Ischemic injury is most apparent in proximal tubules of the S3 segment of the outer stripe of the outer medulla (16, 61). In vehicle-treated animals 24 h postreperfusion, exfoliated cells (Fig. 4A) and protein casts (Fig. 4A) congest tubule lumens to a greater extent than with plerixafor treatment (Fig. 4B). Loss of tubular structure and inflammatory infiltrates (asterisks) also is more extensive as is vascular congestion in the inner stripe (arrow). Significant differences (P < 0.005) are measured between vehicle- and plerixafor-treated groups based on semiquantitative scoring of histological injury (Fig. 4C).
TUNEL staining at 24 h postreperfusion (Fig. 4D) further reveals fewer dead cells with plerixafor treatment either at the single dose of 1 mg/kg or at two doses of 0.1 mg/kg using the dosing regimens described above (P < 0.001). Leakage of plasma proteins into the tubular lumen, a consequence of both vascular and tubular epithelial damage (22, 65), is also significantly less. Protein casts immunoreactive for IgG are 13.75 ± 1.77 profiles/field for vehicle compared with 8.33 ± 1.41 and 8.18 ± 0.526 for the 1 mg/kg (one dose) and 0.1 mg/kg (two doses) groups, respectively (P < 0.001). Similar data are measured after quantitating immunoreactive profiles of albumin (data not shown).
Markers of renal injury were attenuated 5 h postreperfusion and then were sustained at the 24-h time point. qRT-PCR analyses of kidney injury molecule-1 (Kim1) and neutrophil gelatinase-associated lipocalin (Ngal/Lcn2; Fig. 4F) are upregulated with renal IRI and nephrotoxicity (24). However, the fold-change in mRNA expression at 5 and 24 h postreperfusion (Fig. 4, E and F) is significantly lower with drug (P < 0.01).
Thus, based on renal function and mRNA expression of Kim1 and Ngal, as well as injury scoring, quantitation of cell death, and immunoreactive profiles of IgG in the tubule lumen, the data clearly indicate that kidney injury is reduced with plerixafor. This effect is observed with two dosing regimens, either a single dose of 1 mg/kg administered 15 min before ischemic clamping or with two doses of 0.1 mg/kg, the first dose given 15 min before ischemia with a second dose 3 h later (2 h postreperfusion). Based on these data, further experiments were conducted at 1 mg/kg, administered 15 min before ischemia.
Plerixafor Impacts the Immune Response
mRNA expression of proinflammatory chemokines and cytokines.
Because of the proposed role of plerixafor as a modulator of the immune response and to begin to understand the mechanism leading to improved outcome in this model, we examined the expression of proinflammatory chemokines and cytokines known to mediate renal injury. IL-6, a proinflammatory mediator of AKI which acts on neutrophils, lymphocytes, and monocytes to exacerbate renal injury, abruptly increased postreperfusion, declining thereafter (Fig. 5A). Plerixafor treatment resulted in a significant decrease (P < 0.05) in mRNA expression of Il6 at all time points postreperfusion (Fig. 5A). CXCL5 (ENA) and CXCL1 (CINC-1) induce neutrophil recruitment to inflammatory foci. qRT-PCR analyses of these chemokines revealed elevated mRNA expression following IRI, which was significantly reduced with plerixafor (Fig. 5, B and C), suggesting consequent effects on neutrophil chemotaxis and influx. A biphasic response of MCP1/Ccl2 mRNA was measured at 2 and 24 h postreperfusion, with a reduction at the 5-h time point, possibly representing influx of unique macrophage subpopulations (Fig. 5D; Clements M and Zuk A, unpublished observations). Plerixafor attenuated MCP1/Ccl2 mRNA expression at 2 and 24 h postreperfusion compared with vehicle. Il10 mRNA remained constant regardless of treatment (Fig. 5E).
Flow cytometry of kidney leukocytes.
Because tubular epithelial and microvascular endothelial cells can be induced to synthesize and secrete chemokines and cytokines (6) to facilitate renal injury, we examined by flow cytometry of kidney homogenates the number of leukocytes in the kidney following reperfusion. WBCs ubiquitously express CD45, and CD45+ leukocytes were present within the kidney shortly after reperfusion (Fig. 6A). In particular, CD11b+ cells, which represent cells of the neutrophilic and monocytic lineage, were present as expected at 5 and 24 h postreperfusion; however, the number of CD11b+ leukocytes was significantly less with plerixafor (Fig. 6B). Similar data were obtained for HIS+ cells (neutrophils and monocytes, data not shown). The amount of CD4+ lymphocytes was also lower with the drug (Fig. 6C).
Confirmatory data were generated by measuring the activity of MPO, a lysosomal enzyme of neutrophils and monocytes. MPO activity is significantly reduced with plerixafor at 5 and 24 h postreperfusion (Fig. 6D). The percentage of CD68+ monocytes/macrophages in the outer medulla is also significantly lower (data not shown).
Thus the data support that the amelioration in renal function with plerixafor administered immediately before an ischemic insult results from attenuation of proinflammatory cytokine/chemokine production and the number of WBCs appearing in the kidney postreperfusion.
AMD3465 Is Renoprotective
Parallel experiments were carried out with the monocylam CXCR4 inhibitor AMD3465 (5). A dose of 10 mg/kg administered 15 min before ischemic clamping and 2 h later ameliorated the rise in plasma creatinine and BUN (Fig. 7, A and B). Renal histology was improved (Fig. 7C), and cell death and vascular damage were attenuated (data not shown). mRNA expression of the proinflammatory markers Il6 and Cxcl1 were reduced (Fig. 7, D and E), and no change in levels of the anti-inflammatory cytokine Il10 was measured (Fig. 7F). MPO activity significantly decreased (Fig. 7G). Thus the data generated with AMD3465 are consistent with those of plerixafor supporting the observation that CXCR4 antagonism reduces the severity of kidney injury.
CXCL12 Expression Post-IRI
Because CXCL12 is the only ligand for CXCR4 (4, 49), which is expressed by leukocytes (48), HSCs, and HPCs (58), and promotes the chemotaxis of these cell types (48), we investigated the expression of CXCL12 following IRI. CXCL12/SDF-1 is detected in the distal nephron in the normal kidney (arrow, Fig. 8A), as reported by others (50, 59). CXCL12 continues to be expressed in these nephron segments 2 h postreperfusion (arrow, Fig. 8B); however, by 24 h, immunostaining is absent (Fig. 8C). Additionally, qRT-PCR of Cxcl12 mRNA reveals a significant decrease in expression within 24 h postreperfusion (Fig. 8D), confirming the immunofluorescence data.
Long-Term Effects of Plerixafor Treatment
Four weeks postreperfusion, a single dose of plerixafor (1 mg/kg) administered before ischemic injury has long-term effects on fibrosis and inflammation, consequences of renal IRI. Masson's trichrome staining highlights fewer fibrotic areas compared with vehicle treatment (Fig. 9, A–C), which is further confirmed by qRT-PCR analyses of profibrotic genes (Fig. 9D). An ∼50% decrease in mRNA expression of Col3a1, Col4a1, Fn1, and Tgfb1 is measured compared with vehicle controls. mRNA expression of cytokines and chemokines is also reduced by ∼50% (Fig. 9, F and G); these include Il6, Il1b, Tnfa, Cxcl1, Cxcl12, and MCP1/Ccl2. No change in Il10 mRNA levels is measured. The renal injury markers Kim1, Ngal, and Il18 (Fig. 9E) are also attenuated. Interestingly, plasma creatinine and BUN have returned to baseline, and no difference is measured between vehicle, drug, and normal groups (data not shown). Thus an initial reduction in kidney injury measured 24 h postreperfusion has secondary, long-term effects, dampening fibrosis and inflammation which subsequently develop.
The pharmacokinetics of plerixafor in rats was similar to that seen in dogs (9) and humans (34, 55), with rapid and complete absorption following sc administration with a tmax of 0.5 h, with subsequent rapid biphasic clearance from plasma with an initial t½ of 0.9–1.16 h. An increase in circulating WBCs was observed at 0.1, 1.0, and 5.0 mg/kg. The increase in circulating WBCs lagged behind the plerixafor tmax and Cmax, as previously observed (34). However, there was no observed dose-dependent increase in WBCs at these three concentrations, although there was a prolongation in WBC elevation at higher doses. It is therefore difficult to infer a pharmacokinetic-pharmacodynamic relationship between dose and cell mobilization as reported in humans (34). However, it is possible that in the rat a more obvious dose-response relationship may have been seen at higher doses.
Based on the pharmacokinetic profile, we evaluated multiple doses and dosing regimens. Plerixafor was administered before IRI, reminiscent of proposed treatment paradigms for prevention of human AKI resulting from cardiovascular surgery, renal transplantation, and contrast-induced nephropathy. AKI leads to high mortality and associates with progression to CKD/ESRD and thus represents a major unmet medical need. A single dose of plerixafor at 1 mg/kg administered 15 min before ischemic clamping as well as at two doses of 0.1 mg/kg administered 15 min before ischemia and 3 h later were renoprotective. Interestingly, no pharmacological effect was observed at a high dose of 5 mg/kg or at a low dose of 0.01 mg/kg. Furthermore, no effect was measured when the CXCR4 antagonist was administered only postreperfusion (data not shown), arguing the window to achieve a favorable response in this model is narrow. Thus a positive effect on outcome following IRI in the kidney is demonstrated with acute administration of plerixafor. This is similar to the reported effect of a single injection in a model of acute myocardial infarction (28).
However, plerixafor has been reported to worsen renal function following AKI when administered continuously by an Alzet minipump at 21 mg·kg−1·day−1 starting 24 h before ischemic clamping and continuing for 8 days postreperfusion (50). This is consistent with observations in a mouse model of myocardial infarction in which continuous CXCR4 blockade increased adverse cardiac remodeling (28). The discrepancy in outcome between acute and chronic dosing raises the question of the mechanism of these different pharmacological effects. The CXCR4/CXCL12 axis mediates the trafficking of WBCs, HSCs, and HPCs (1, 48). Acute inhibition of CXCR4/CXCL12 interaction mobilizes these cells as we and others have shown (48). In this setting, mobilization of stem cells has been proposed to contribute to the beneficial effect of plerixafor in disease models (11, 26, 28); however, an increase in circulating WBCs has the potential to exacerbate an inflammatory response. Conversely, blocking this interaction may also inhibit CXCL12-mediated trafficking and homing of CXCR4-expressing inflammatory cells (14, 46). Continuous administration of plerixafor, to the contrary, has been shown to block the mobilization and/or recruitment of putative reparative cells, leading to a worsened outcome in models of AKI (50) and acute myocardial infarction (28); WBC mobilization and trafficking, however, were not evaluated. Thus the pharmacological effect of CXCR4 inhibition in disease models is an intricate balance between mobilization of reparative cells and mobilization and blockade of inflammatory cell trafficking, emphasizing the complex pharmacodynamic-pharmacokinetic interaction implicit in CXCR4 inhibition. This could explain the lack of effect at 5 mg/kg in our study, where we see prolonged mobilization of WBCs, which may counteract the positive effect of the blockade of inflammatory cell trafficking to the kidney. Overlaid on this is the cross talk with other inflammatory pathways such as the CXCL12-mediated regulation of VEGF expression, which in turn can mediate other factors including matrix metalloproteinase (MMP) 9 (28). It is the modulation of these poorly understood interactions which may, in part, explain the observed pharmacological differences seen with different dosing regimens of plerixafor, acute vs. chronic, in AKI.
For our studies, we defined the optimal dose and timing of administration and thus were able to measure a beneficial effect on multiple end points. The renoprotective effect was further confirmed with a similar dosing regimen of AMD3465, a CXCR4 antagonist with comparable pharmacokinetic and pharmacodynamic properties to plerixafor. A decrease in cell death was measured with both antagonists; however, it remains to be determined which cell death pathway(s) was affected. Various cell death pathways are activated in renal IRI, including apoptosis and regulated necrosis. The latter includes mitochondrial permeability transition (42), necroptosis (36, 42, 43), and pyroptosis (63). TUNEL staining, which detects DNA fragmentation and has been used to measure apoptosis, can also be an indicator of necrotic cell death (42, 44). Thus, although plerixafor was renoprotective and decreased TUNEL-positive cells, it remains to be determined the cell death pathway(s) that was impacted in this in vivo model.
This is the first study to demonstrate that, in the rat, hematopoietic stem/progenitor cells are mobilized with plerixafor; however, our data do not suggest a stem cell-mediated effect. First, based on colony-forming unit assays, the number of mobilized cells (1.09 ± 0.24 × 104/200-g rat) was ∼250 times less than the number exogenously administered by Li et al. (38) to promote kidney repair following IRI (2.5 × 106/mouse). Similarly, the numbers of EPCs and mesenchymal stem cells required to mitigate renal IRI were 1 × 106/mouse (51) and 1.5 × 106/rat (35), respectively. Thus, even though we are comparing different cell types, the number of cells endogenously mobilized by plerixafor compared with the number of cells exogenously administered (35, 38, 51) suggests that the number of mobilized cells and/or their potency is insufficient to impact acute renal injury. Furthermore, the beneficial pharmacological effect of plerixafor seen in this model of AKI was at the lower doses (0.1 and 1 mg/kg). This together with the lack of a dose-response relationship for cell mobilization further suggests that HSC mobilization is not the mechanism whereby plerixafor protects against IRI. Recent data in fact have challenged the concept of HSCs as a therapeutic option for AKI (10).
Second, we were unable to measure an increase by qRT-PCR or immunohistochemistry of CXCL12/SDF-1 (56, 59) in kidney tubules postreperfusion, further arguing against a stem cell-mediated mechanism. The decrease in CXCL12/SDF-1 correlates with a loss of renal function, cell death, and inflammation (56), yet others have shown to the contrary that CXCL12/SDF-1 is elevated in renal tubular epithelial cells postreperfusion and leads to renoprotection by providing a chemotactic gradient for exogenously administered CXCR4-expressing bone marrow-derived cells (59). The basis for the difference in CXCL12/SDF-1 expression is unclear, but perhaps can be explained by variations in the animal models themselves and/or physiological differences in timing of tissue samples with disease onset. A recent study, however, reports that CXCL12/SDF-1 is not essential for trafficking of HSCs to the kidney (56, 57).
A more plausible explanation for the lack of an increase in renal CXCL12/SDF-1 is that CXCL12 is a labile chemokine, degraded by diaminopeptidases, MMPs, and other enzymes (13, 47). It is these properties which likely explain why CXCL12/SDF-1 has limited therapeutic utility, unless engineered to be protease resistant (53) or structurally modified to increase biological activity (60). Indeed, CXCL12/SDF-1 is rapidly degraded by proteases expressed in ischemic tissues (53). In the kidney, diaminopeptidase localizes to the apical plasma membrane of proximal tubular epithelial cells, which is shed with the apical cytoplasm into the tubular lumen (16, 61) within 1–3 h postreperfusion (64). MMP-2 and MMP-9 increase within 2–6 h postreperfusion (33), and enzymatic activity is detected in the interstitial space surrounding damaged tubules (3, 33). It is therefore likely that in the complex milieu of renal tubular and perivascular injury, where there is disruption of both the epithelial and endothelial permeability barriers, CXCL12/SDF-1 is rapidly proteolyzed, preventing accumulation in the injured areas to form a chemotactic gradient for newly mobilized CXCR4-expressing HSCs or HPCs. In fact, our immunofluorescence and mRNA expression data show a decline in CXCL12/SDF-1 2 h postreperfusion, persisting at 24 h. However, at 4 wk postreperfusion, when the kidney has recovered and fibrosis is well underway, an approximately threefold increase in CXCL12 is measured by qRT-PCR with the same probes and primers used in normal kidneys at the 24-h time point. At 4 wk, elevated CXCL12 may be chemotactic for CXCR4-expressing leukocytes or bone marrow fibrocytes (52), both of which may contribute to the inflammation and fibrosis that develop long term in this in vivo model.
Last, plerixafor has been suggested to mobilize EPCs and circulating proangiogenic CD11b+ cells (CACs) in mice (11, 28). We attempted to identify rat EPCs with plerixafor treatment but were unsuccessful (Zuk A and Vincent K, unpublished observations); markers for rat EPCs are not as well defined as for the mouse, and the low numbers measured by cell culture and flow cytometry were unreliable, below the level of sensitivity. Furthermore, plerixafor treatment resulted in a reduction of CD11b+ monocytes in the kidney, arguing against a possible role of CD11b+ CACs in this model of AKI. These observations, together with our data further suggest that mobilized HSCs or proangiogenic cells are not the mechanism whereby plerixafor prevents IRI in this in vivo model.
In the absence of a stem cell-mediated mechanism, we investigated whether the observed beneficial effect of plerixafor in this model had an immunomodulatory basis. By qRT-PCR analyses, the mRNA expression of proinflammatory chemokines Cxcl1 and Cxcl5 was ∼50% lower in kidneys of treated animals. This reduction was evident at time points early after reperfusion, suggesting that the influx of inflammatory cells known to express CXCR4 (48) was affected. Indeed, flow cytometry analyses confirmed a significantly lower renal infiltration of CD45+ leukocytes, including CD11b+ neutrophils/monocytes, within 5 h postreperfusion, persisting at the 24-h time point. The reduction in neutrophil/monocyte infiltration, key mediators of the innate immune response, was further confirmed by a decrease in MPO activity. These data agree with studies in which reducing neutrophil infiltration (2, 31) or monocyte influx (39) attenuates renal IRI (30). There was a trend toward a decrease in CD4+ T cells 5 h postreperfusion, becoming statistically significant at 24 h. Monocytes, neutrophils, and T cells are the first immune cells to infiltrate the kidney postreperfusion, and our data show that despite the elevation of WBCs in the peripheral circulation with plerixafor treatment, leukocyte influx was significantly diminished in ischemic kidneys receiving plerixafor.
In addition to this report, pharmacological benefits of plerixafor have been demonstrated in other in vivo models in which inflammation is a significant pathophysiological component. These include mouse models of collagen-induced arthritis (14, 46), allergen-induced asthma (45), and myocardial infarction (28). In all instances, there was a reduction in leukocytic infiltration into the damaged tissue; however, the question is raised as to the mechanism of this immunomodulatory property. Plerixafor uniquely and specifically binds CXCR4 and is thus able to have a direct effect on the trafficking of CXCR4-expressing inflammatory cells to the site of injury (21). It is also likely that CXCR4 inhibition has functional consequences on other leukocyte receptors. In fact, competition binding experiments with human CD4+ T lymphocytes and human monocytes illustrate that plerixafor-CXCR4 interaction negatively regulates the ability of CCR2 to bind the chemokine ligands MCP-1 and MCP-3, and CCR5 to bind MIP-1β. Similarly, CCR2 binding to MCP-1 is partially inhibited by MIP-1β, a CCR5 ligand, and SDF-1 and plerixafor, which are CXCR4-specific ligands (54). It is proposed that these interactions are mediated by receptor hetero-oligomerization; hence plerixafor, by binding CXCR4 on the leukocyte plasma membrane, could negatively regulate the functional properties of other receptors without direct binding and in this way dampen the inflammatory response characteristic of reperfusion injury. Thus the mechanism of improved outcome in this model is likely mediated by plerixafor's direct effect on reduction in immune cell trafficking to the kidney, possibly in recruitment from the periphery or inhibition of leukocyte adhesion and rolling to endothelial cells (58). Since we did not detect CXCL12/SDF-1 in the kidney following IRI, other chemokines expressed by resident cells (6) likely mediate immune cell recruitment and infiltration; however, plerixafor could diminish this response by negatively regulating receptors to other chemokines (54), impacting their ability to bind ligand.
Our data also show that dampening the immune response following IRI mitigates the progressive chronic insult which continues even after renal function has returned to baseline. Four weeks postreperfusion as fibrosis develops, mRNA expression of fibrotic genes are ∼50% lower in those kidneys which received plerixafor. mRNA expression of proinflammatory chemokines and cytokines is similarly reduced as are genes of endothelial activation [Vcam1, Vwf (von Willebrand factor), Selp (P-selectin); data not shown]. Macrophages are key effectors of renal fibrosis (18) with recent data also implicating CD4+ T cells (18). We did not analyze macrophage phenotypes in this study; however, the diminished infiltration immediately postreperfusion of monocytes/macrophages and T cells with plerixafor treatment, which is sustained at 24 h, likely explains the long-term reduction in fibrosis.
Additionally, 4 wk postreperfusion, mRNA expression of Kim1, Ngal, and Il18 are expressed at levels half that of vehicle-treated kidneys, suggesting ongoing structural damage and inflammation are reduced with plerixafor. The observation that these promising biomarkers of AKI continue to be expressed long after plasma creatinine has returned to baseline further supports the view held by some that plasma creatinine as a renal biomarker is inadequate (62) and the continued expression of KIM-1, NGAL (32), and IL-18 can prove useful in the identification of persistent renal injury in the progression of AKI to CKD.
In summary, we demonstrate in an in vivo rat model that a single injection of plerixafor mitigates the loss of renal function characteristic of AKI and improves renal histology, dampens cell death, and reduces vascular leak. Based on our data, a cell-mediated reparative effect, either by HSCs or proangiogenic cells such as EPCs or supportive CD11b+ CACs, appears unlikely. Rather, direct effects on immunomodulation of leukocyte infiltration into the kidney concomitant with a reduction in proinflammatory cytokines and chemokines appear to mediate the observed beneficial response. Long term, development of chronic injury in this model is minimized. The data presented here suggest that plerixafor may be a potential option for prevention of renal IRI. Pending evaluation of the CXCR4-CXCL12 pathway in human AKI by analysis of biomarkers relevant to this pathway, clinical indications may include renal transplantation and cardiovascular surgery.
A. Zuk, M. Gershenovitch Y. Ivanova, S. P. Fricker, and S. Ledbetter are employees of Genzyme. R. T. MacFarland is an employee of Pfizer, Inc.
Author contributions: A.Z., M.G., Y.I., R.T.M., S.P.F., and S.R.L. provided conception and design of research; A.Z., M.G., Y.I., and R.T.M. performed experiments; A.Z., M.G., Y.I., R.T.M., S.P.F., and S.R.L. analyzed data; A.Z., M.G., Y.I., R.T.M., S.P.F., and S.R.L. interpreted results of experiments; A.Z., M.G., R.T.M., and S.P.F. prepared figures; A.Z., M.G., Y.I., and S.P.F. drafted manuscript; A.Z. and S.P.F. edited and revised manuscript; A.Z., M.G., Y.I., R.T.M., S.P.F., and S.R.L. approved final version of manuscript.
Present address of R. T. MacFarland: Pfizer, Inc., 10646 Science Center Dr., San Diego CA 92121.
We thank Nibedita Chattopadhyay and Juanita Campos-Rivera for technical assistance, Meghan Clements, Christopher Chaber, and Jianhua Huang for help with the in vivo surgical studies, Melanie Ruzek for thought-provoking discussions, and Yves Sabbagh for help with preparation of figures.
- Copyright © 2014 the American Physiological Society