Urine podoplanin heralds the onset of ischemia-reperfusion injury of the kidney

Abstract

Ischemia-reperfusion injury represents one of the most common causes of acute kidney injury, a serious and often deadly condition that affects up to 20% of all hospitalized patients in the United States. However, the current standard assay used universally for the diagnosis of acute kidney injury, serum creatinine, does not detect renal damage early in its course. Serendipitously, we found that the immunofluorescent signal of the constitutive podocyte marker podoplanin fades in the glomerulus and intensifies in the tubulointerstitial compartment of the kidney shortly after ischemia-reperfusion injury in 8- to 10-wk-old male C57Bl/6j mice. Therefore, we sought to define the appearance and course of the podoplanin-positive signal in the kidney after ischemia-reperfusion injury. The tubulointerstitial podoplanin-positive signal increased as early as 2 h but persisted for 7 days after ischemia-reperfusion injury. In addition, the strength of this tubulointerstitial signal was directly proportional to the severity of ischemia, and its location shifted from the tubules to interstitial cells over time. Finally, we detected podoplanin in the urine of mice after ischemia, and we observed that an increase in the urine podoplanin-to-creatinine ratio correlated strongly with the onset of renal ischemia-reperfusion injury. Our findings indicate that the measurement of urine podoplanin harbors promising potential for use as a novel biomarker for the early detection of ischemia-reperfusion injury of the kidney.

INTRODUCTION

Acute kidney injury (AKI) is a common but costly condition that affects nearly 20% of hospitalized patients in the United States, and it imposes a significant burden on the financial resources and infrastructure of the American health care system by prolonging hospital admissions (8, 26, 28). Although AKI strains the resources of hospitals universally, timely diagnosis of this condition remains elusive, since the use of serum creatinine (Cr), the most widely used biomarker for AKI in the clinical setting, does not permit its early detection (5). Therefore, the inadequacy of serum Cr has led to an exigent need for the identification of a biomarker for AKI that diagnoses the condition reliably and immediately after its onset.

Ischemia-reperfusion injury (IRI) refers to the organ tissue damage mediated by free radicals and reactive oxygen species that occurs after periods of disrupted blood flow, and IRI of the kidney accounts for almost 50% of the cases of AKI (6). Prior research has focused on the impact of IRI on the tubular compartment of the kidney, since patients are known to suffer acute tubular necrosis (ATN) after periods of ischemia. However, despite widespread awareness among nephrologists that tubuloglomerular feedback after ischemic AKI results in vasoconstriction of the afferent arteriole and a subsequent decrease in glomerular blood flow, the acute impact of IRI on the glomerulus remains relatively unexplored (19). Although one group found that mild IRI in rats resulted in albuminuria, no change in the morphology of the glomerular endothelial cells or podocytes was observed (2).

Podocytes express the membrane glycoprotein podoplanin (PDPN), mainly along their urinary surfaces (7). However, few precise functional roles for PDPN, also known as gp38, Aggrus, and T1α, have been elucidated (4).

We found unexpectedly that IRI of the kidney in a mouse model resulted in a decrease in the signal of PDPN in the glomerulus. Therefore, we sought to determine whether the expression of PDPN by the podocytes simply diminished or whether PDPN was released in the urine, a finding that could serve as a springboard for future investigations of PDPN as a novel biomarker of IRI. In addition, we found incidentally that the PDPN⁺ signal increased in the tubulointerstitial compartment of the kidney at later time points after IRI, its intensity increased with the severity of ischemia, and the distribution of its expression changed over time. Therefore, we characterized the source of the PDPN⁺ signal further to clarify the evolution of its distribution in the kidney after IRI, and we evaluated the association between the level of PDPN in the urine and onset of IRI.

MATERIALS AND METHODS

Animals.

Male C57Bl/6J mice (8–10 wk old, Jackson Laboratory, Bar Harbor, ME) were used in all experiments. All mice resided in groups of five or less in pathogen-free cages in the animal care facility at Brigham and Women’s Hospital. They received irradiated food and water ad libitum. At the designated euthanasia time points, mice were euthanized by intraperitoneal injection of ketamine-xylazine, monitored for the loss of vital signs, and underwent cervical dislocation. Urine was collected, and blood was drawn at the time of euthanasia. All animal experiments and methods adhered to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of Brigham and Women’s Hospital (protocol no. 2016N000167).

Antibodies.

The following antibodies were used for immunofluorescence staining: goat anti-PDPN (1:200, catalog no. AF3244, R&D Systems), Syrian hamster anti-PDPN (1:200, catalog no. 127401, BioLegend), goat anti-nephrin (1:100, catalog no. AF3159, R&D Systems), rat anti-podocalyxin (1:100, catalog no. MAB1556-SP, R&D Systems), rabbit anti-neural/glial antigen 2 (NG2; 1:100, catalog no. AB5320, EMD Millipore), fluorescein-labeled Lotus tetragonolobus lectin (LTL; 1:200, catalog no. FL-1321, Vector Laboratories), rabbit anti-cleaved caspase-3 (1:200, catalog no. 9661S, Cell Signaling), rabbit anti-platelet-derived growth factor receptor-α (PDGFR-α; 1:200, catalog no. ab96806, Abcam), rabbit anti-desmin (1:200, catalog no. ab32362, Abcam), mouse anti-α-smooth muscle actin (α-SMA; 1:200, catalog no. A2547, Sigma), and rabbit anti-fibronectin (1:200, catalog no. ab2413, Abcam).

IRI of the kidney.

IRI was performed in mice as previously described (16). Briefly, 8- to 10-wk-old male C57Bl/6j mice were anesthetized by intraperitoneal injection of ketamine (80–100 mg/kg) and xylazine (5–10 mg/kg). Flank incisions were made, through which the kidneys were exposed. Nontraumatic microaneurysm clamps (Roboz Surgical Instruments) were used to clamp both renal pedicles and induce bilateral ischemia for the designated duration. Body temperature was maintained at 36.5–37.3°C. Reperfusion of the kidneys was verified through visual observation after removal of the clamps. The above steps were repeated, except for the clamping of the renal pedicle, in mice randomized to the sham group included for the biomarker analysis experiments. Each group in all experiments consisted of five mice.

Cisplatin injury.

Cisplatin (20 mg/kg, Sigma-Aldrich) was administered intraperitoneally to four male C57Bl/6j mice. After the injection (4 days), mice were euthanized and their kidneys were harvested and sectioned.

Immunofluorescence and immunohistochemical staining.

Immunofluorescence staining of the kidneys was performed on frozen sections obtained at the time of euthanasia, and immunohistochemical staining was performed on paraffin sections with a routine protocol using VECTASTAIN ABC Kits (HRP) (Vector Laboratories) and 3,3-diaminobenzidine (DAB) peroxidase as the colorizing substrate. Frozen sections were fixed with cold acetone for 1 min and washed with PBS for 10 min. Next, sections were incubated with blocking solution (3% BSA in PBS + 0.1% saponin for intracellular staining, 3% BSA in PBS for others) for 30 min. The primary antibody diluted in blocking solution was then applied to the sections for 1 h. Sections were washed three times in PBS for 5 min each. Secondary antibody diluted in PBS was applied for 30 min, and sections were then washed again three times with PBS. Sections were mounted using a drop of mounting medium with DAPI (Vectashield). Fluorescent sections were visualized using the microscope Evos FL Auto 2 by Invitrogen (ThermoFisher Scientific), and all images assessed in each comparison of experimental and control specimens were captured using the same exposure time and gain. Light micrographs of sections stained by immunohistochemistry and hematoxylin and eosin (H&E) were captured by the AmScope light microscope. Mean fluorescence intensity of PDPN in the glomeruli and tubules was determined by comparing the signals of each channel using ImageJ software.

Western blot analysis.

The protein concentrations of lysates from the kidney medulla of three naïve mice and three mice that underwent IRI of 26 min and euthanasia 24 h later were measured using the Bradford assay. Equal amounts of protein were separated by SDS-PAGE and transferred to a polyvinylidene difluoride (PVDF) membrane. GAPDH was used as the loading control. Membranes were immunoblotted with the following specific antibodies: anti-PDPN (Novus), anti-GAPDH (Santa Cruz Biotechnology), anti-goat HRP (Abcam), and anti-mouse HRP (Jackson ImmunoResearch) using standard protocols. Blots were developed with West Dura Chemiluminescent Substrate using a Bio-Rad ChemiDoc Imaging System. Band density analysis was performed using ImageJ software.

Cell culture.

HK-2 human kidney proximal tubule cells (CRL-2190, adult male donor, American Type Culture Collection) were cultured in DMEM (Lonza) supplemented with 10% FBS and incubated at 37°C and 5% CO₂. Confluent monolayers (80%) of HK-2 cells grown in six-well plates were treated with or without 500 μmol/l H₂O₂ (Fisher Scientific) diluted in serum-free media. After exposure (30 min), cells were washed three times with PBS and harvested for further analysis.

Quantitative PCR.

RNA was isolated from HK-2 cells with TRIzol (ThermoFisher Scientific), and first-strand cDNA was synthesized using 2 μg RNA and High-Capacity Reverse Transcriptase (ThermoFisher Scientific). RT-PCR was performed using SYBR Green PCR reagents (Bio-Rad, Hercules, CA). RNA levels were normalized to the level of GAPDH and calculated as the ΔΔ threshold cycle. Primers used for RT-PCR were as follows: GAPDH, forward 5′-GGATTTGGTCGTATTGGG-3′ and reverse 5′-GGAAGATGGTGATGGGATT-3′; and PDPN, forward 5′-CGAAGATGATGTGGTGACTC-3′ and reverse 5′-CGATGCGAATGCCTGTTAC-3′. All quantitative PCRs were performed in triplicate.

ELISA.

Urine samples were collected directly via pipette from the urinary meatus at the time of euthanasia. Next, urine samples were centrifuged for 10 min at 1,000 g. The supernatant was collected and diluted 1:4 in PBS. Urine PDPN was measured by sandwich ELISA as per the manufacturer’s protocol (LifeSpan BioSciences).

Cr measurement.

Urine Cr was measured by an enzymatic assay kit (Crystal Chem), as per the manufacturer’s protocol.

Statistical analysis.

Two-sided, unpaired Student’s t-tests were used for all comparisons in this study, except for the analysis of biomarker utility, and were performed using Prism software (GraphPad). Differences were considered statistically significant when P < 0.05.

To test the utility of the urine PDPN-to-Cr ratio as a biomarker, parallel boxplots for the PDPN-to-Cr ratio in urine collected from mice 2 h after 30 min of IRI and those that underwent sham surgery (n = 5 mice/group) were created to indicate possible violation of the normality assumption. Statistical analysis was performed using the Wilcoxon two-sample rank-sum test.

RESULTS

The PDPN⁺ signal increases in the tubules after AKI.

Examination of the kidneys of mice at 24 h after IRI with an ischemia time of 26 min incidentally revealed the presence of PDPN in the tubular compartment of the kidney, including the proximal tubule, distal tubule, and medullary tubules. Close inspection of the proximal tubule epithelium, identified by the constitutive marker megalin, indicated that PDPN was found inside the cytoplasm of these cells, but it was not located prominently inside the lumen of the tubule (Fig. 1A). To investigate whether the increased PDPN⁺ signal was because of gene expression by proximal tubule cells, we exposed HK-2 human proximal tubule cells in vitro to H₂O₂ and performed quantitative PCR for PDPN expression. However, no expression of PDPN was observed by these cells, neither in the presence nor absence of H₂O₂ (data not shown).

Fig. 1.

The podoplanin (PDPN)⁺ signal increases in the tubules after acute kidney injury (AKI). A: immunofluorescence micrograph of the kidney cortex after ischemia-reperfusion injury (IRI) with an ischemia time of 26 min revealed a PDPN⁺ signal (red) in the megalin⁺ proximal (green) and megalin⁻ distal tubules. The close-up images (bottom) indicate that PDPN can be found inside the cytoplasm of the proximal tubular (PT) epithelium, whereas it outlined the lumen of the distal tubule (DT). DAPI (blue) was used as a nuclear marker. Scale bar = 50 μm. B: immunofluorescence micrograph of the kidney cortex after IRI demonstrating a population of PDPN⁺ interstitial cells (red, marked by white asterisk) that is distinct from the PDPN⁺ tubular epithelium, as indicated in the close-up image (bottom). DAPI (blue) was used as a nuclear marker. Scale bar = 50 μm. C: immunofluorescence micrograph of the kidney medulla indicating the PDPN⁺ signal (red) in the medullary tubules. DAPI (blue) was used as a nuclear marker. Scale bar = 50 μm. D: immunofluorescence micrograph of transverse section of kidney medulla displaying the clear separation between PDPN⁺ tubules (red) and podocalyxin⁺ blood vessels (green). DAPI (blue) was used as a nuclear marker. Scale bar = 25 μm. E: light micrograph of kidney tissue harvested from mouse 4 days after cisplatin administration contained PDPN⁺ material (brown) in the lumens of the tubules. Scale bar = 50 μm.

In contrast to the pattern observed by the proximal tubule cells in the kidney, PDPN outlined the contour formed by the luminal surface of the distal tubular epithelium, which indicated that it was located primarily inside the lumen of the distal tubule and not inside the cytoplasm of these cells (Fig. 1A). In addition, PDPN was expressed by a population of cells within the interstitium, alongside the signal observed within the tubular epithelium (Fig. 1B). PDPN was also found in the lumen of the medullary tubules, as indicated by immunofluorescence staining of both longitudinal and transverse sections of the medulla (Fig. 1, C and D). To investigate whether the PDPN⁺ signal can also be found in other mouse models of AKI, we stained the kidneys of mice that underwent cisplatin-induced injury, and we observed that PDPN was present inside material within the tubules of these mice 4 days after cisplatin administration (Fig. 1E).

The PDPN⁺ signal fades in the glomerulus but increases in the tubular compartment gradually after IRI.

Mouse kidneys were stained for the podocyte markers PDPN and nephrin, the vascular endothelial cell marker podocalyxin, and the mesangial cell marker NG2 to examine changes in the signals of these proteins in the glomerulus over 24 h after IRI with an ischemia time of 26 min (Fig. 2). The PDPN⁺ signal diminished gradually at 2, 6, and 24 h after IRI compared with a naïve kidney (Fig. 2A, row 1). However, no significant change in the fluorescent signals of nephrin, podocalyxin (a protein expressed both by podocytes and endothelial cells, especially after glomerular damage), or NG2 was noted (Fig. 2A, rows 2–4). Caspase-2 was not found to be present in the glomerulus, indicating that the podocytes did not undergo apoptosis despite the loss of the PDPN⁺ signal (Fig. 2A, row 5). The mean fluorescence intensity of the PDPN⁺ signals was quantified and indicated a gradual loss of fluorescence through 24 h after IRI (Fig. 2B).

Fig. 2.

The podoplanin (PDPN)⁺ signal decreases in the glomerulus but increases in the tubular compartment gradually after ischemia-reperfusion injury (IRI). A: immunofluorescence micrographs of the glomerulus at different time points after IRI with an ischemia time of 26 min displaying the decrease in the PDPN⁺ signal (red, row 1), whereas the signals for the podocyte marker nephrin (green, row 2), vascular endothelial cell marker podocalyxin (green, row 3), and mesangial cell marker neural/glial antigen 2 (NG2; green, row 4) remained strong. The absence of caspase-2 (row 5) at all time points indicates the lack of cellular apoptosis. Scale bar = 50 μm. B: the mean fluorescence intensity (MFI) of podoplanin in the glomeruli decreased significantly and progressively through 24 h after IRI (n = 10 glomeruli from a total of 3 mice/group). C: immunofluorescence micrographs of kidney revealing the strengthening in the PDPN⁺ signal (red) in the tubular compartment of the kidney at different time points through 24 h [medulla (m) and cortex (c)] after IRI. DAPI (blue) was used as a nuclear marker. Scale bar = 200 μm. D: the MFI of PDPN in the tubular compartment increased significantly through 24 h after IRI (n = 3–5 microscopic fields from a total of 3 mice/group). E: immunofluorescence micrographs of the kidney medulla using a different anti-PDPN antibody (Syrian hamster IgG, clone 8.1.1, BioLegend) demonstrating the similarly increased signal at 24 h post-IRI compared with medullas of naïve mice. DAPI (blue) was used as a nuclear marker. Scale bar = 200 μm. F: Western blot of the kidney medulla demonstrating the increased band density of PDPN from kidney medulla lysates harvested from mice at 24 h post-IRI compared with medullas from naïve mice (P = 0.0916 as per a two-sided unpaired Student’s t-test). GAPDH was used as a loading control. G: light micrographs of the hematoxylin and eosin (H&E)-stained kidney cortex at 2 h (top) and 6 h (middle) after IRI displaying sloughing of the tubular epithelium, whereas light micrograph at 24 h after IRI (bottom) revealed severe tubular injury, including tubular cast formation (asterisk). Scale bar = 75 μm. *P < 0.05 and ***P < 0.005 per a two-sided unpaired Student’s t-test.

The PDPN⁺ signal in the tubular compartment of the kidney persisted at 2, 6, and 24 h after ischemia, especially in the medulla, as indicated by mean fluorescence intensity (Fig. 2, C and D). We verified that the antibody that we used to detect PDPN in these kidneys (goat anti-PDPN, R&D Systems) was specific for PDPN by confirming that a second antibody against PDPN (Syrian hamster anti-PDPN, clone 8.1.1, BioLegend) also demonstrated a similar increase in fluorescent signal in the medulla at 24 h post-IRI compared with naïve mice (Fig. 2E). In addition, Western blot analysis performed with lysates from the medullas of kidneys harvested from mice 24 h post-IRI revealed an increase in PDPN compared with medullas of naïve mice (Fig. 2F).

H&E stains of the kidney cortex performed at each time point after IRI indicated increasing signs of renal injury, including sloughing of the tubular epithelium in the tubular lumen at 2 and 6 h, followed by tubular cast formation at 24 h (Fig. 2G).

Prolongation of ischemia time correlates with increase in PDPN⁺ staining, and the glomerular PDPN⁺ signal returns by 7 days after IRI.

Next, we sought to determine whether increasing the severity of IRI by prolonging the ischemia time correlates with a higher PDPN⁺ signal in the kidney (Fig. 3). Indeed, incremental increases of the PDPN⁺ signal in the tubules of the cortex and medulla were observed after 15, 20, and 26 min of ischemia, as quantified by mean fluorescence intensity (Fig. 3, A and B).

Fig. 3.

Prolongation of ischemia time correlates with the increase in podoplanin (PDPN)⁺ staining, and the glomerular PDPN⁺ signal returns by 7 days after ischemia-reperfusion injury (IRI). A: immunofluorescence micrographs of kidney demonstrating the positive correlation between the strength of the PDPN⁺ signal (red) and severity of IRI. B: the mean fluorescence intensity (MFI) of the PDPN⁺ signal in tubules of the cortex and medulla increased significantly with the worsening severity of IRI (n = 5 microscopic fields from a total of 3 mice/group). C: the MFI of the PDPN⁺ signal was not significantly different in the glomeruli of naïve kidneys and kidneys 7 days after IRI with an ischemia time of 26 min (n = 10 glomeruli from a total of 3 mice/group). D: immunofluorescence micrographs of glomeruli 7 days after IRI showing the return of a strong PDPN⁺ signal (red, top left) and persistence of signals for the podocyte marker nephrin (green, top right), vascular endothelial cell marker podocalyxin (green, bottom left), and mesangial cell marker neural/glial antigen 2 (NG2; green, bottom right). DAPI (blue) was used as a nuclear marker. Scale bar = 50 μm. E: immunofluorescence micrograph of a kidney 7 days after IRI showing the very intense PDPN⁺ signal (red) in the medulla (m) compared with the cortex (c). DAPI (blue) was used as a nuclear marker. Scale bar = 200 μm. F: the MFI of the PDPN⁺ signal was significantly higher in the medulla of kidneys 7 days after IRI compared with naïve kidneys (n = 5 microscopic fields from a total of 3 mice/group). G: immunofluorescence micrographs of kidney tissue indicating that PDPN (red) does not costain with stromal cell markers such as platelet-derived growth factor receptor-α (PDGFR-α; green, 1st image of the top row), NG2 (green, 2nd image of the top row), fibronectin (FN; green, 3rd image of the top row), desmin (green, 4th image of the top row), and α-smooth muscle actin (α-SMA; green, 5th image of the top row) in the naïve kidney. PDPN did partially costain with PDGFR-α (green, 1st image of the bottom row), NG2 (green, 2nd image of the bottom row), FN (green, 3rd image of the bottom row), and desmin (green, 4th image of the bottom row), but it surrounded and did not costain with α-SMA (green, 5th image of the bottom row). Scale bar = 25 μm. **P < 0.01 and ***P < 0.005 per a two-sided unpaired Student’s t-test.

Next, we examined the presence of PDPN in the kidney 7 days after an ischemia period of 26 min (day 7). The strength of the PDPN⁺ signal in the glomerulus at this time point resembled the naïve kidney, thereby indicating that this change is reversible (Fig. 3C). The nephrin, podocalyxin, and NG2 signals also remained strong (Fig. 3D). In contrast to the reversal of changes in the glomerulus, the PDPN⁺ signal increased markedly in the medulla of the kidney at day 7, and this signal was found increasingly in the interstitium, although it also remained strong in the tubules (Fig. 3E). The mean fluorescence intensity of the medullary PDPN⁺ signal increased significantly in the day 7 kidneys compared with naïve kidneys (Fig. 3F).

Next, the cellular signal of PDPN in the interstitium was examined more closely, and costaining with several common markers of stromal cells was performed to uncover the identity of these PDPN-expressing cells (Fig. 3G). The few PDPN-expressing cells in the interstitium of the naïve kidney were not found to costain with any of these markers, including PDGFR-α, NG2, fibronectin (FN), desmin, and α-SMA (Fig. 3G, top row). However, in the day 7 kidney, some costaining was observed between PDPN and PDGFR-α, NG2, FN, and desmin (Fig. 3G, bottom row). Nonetheless, the majority of PDPN⁺ cells did not costain with these markers, and, peculiarly, PDPN⁺ cells appeared to form a ring around α-SMA⁺ fibroblasts.

Although H&E staining of the kidney cortex 7 days after IRI indicated persistent sloughing of the tubular epithelium, it revealed no significant sign of injury in the glomerulus, and the tubular casts found in the cortex at the 24-h time point were no longer observed (data not shown).

PDPN can be detected in the urine immediately after IRI and increases with the severity of ischemia.

Next, we measured the levels of PDPN in the urine of mice at 2 h after bilateral IRI of varying severity to test its efficacy as a biomarker of early kidney damage (Fig. 4). ELISA was performed to measure PDPN and Cr in the urine of mice that received 10 and 30 min of ischemia, and these mice were compared with a control group that received sham surgery (Fig. 4A). The urine PDPN-to-Cr ratio was calculated to normalize the excretion of PDPN for the glomerular filtration rate. The urine PDPN-to-Cr ratios of the mice were higher after 30 min of ischemia compared with mice that received 10 min of ischemia or sham surgery (Fig. 4A).

Fig. 4.

Podoplanin (PDPN) can be detected in the urine immediately after ischemia-reperfusion injury (IRI) and increases with severity of ischemia. A: the urine PDPN-to-creatinine (Cr) ratio as measured by ELISA 2 h after IRI increased without significance after 10 min of IRI, but it increased significantly after 30 min of IRI compared with mice that underwent sham surgery (n = 5 mice/group). B: parallel boxplots of urine PDPN-to-Cr ratios from mice that underwent sham surgery and those that received 30 min of ischemia demonstrating the significantly higher median normalized PDPN in the ischemic group compared with the sham surgery group (n = 5 mice/group, P = 0.003968 per a Wilcoxon rank-sum test). **P < 0.01 per a two-sided unpaired Student’s t-test.

Figure 4B shows the parallel boxplots of the PDPN-to-Cr ratio of the mice in the sham surgery group and the group that received 30 min of ischemia (n = 5 mice/group). The boxplots indicate possible violation of the normality assumption. Hence, we applied a Wilcoxon rank-sum test to challenge the null hypothesis that the median normalized PDPN of the ischemic group was greater than that of the sham group. The resulting P value of 0.003968 led to rejection of the null hypothesis.

H&E staining of kidney after different durations of ischemia revealed expectedly signs of increased injury to the tubules, such as cast formation, after 30 min of ischemia compared with 10 min (data not shown).

DISCUSSION

PDPN is a transmembrane glycoprotein that binds to the ezrin/radixin/moesin (ERM) family of intracellular proteins that is responsible for the cross-linking of actin with the cell membrane, so its activity can produce rearrangements in the actin cytoskeleton (17, 18). However, despite the publication of several studies focusing on its activity, the function of PDPN in podocytes remains mysterious (10, 12, 14, 15, 23, 27, 29). In particular, no previous study has investigated the relationship between IRI and the expression of PDPN in the kidney.

In the present study, we identified two separate and distinct patterns illustrated by the PDPN⁺ signal in the kidney after IRI. The first pattern, the PDPN⁺ signal in the tubules, occurs immediately after IRI, and its appearance correlates with the decline in the glomerular PDPN⁺ signal. This temporal relationship raises the possibility that these two phenomena are linked. PDPN has been previously identified in exosomes and microvesicles shed in vitro by PDPN-transfected Madin-Darby canine kidney cells, PDPN-transfected human melanoma SK-MEL-28 cells, and human HN5 squamous cell carcinoma cells (9). The expression of PDPN was also implicated in the promotion of extracellular vesicle formation (9). Podocytes express PDPN constitutively, and, therefore, they may shed PDPN constantly in microvesicles. Because 90% of the PDPN in the podocyte is expressed on the cell membrane that faces the urinary surface (7), the podocyte may encapsulate some of these proteins in extracellular vesicles and expel them in the urine. This concept could explain the detection of PDPN in the urine within 2 h of IRI. The lack of expression of PDPN by proximal tubule cells that we verified both under normal conditions and ischemic conditions in vitro substantiates the possibility that PDPN is internalized by these cells from the urine. Notably, in vitro experiments conducted to replicate ischemic conditions do not reflect completely the in vivo scenario of renal IRI, and measuring gene expression of proximal tubular epithelial cells microdissected from the kidney after IRI may provide a more accurate representation. Moreover, the effect of IRI on the possible internalization of PDPN by the proximal tubule epithelium remains unstudied and provides the basis for an exciting future topic of research.

The second pattern that characterizes the PDPN⁺ signal in the kidney after IRI is exemplified by the expression of PDPN by spindle-shaped cells in the interstitium of the medulla. The morphology of these cells resembles fibroblasts, although their molecular signature appears unique. The majority of these PDPN⁺ cells do not express the typical fibroblast marker PDGFR-α, nor do they express α-SMA, a marker of activated fibroblasts. In addition, they do not express the pericyte marker NG2. These PDPN⁺ cells may arise from resident fibroblasts in the kidney or they may migrate to the kidney from another organ. Others have demonstrated that the expression of PDPN is associated with epithelial-to-mesenchymal transition, and it correlates with a phenotypic change associated with increased cellular mobility and invasiveness in distant tissues, such as lymph nodes (17, 24, 25). This property could suggest a distant origin for these cells, such as the bone marrow, which is the site of origination for progenitor cells during processes in which neovascularization plays an important role (1, 3, 21, 22).

Fibroblastic reticular cells are PDPN⁺ mesenchymal cells that reside in the stromal compartment of the lymph node, and they mold the identity of the inflammatory milieu within this secondary lymphoid organ (11, 20). Our laboratory has found that, during the early phase of a classic mouse model of glomerulonephritis, PDPN⁺ fibroblastic reticular cells undergo a transformation into a proinflammatory phenotype, since they form tracks that surround and support the growing lymphatic and blood vasculature in the expanding lymph node (13). The perivascular location of PDPN⁺ cells in the interstitial compartment of the kidney after IRI resembles the pattern that these fibroblastic reticular cells adopt during activation of the immune response in glomerulonephritis, a similarity that suggests perhaps that these cells support the growing vasculature in the kidney during the repair phase after IRI. Whether these PDPN⁺ cells in the kidney originate in the lymph node remains unexplored and represents an important subject for future investigation.

One limitation of our study is the sole use of male mice, which tend to be more susceptible to IRI than female mice. We chose to study a single sex in our experiments to ensure consistency across the various time points and degrees of ischemia that we assessed. Future investigations of the utility of PDPN as an early signal of IRI should include studies of female mice to ensure that an increase in the PDPN⁺ signal is present, regardless of the sex of the animal. However, the appearance of PDPN within material found in the tubules of mouse kidneys after cisplatin-induced injury suggests that this phenomenon may also be relevant to other experimental models of AKI.

IRI of the kidney places a substantial burden on health care systems worldwide, but the current ubiquitous use of serum Cr does not provide a sensitive method for its prompt diagnosis. Although the glomerulus does not appear to undergo a structural change after IRI, its expression of PDPN decreases as this protein is released in the urine. In this study, we have demonstrated the potential utility of urine PDPN for the early detection of IRI in mice. Translation to human studies through detection of PDPN in the urine of patients with renal IRI represents an important next step that can be pursued through high-sensitivity proteomic analysis by methods such as multidimensional chromatography, lectin purification, or data-independent acquisition mass spectrometry. Therefore, measurement of the urine PDPN-to-Cr ratio represents a promising possible solution to the current important unmet clinical need for the expedient identification of renal IRI.

GRANTS

This work was supported by a research grant from Dialysis Clinic (to V. Kasinath) and by National Institutes of Health Grants T32-DK-007527 (to V. Kasinath) and R01-AI-126596 (to R. Abdi).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

V.K., O.A.Y., and R.A. conceived and designed research; V.K., O.A.Y., M.U., M.Y., L.J., X.L., and T.I. performed experiments; V.K., O.A.Y., W.Q., S.E., and R.A. analyzed data; V.K., O.A.Y., and R.A. interpreted results of experiments; V.K., O.A.Y., W.Q., and SE prepared figures; V.K., O.A.Y., and W.Q. drafted manuscript; V.K., O.A.Y., and R.A. edited and revised manuscript; V.K., O.A.Y., M.U., M.Y., X.L., W.Q., S.E., T.I., and R.A. approved final version of manuscript.