Abstract
Background Ischemic AKI lacks a urinary marker for early diagnosis and an effective therapy. Differential nucleophosmin (NPM) phosphorylation is a potential early marker of ischemic renal cell injury and a therapeutic target.
Methods Differential NPM phosphorylation was assessed by mass spectrometry in NPM harvested from murine and human primary renal epithelial cells, fresh kidney tissue, and urine before and after ischemic injury. The biologic behavior and toxicity of NPM was assessed using phospho-NPM mutant proteins that either mimic stress-induced or normal NPM phosphorylation. Peptides designed to interfere with NPM function were used to explore NPM as a therapeutic target.
Results Within hours of stress, virtually identical phosphorylation changes were detected at distinct serine/threonine sites in NPM harvested from primary renal cells, tissue, and urine. A phosphomimic NPM protein that replicated phosphorylation under stress localized to the cytosol, formed monomers that interacted with Bax, a cell death protein, coaccumulated with Bax in isolated mitochondria, and significantly increased cell death after stress; wild-type NPM or a phosphomimic NPM with a normal phosphorylation configuration did not. Three renal targeted peptides designed to interfere with NPM at distinct functional sites significantly protected against cell death, and a single dose of one peptide administered several hours after ischemia that would be lethal in untreated mice significantly reduced AKI severity and improved survival.
Conclusions These findings establish phosphorylated NPM as a potential early marker of ischemic AKI that links early diagnosis with effective therapeutic interventions.
Ischemia/reperfusion injury is a leading cause of human AKI.1–3 AKI occurs in 7%–10% of hospitalized patients at average risk for renal injury4 and between 23% and 74% of high-risk patients.5,6 Even a modest increase in serum creatinine of >0.5 mg/dl increases the length of stay and hospital cost as well as morbidity and mortality.4 In fact, mortality has been reported to be as high as 80% when the serum creatinine acutely rises by >2 mg/dl.4 In children and adults, a single AKI episode risks progressive kidney injury resulting in ESRD.7,8 Less than 60% of children survive >3–5 years after a single AKI episode.9 Surviving children have a 60% chance of developing high BP, proteinuria, and/or CKD that ultimately requires dialysis or kidney transplantation.9 This sobering reality is partly due to the facts that current AKI diagnostic tests of organ dysfunction are insensitive and that ischemic renal injury lacks effective treatment.10
Tissue ischemia/reperfusion causes proximal tubule epithelial cell (PTEC) injury, a major contributor to organ failure.1,3,11 Experimental evidence implicates Bax, a quintessential BCL2 proapoptotic protein, as an important cause of regulated PTEC death during hypoxic or ischemic stress and contributor to AKI.11–15 We have recently discovered that nucleophosmin (NPM), a highly conserved, ubiquitously expressed nucleolar protein essential for mammalian cell survival,16–18 also facilitates PTEC death in a Bax-dependent manner.13 In normal cells, NPM acts as a protein chaperone that shuttles between the nucleus and cytosol to promote protein synthesis, ribosomal biogenesis, and cell proliferation.13,17,19,20 During ischemic stress, however, NPM rapidly enters the cytosol and complexes with conformationally activated Bax, and together, NPM and Bax cause mitochondrial injury and cell death.13 NPM exists in two forms: large multimers restricted to the nucleus and monomers capable of entering the cytosol.21 In general, oligomeric NPM promotes cell proliferation, whereas monomeric NPM enhances death21 after proapoptotic insults that activate Bax in both cancer22,23 and noncancer13 cells. The post-translational modifications that convert NPM from its essential role in normal cell housekeeping to a cytotoxin during stress are completely unknown. In this study, a phosphoproteomics approach is used to identify the post-translational phosphorylation changes that convert NPM from an essential component of normal cell machinery to a toxic Bax chaperone during ischemic AKI in murine and human cells, fresh kidney tissue, and urine.
We tested the hypothesis that phosphorylated NPM is both a marker of acute renal cell injury and an effective target for peptide therapeutics designed to ameliorate NPM-Bax–mediated tissue injury. We discover that virtually identical stress-induced, site-specific phosphorylation changes (i.e., phosphorylation and dephosphorylation) render NPM toxic in isolated murine and human renal cells as well as fresh kidney tissue and urine. Furthermore, NPM-Bax–mediated renal cell death and AKI may be remediable to therapeutic intervention. Our results suggest that diagnostics on the basis of differential NPM phosphorylation can be biologically linked to effective AKI therapeutics, even after the insult has occurred.
Methods
Animals
All animals were maintained under the guidance and policies of the Boston University Animal Core facility using National Institutes of Health (NIH) and Institutional Animal Care and Use Committee guidelines. Male and female C57BL/6J mice were purchased from Jackson Laboratory (Bar Harbor, ME), and they were housed and bred at the Boston University Animal Core facility under license number AN-15024. Four-week-old mice with mature kidneys afford more consistent results with regard to AKI severity and have superior survival after anesthesia. Animas were handled in a manner consistent with the NIH guide for their care and use under Institutional Animal Care and Use Committee approval.
Cell Culture
Primary culture of human donor kidney and murine kidney proximal tubular cells was performed as previously described.13,24 Briefly, renal cortices were harvested. Individual cortical samples were minced and digested by collagenase IV (1 mg/ml); then, they were plated onto culture dishes in a renal epithelial cell–compatible medium (ATCC PCS-400–030). Cultures of cells that exhibited proximal tubule characteristics were maintained for 5–7 days in 5% CO2 at 37°C.13,24,25 HEK293NT lentiviral packaging cells (System Biosciences, Mountain View, CA) were maintained in DMEM with high-glucose medium (Life Technologies BRL, Carlsbad, CA) containing 10% bovine serum and 10% penicillin-streptomycin at 37°C in a 5% CO2 incubator.
Ischemic Stress
In Vitro
Exposure to metabolic inhibitors is an established model of ischemic stress that reduces ATP content to <10% of baseline values until recovery is initiated.24,26 To achieve ATP depletion, cells were washed three times in glucose-free DMEM (Invitrogen, Carlsbad, CA) followed by incubation in glucose-free DMEM containing sodium cyanide (5 mM) and 2-deoxy-glucose (5 mM) for the indicated times.26 Recovery was initiated by replacing the stress medium described above with complete primary cell culture medium containing glucose and substrates that support the rapid restoration of cell ATP.27
In Vivo
Mice were subjected to bilateral renal pedicle clamping after being anesthetized with 2,2,2-tribromoethanol. A midline laparotomy was performed, and nontraumatic vascular clamps were placed on both renal pedicles for 25–30 minutes, an insult that produces severe AKI.13,28,29 Slight variation in the clamp time was required to achieve similar degrees of renal dysfunction in the relatively large number of animals studied. The clamp was then removed, and reperfusion was confirmed by visual inspection of the kidneys. Sham ischemia was performed by encircling each renal pedicle with a nonocclusive ligature. The first postischemic urine was collected by gentle compression of the inferior abdominal wall.
Hypoxic Stress
To simulate insults encountered during clinical AKI, PTECs were maintained under hypoxic conditions for 70 minutes as previously described, and this is sufficient to induce cell death.15 Briefly, cells were washed with PBS, transferred to a sealed anaerobic chamber filled with 95% nitrogen and 5% CO2 and incubated in Krebs–Ringer bicarbonate buffer (catalog no. k4002; Sigma-Aldrich). This buffer was pre-equilibrated with 95% nitrogen and 5% CO2.
Human Kidneys
Samples of renal cortex were harvested for primary cell culture, and assessment of cytosolic NPM content from paired, nontransplantable human kidneys in transportable chambers perfused with chilled University of Wisconsin machine preservation solution at 4°C was performed.30 In both kidney pairs, one kidney was normally perfused, whereas the other experienced perfusion failure and was, therefore, subjected to a prolonged period of warm ischemia.
Mapping of NPM Phosphorylation Sites by Mass Spectrometry
NPM was purified from human and murine primary renal cells and fresh kidney tissue homogenates as well as control and postischemic urine by immunoprecipitation using an established method.13 Enzyme digestion with glutamic acid and/or trypsin was performed to obtain fragments appropriate for subsequent analysis. Tandem mass spectrometry spectra obtained by fragmenting a peptide by collision-induced dissociation were acquired using a capillary liquid chromatography/tandem mass spectrometry system that consisted of a Surveyor HPLC pump, a Surveyor Micro AS autosampler, and an LTQ linear ion trap mass spectrometer. Detection and mapping of phosphorylation sites were achieved by database searching of tandem mass spectra of proteolytic peptides with specified phosphate modification (P=+80 D) on serine, threonine, and tyrosine residues against the current mouse and human protein databases. A detailed liquid chromatography/tandem mass spectrometry approach for phosphoproteomics has been previously described.31–33 Mass spectrometry was performed in primary human and renal cells at baseline and after simulated ischemic stress, in fresh human and murine kidney tissue homogenates before and after frank ischemia, and in human and murine urine from patients without AKI and patients with AKI before and after experimental AKI, respectively. At least three replicates of each purified NPM sample harvested from renal cells, kidney tissue, and fresh urine were subjected to mass spectrometry.
NPM Phosphomutant Construction and Infection
Human NPM cDNA (catalog no. 34553; Addgene, Cambridge, MA) and pCDH lentivector systems (System Biosciences, Mountain View, CA) were used to generate NPM wild-type, “NPM stress mimic” (T86E-S88E-T95E-T234A-S242A), and “NPM normal mimic” (T86A-S88A-T95A-T234E-S242E) lentiviral expression plasmids by PCR-based mutagenesis with a flag tag fused to the 5′ open reading frame. PCR products were cloned into pCDH vector BamHI/EcoRI sites. All constructs were confirmed by DNA sequencing before cotransfection with pPACKG1 lentivector packaging plasmids into HEK293NT cells to generate lentipseudoviral particles. The resultant pseudoviruses were harvested and purified, and titers were determined according to the manufacturer’s instructions. Cells were exposed to virus (MOI: 5–8) diluted in Opti-MEM (Invitrogen). After 6 hours incubation with virus-infected medium, a complete culture medium was substituted for an additional 16-hour incubation period.
NPM Knockdown
Clustered regularly interspaced short palindromic repeats (CRISPR) interference was used to transiently suppress NPM. After 60 hours under routine cell culture conditions, primary cells were transfected with two sgRNA targeting adjacent to the NPM transcription start site (synthesized by Synthgo, Menlo Park, CA) using TransIT-2020 Transfection Reagent using the manufacturer’s protocol. Six hours later, five MOI of dcas9-KRAB lentivirus (catalog no. k204; abm.com, Richmond, BC, Canada) were added. The medium was then changed, and culture continued for 48 hours. The magnitude of NPM suppression was assessed by immunoblot analysis.
Dot Blot Analyses
Because relatively low NPM concentrations in urine are insufficient for routine immunoblot due to volume limitations of sample loading, a dot blot assay standard method and protocol were used (http://www.abcam.com/ps/pdf/protocols/dot%20blot%20protocol.pdf). In the dot blot assay, NPM can be easily detected, because a much larger urine volume can be loaded. Importantly, identical anti-NPM antibodies were used in both the immunoblot and dot blot assays.
Renal-Targeted NPM Peptides
For in vitro testing in primary human PTECs, peptides (200 μM) were tested that replicate the Bax domain responsible for binding NPM,34 or two peptides designed to interfere with NPM function were compared with a random nonspecific sequence of the same length (control). Peptides intended to interfere with NPM function overlap key phosphorylation sites located at the NPM amino terminus (peptide 2) or carboxy terminus (peptide 3) detected by mass spectrometry. To improve cell uptake, all peptides contained a cell-penetrating sequence13. For in vivo experiments, a renal-targeting sequence (NLS) was added to enhance renal peptide uptake.35 Peptides were commercially synthesized (Biomatik, Wilmington, DE) as follows: NPM-Bax blocking peptide: Ac-KKKRKV-(βA)-TVTIFVAGVLTASLTIWKKMG-COOH; peptide 2: AC-PKKKRKV-(βA)-TLKMSVQPTVSLGGFEITPPVVLRLK-COOH; and peptide 3: AC-PKKKRKV-(βA)-ESFKKQEKTPKTPKGPSSVED-IKAK-COOH. A single dose (100 μg/g body wt of the NPM-Bax blocking or control peptide) was administered by tail vein injection immediately after release of the pedicle clamps (0 time point) as well as 1, 2, 3, 4, 5, and 6 hours later in each experimental group. A total of 42 control and 56 experimental animals were studied.
Cell Viability or Kidney Function
Transient exposure to metabolic inhibitors replicates ischemia in vivo by activating Bax and causing cell death.13,29,36–38 Cell viability was assayed using a modified colorimetric technique (MTT assay).37 The number of surviving cells is expressed as a percentage of viable control cells detected at baseline. Serum BUN and creatinine levels were measured in tail vein blood samples for 7 days after ischemia using QuantiChrom BUN or creatinine assay kits (BioAssay Systems, Hayward, CA) according to the manufacturer’s instructions.
Protein Isolation and Immunoprecipitation
Cell, tissue, and urinary proteins were harvested using PD buffer containing 40 mM Tris-HCl, pH 8.0, 500 mM NaCl, 0.1% Nonidet P-40, 6 mM EGTA, and a protease inhibitor cocktail (Set I; Calbiochem, San Diego, CA). NPM was immunoprecipitated overnight with NPM antibodies (catalog no. B0556; Sigma-Aldrich) in the presence of protein A/G–agarose beads at 4°C. NPM accumulation was measured in cytosolic samples extracted with low-dose digitonin (8 μg/ml).13 Cytosolic proteins were extracted from renal cortical tissue samples using an extraction kit (Biochain Institute, Newark, CA). Briefly, 50–100 mg of kidney cortex tissue was minced on ice and spin washed with 1 ml ice-cold wash buffer before adding 100–200 μl of buffer C. The mixture was rotated at 4°C for 20 minutes and then centrifuged at 18,000×g at 4°C for 20 minutes before being harvested. Protein levels were measured with the bicinchoninic acid assay (Thermo Scientific, Rockford, IL).
Immunohistochemistry
NPM was visualized as previously reported,13 and the protocol was performed according to the manufacturer’s protocol (catalog no. B0556).
Dot Blot Assay
NPM was detected in murine urine using a dot blot assay as previously reported39 using an antibody directed against total NPM.
Statistical Analyses
Data are expressed as means with SEM. Differences between groups were determined by a two-tailed t test using Excel (Microsoft, Redmond, WA). Comparisons involving more than two groups were determined by a two-way ANOVA followed by a Holm–Sidak post hoc test for nonparametric data. P<0.05 was considered significant.
Results
Stress Causes NPM Translocation in Both Cells and Tissues
NPM translocation from the nucleus to the cytosol promotes cell death.40,41 To determine the extent to which ischemic stress causes NPM translocation, an early step in the NPM-mediated death pathway,42 PTECs were subjected to transient ATP depletion, an established model of metabolic stress that resembles ischemia.24,26 In resting primary human PTECs, >95% of total cell NPM localized to the nuclear (i.e., noncytosolic) fraction. In contrast, about 15% of total cell NPM accumulated in the cytosol after stress. Compared with normal cells, only about 80% of NPM remained intracellular (15% cytosolic plus 65% nuclear) (Figure 1A), suggesting that extracellular NPM leakage or NPM degradation occurred. Stress caused marked cytosolic NPM translocation in ATP depleted primary murine PTECs (Figure 1B) or renal homogenates harvested from grossly ischemic, nontransplantable human kidneys ex vivo compared with well perfused kidneys harvested from the same donor (Figure 1C). Cytosolic NPM translocation was also detected in murine and human primary PTECs after transient hypoxia (Figure 1D). Densitometric analysis confirmed the magnitude and reproducibility of NPM translocation. Both de-energization and hypoxia contribute to renal epithelial cell injury and organ failure during human AKI.3
Stress causes cytosolic nucleophosmin (NPM) translocation. (A) Quantitative assessment of NPM in the nuclear and cytosolic cell fractions of primary human proximal tubule epithelial cells (PTECs) subjected to 60 minutes ATP depletion; identical amounts of proteins were loaded in each lane, and densitometric analysis was used to assess the relative amount of NPM present in the cytosolic and noncytosolic fractions before (Base) and after ATP depletion (ATP Depl; n=3). NPM accumulation detected in the cytosolic fraction of (B) renal cortical homogenates harvested from paired donor kidneys with either normal perfusion (normal) or perfusion pump failure (ischemic; results represent four kidneys from two human donors), (C) primary murine or human PTECs subjected to ATP depletion ×60 minutes, and (D) murine and human PTECs subjected to 70 minutes of hypoxia. In B–D, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) serves as a loading control; densitometry represents three independent experiments in each study. Cytosolic fractions were harvested using digitonin (see Methods); noncytosolic extracts were harvested by exposing cells to RIPA buffer after extracting the cytosolic fraction. CTL, control; RDU, relative density unit.
Stress Induces Site-Specific NPM Phosphorylation Changes
To identify the post-translational events that regulate NPM toxicity, the effect of metabolic stress on differential NPM phosphorylation was examined by mass spectrometry in purified NPM harvested from human and murine kidney cortex and human and murine primary PTECs after ischemic stress as well as after murine and human AKI. Representative tracings are shown (Figure 2, A–C). Digestion of NPM from cells and tissue with trypsin and glutamic acid afforded positive identification of 269 of 292 (92%) potential NPM residues and yielded 100% all known serine, tyrosine, and threonine residues capable of undergoing ischemia-induced phosphorylation or dephosphorylation. Despite a two-amino acid difference in overall length, the amino acid sequences are >94% similar in murine and human NPM, and the phosphorylation consensus sequences are identical. Compared with control renal tissue and cells, five distinct phosphorylation changes at T86, S88, T95, T234, and S242 were observed after ischemic stress (Figure 2D). Remarkably, only a single event (dephosphorylation at T86 in murine kidney tissue) differed from site-specific phosphorylation changes detected in other cell or tissue samples subjected to ischemic stress. The pattern of urinary NPM phosphorylation identically matched phosphorylation at four of five serine/threonine residues detected in ischemic cells and renal tissue (Figure 2D). Only the peptide fragment containing S240 was not detected in NPM purified from postischemic murine urine. In a patient with Kidney Disease Improving Global Outcomes criteria for stage 3 AKI,43 differential phosphorylation at residues T86, S88, and S242 was detected (Figure 2D). The pattern of NPM phosphorylation in human urine during AKI resembled NPM phosphorylation changes in ischemic cells, tissue, and murine urine after experimental AKI. Total NPM was readily detected in murine urine within 6 hours after renal ischemia and disappeared within 48 hours after injury (Figure 3).
Mass spectrometry reveals consistent differential phosphorylation changes in NPM harvested from in renal cells, kidney tissue, and urine. Representative tracings used to perform differential phosphoproteomic analysis of purified NPM harvested from the renal cortex of an ischemic human kidney are shown. (A) Liquid chromatography/mass spectrometry base peak peptide profile of an NPM amino acid fragment. (B and C) Liquid chromatography/tandem mass spectrometry (MS/MS) peptide collision-induced dissociation fragment ions providing de novo peptide sequence and specific serine phosphorylation sites represented by loss of −167 D (S-87 + P-80 D). (D) Summary of five differentially NPM-phosphorylated sites altered in a virtually identical manner by ischemic stress in primary murine cells or human proximal tubule epithelial cells (PTECs), murine kidney (mKdy), and human kidney (hKdy) versus control (CTL; murine PTEC, mKdy, and hKdy controls were identical and are shown together). Murine AKI urine (mAKI), human AKI urine (hAKI), and human non-AKI urine (hNonAKI) are also shown. Phosphorylation absent is indicated by a cross; phosphorylation present is indicated by a star. Asterisks show homology between human NPM and murine NPM at residues T234/T232 and S242/S240, respectively. The kinase consensus sequences for each of the five phosphorylation NPM sites are identical between these two species. N/D, not detected.
Renal ischemia increases urinary nucleophosmin (NPM). Marked accumulation of total NPM in murine urine was detected by dot blot analysis 6 and 12 hours after 25 minutes of transient bilateral renal ischemia, but it disappeared within 48 hours postischemia. In contrast, no urinary NPM was detected at the same time points in sham-operated animals.
NPM Phosphomimic Proteins Mediate NPM Toxicity
To determine the biologic role of differential NPM phosphorylation at these five serine/threonine sites and the contribution to cytotoxicity, two flag-tagged NPM phosphomutant proteins as well as wild-type NPM were generated and introduced into human proximal tubule cells. These two mimic proteins replicated the phosphorylation state of NPM in either its normal or stressed configuration at the five differentially phosphorylated sites identified by mass spectrometry (Figure 4). Specific bioassays were then used to assess the effect of these five post-translational phosphorylation changes on toxic NPM behaviors, including translocation, deoligomerization, NPM-Bax complex formation, mitochondrial NPM and Bax accumulation, and cell death. Both the wild-type and normal NPM phosphomimic proteins accumulated in a nucleolar pattern within resting cells (Figure 5A, left and center panels). In contrast (and in the absence of ischemic stress), only the NPM stress mimic protein localized to the cytosol (Figure 5A, right panel). In resting cells, wild-type and normal mimic NPM formed large oligomers. In contrast, the stress mimic NPM protein accumulated as small monomers known to interact with conformationally active Bax (Figure 5B). After ATP depletion sufficient to activate Bax, NPM-Bax complex formation was substantially greater in cells that contained stress mimic NPM (Figure 5C). Similarly, the accumulation of both stress mimic NPM and active Bax accumulation in isolated mitochondria after ischemic stress exceeded that observed with wild-type or normal mimic NPM (Figure 5D). Compared with control or cells that expressed either empty vector (an added transfection control) or wild-type NPM, postischemia survival was significantly lower in cells containing stress mimic NPM (Figure 5E) (P<0.05). Unexpectedly, normal mimic NPM expression significantly reduced cell death (P<0.05). Compared with wild-type NPM, the expression of normal mimic NPM also decreased NPM-Bax complex formation (Figure 5C) and the accumulation of NPM and Bax in isolated mitochondria after ischemic stress (Figure 5D).
Differentially phosphorylated nucleophosmin (NPM) sites localize near its structural and functional domains. Location of NPM phosphosites detected in postischemic proximal tubule epithelial cells (PTECs), kidney tissue, and urine. Therapeutic peptide 2 (representing NPM amino acids 78–103) and 3 (representing NPM amino acids 226–246) were designed to interfere with NPM functions at the indicated phospho-sites differentially phosphorylated during ischemic stress. Site-specific amino acid substitutions were made to render each site either constitutively phosphorylated (E substituted for S or T) or constitutively dephosphorylated (A substituted for S or T) to replicate the phosphorylation status of wild-type NPM under resting or ischemic stress conditions (as described in Figure 2D).
Phosphomimic nucleophosmin (NPM) proteins that replicate stress-induced phosphorylation changes regulate NPM toxicity. The biologic behavior of nucleophosmin (NPM) phosphomimic proteins assessed using functional bioassays performed in primary murine proximal tubule epithelial cells (PTECs) that express flag-tagged wild type (WT), a normal (Nor-M or NrNPM) NPM mimic protein, or a stress (Stre-M, StNPM, or Stress-M) NPM mimic protein that replicates differential NPM phosphorylation. (A) NPM localization in intact renal cells in the absence of stress. (B) NPM oligomers and monomers detected in lysates harvested from nonstressed cells in a native gel; total NPM content is unchanged in each lane, and the increase in monomeric NPM is accompanied by a reciprocal decrease in the oligomeric form of NPM. (C) NPM-Bax complex formation assessed after immunoprecipitation (IP) of flag-tagged NPM followed by immunoblot (IB) of conformationally activated Bax (Bax 6A7) after ischemic stress; input protein amounts were similar in each sample. (D) Accumulation of NPM and 6A7 Bax in isolated PTEC mitochondria after ischemic stress; VDAC, a mitochondrial membrane protein, serves as loading control. (E) Cell survival after 60 minutes ATP depletion (n=6). EV, empty sgOpti vector; RDU, relative density unit. *P<0.05 for Nor-M and Stress-M NPM phosphomimics versus control (CTL).
NPM Is a Relevant Target for Ameliorating Renal Cell Injury
To assess the causal role of NPM in renal cell death, primary human PTECs were subjected to ischemic stress before and after NPM knockout using an inducible CRISPR system. CRISPR reduced NPM expression by 80%–90% (Figure 6, inset) and significantly increased cell survival after ischemia (Figure 6).
Nucleophosmin (NPM) suppression increases proximal tubule epithelial cell (PTEC) survival after ischemic stress. An inducible clustered regularly interspaced short palindromic repeats-based system suppressed NPM expression after doxycycline exposure (Cri; 2 mg/ml 372 hours; inset) versus empty sgOpti vector (EV; inset). NPM suppression significantly improved human PTEC survival after 60 minutes of ATP depletion. CRISPRi, clustered regularly interspaced short palindromic repeats interference. *P<0.05 EV versus Cri (n=4).
Cell and Animal Survival after Ischemic Stress Is Improved by Peptides Designed to Interfere with NPM Function
Identification of the key events that regulate NPM cytotoxicity raised the possibility that peptides designed to interfere with stress-induced NPM phosphorylation or NPM-Bax complex formation might have therapeutic potential. To test this hypothesis, peptides that replicate the phosphorylation consensus sequences altered by ischemic stress (Figure 2D) at the NPM amino (amino acids 78–103) or carboxy termini (amino acids 226–246) or likely to interfere with NPM-Bax complex formation34 were generated and tested. After introduction into human PTECs, all three peptides significantly improved postischemic cell survival (Figure 7).
Therapeutic peptides improve renal cell survival after ischemic stress. Peptides designed to reduce nucleophosmin (NPM)-Bax complex formation (Bax peptide) or interfere with phosphorylation changes that regulate NPM toxicity (peptides 2 and 3) significantly improve cell survival after ischemic stress (n=4; upper panel). Peptide 2 replicates regulatory phosphorylation sites located at the amino terminus (T86, S88, and T95). Peptide 3 replicates regulatory phosphorylation sites located at the carboxy terminus (T234 and S242). The amino acid sequences of each of the three peptides designed to interfere with NPM function are shown (lower panel). *P<0.05 versus control peptide (CTL).
To assess its efficacy in treating severe ischemic AKI, a single peptide dose was intravenously administered to mice 0–3 hours after releasing bilateral renal pedicle clamps after 28 minutes of ischemia, an insult sufficient to induce lethal organ failure. All control mice that received a random peptide (six of six) died within 72 hours after pedicle clamp release (Figure 8A). Before death, these animals had BUN concentrations that averaged 120 mg/dl, a marked degree of azotemia. In contrast, animals receiving a single dose of a renal targeted peptide designed to interfere with Bax-NPM complex formation exhibited significantly lower BUN levels and less severe organ failure on days 1–7 postischemia. Peptide treatment also reduced serum creatinine by nearly 50% on day 1 postischemia, a time point associated with marked BUN accumulation (Figure 8B). Remarkably, seven of eight peptide-treated animals (88%) survived severe ischemic AKI. Peptide administration at 0, 1, or 2 hours postischemia also provided significant protection against lethal renal ischemia (P<0.05 versus control), but no protection was observed if the same peptide dose was administered 4, 5, or 6 hours after transient renal ischemia (P>0.05) (Table 1). In peptide-treated animals, recovery of renal function was complete, and it was sustained for at least 4 weeks. Blocking peptide treatment markedly reduced Bax-NPM interaction in renal epithelial cells in vitro (Figure 9A) as well as postischemic renal homogenates (Figure 9B).
Therapeutic peptide treats ischemic AKI. (A) After 28 minutes of bilateral renal pedicle clamping, a single dose of blocking peptide was intravenously administered to each of seven paired animal groups at 0, 1, 2, 3, 4, 5, or 6 hours (n=8 animals for each group; total of 56 mice) or control (n=6 animals for each group; total of 42 mice). Blocking peptide administered within 3 hours postischemia significantly improved renal function (i.e., lowered serum BUN) on days 1–7 and dramatically improved animal survival (P<0.05 versus control). The magnitude of renoprotection was similar with peptide given at 0, 1, , and 3 hours postischemia (P<0.05 for 0 versus 1 or 3 hours); only data for blocking and control peptides administered 3 hours after ischemia are shown. (B) Serum creatinine on day 1 (the time of peak kidney dysfunction) was significantly reduced when the blocking peptide was administered 3 hours after ischemia; *P<0.05 versus control (CTL).
Renoprotective effect of peptide therapy for AKI
Blocking peptide decreases nucleophosmin (NPM)-Bax interaction. Active Bax immunoprecipitates harvested from (A) human proximal tubule epithelial cell (PTEC) lysates after 60 minutes of ATP depletion plus 30 minutes of recovery or (B) renal cortical homogenates after ischemia in the absence of a peptide (None), a control peptide (CTL), or a blocking peptide (Block) injected 3 hours after bilateral renal ischemia. Similar amounts of Bax were detected in each lane, and input protein amounts for each sample were identical. IB, immunoblot; IP, immunoprecipitation.
Discussion
This manuscript describes the discovery of the post-translational modifications that convert NPM from a highly conserved protein essential to cell survival to a cytotoxin and shows that these post-translational phosphorylation changes link early AKI diagnostics with effective therapeutics. Mass spectrometry of purified NPM reveals five distinct serine/threonine sites that regulate NPM toxicity during ischemic stress in both primary renal cells and intact tissues of two genetically divergent mammals, showing conservation of this mediator in the stress-induced cell death pathway. Expression of an NPM phosphomimic protein that replicates the differential phosphorylation changes detected in primary renal cells, tissue, and urine after an ischemic insult reproduces NPM translocation and deoligomerization, key behaviors of the wild-type NPM observed during ischemic stress. Interestingly, two of five phosphosites identified in our study (S88 and T95) also regulate NPM deoligomerization and cytosolic translocation in human cancer cells, respectively,40,41 and they determine responsivity to therapy in patients with acute myelogenous leukemia by enhancing their sensitivity to chemo- and radiation-induced apoptosis.44,45 Semiquantitative analysis of peptide abundance by mass spectrometry revealed that >80% of cytosolic NPM was differentially phosphorylated in the toxic pattern after stress. This suggests that differential phosphorylation positively correlates with translocation to the cytosol. In addition to S88 and T95, this study also discovers three other stress-induced differential phosphorylation changes that regulate NPM toxicity in ischemic renal cells (outlined in Figure 10). The proposed intracellular events are on the basis of the observation that our NPM phosphomimic protein (with all five stress-induced phosphorylation changes) undergoes (1) cytosolic NPM translocation (Figures 1 and 5A), (2) deoligomerization of large NPM pentamers to monomers capable of traversing the nuclear pore46 (Figure 5B), (3) complex formation with conformationally activated Bax (Figure 5C), and (4) coaccumulation with Bax in isolated mitochondria (Figure 5D) and cell death (Figure 5E).
Five phosphorylation changes mediate NPM toxicity during ischemic stress. Based on the results of our NPM bioassays (Figure 5), we propose that phosphorylation changes at S88 and T9540,41 regulate steps 1–3 in the nucleophosmin-Bax cell death pathway. T95 phosphorylation also regulates cell death in response to chemotherapy and radiation therapy in acute myelogenous leukemia.40,41 Release from nucleolar binding sites is assumed to occur before NPM deoligomerization or cytosolic translocation.45,75 The function of NPM-T86, T234, and S242 phosphorylation changes during ischemic cell death is presently unknown.
Alterations in NPM phosphorylation alone, even in the absence of cell stress, are sufficient to cause cytosolic NPM accumulation and deoligomerization. In contrast, NPM-Bax complex formation and mitochondrial NPM-Bax accumulation require both site-specific NPM phosphorylation changes and stress-induced conformational Bax activation. Our laboratory recently showed that mitochondrial Bax accumulation is NPM dependent13 and that the NPM-Bax complex forms only after Bax undergoes conformational activation,13,34,47 exposing the amino-terminal 6A7 epitope.48,49 Although increasing cytosolic NPM significantly promotes Bax-mediated cell death, increasing cytosolic NPM per se is nontoxic.13 These observations support the hypothesis that both NPM and Bax are both required to cause stress-induced cell death, and they are analogous to Knudson “two-hit hypothesis” of cancer cell transformation.50 In a biologic twist of fate, Bax has also been implicated in releasing nuclear NPM into the cytosol,51 further linking their toxic behaviors. Effective reagents in our experimental models are likely to be therapeutic for Bax-mediated renal injury in humans, because virtually identical post-translational NPM modifications at identical consensus sequences were detected in ischemic murine and human cells and tissue and were partially replicated in urine during clinical and experimental AKI.
Several laboratories have shown that Bax-regulating kinases, including protein kinase B (Akt13,52), glycogen synthase kinase 3β13,53, and Jun N-terminal kinase,54 activate Bax. In renal cells, conformational Bax activation is primarily due to the loss of Akt-mediated Bax serine184 phosphorylation during ischemic stress.13 Surprisingly, the normal NPM phosphomimic protein significantly improved cell survival after stress. This is likely due to the fact the normal phosphomimic protein inhibited both stress-induced NPM-Bax complex formation (Figure 5C) and NPM-Bax accumulation in isolated mitochondria (Figure 5D) compared with either the wild-type or stress mimic NPM. These observations suggest that the normal NPM phosphomimic mutant protein acts as a competitive inhibitor of endogenous NPM and may itself be a therapeutic agent.
These data also suggest that differential NPM phosphorylation is a potential urinary marker for renal epithelial cell injury as well as an effective target for peptide-based therapeutics directed against NPM. Within hours of renal ischemia (and at the time of first urine collection), both total and phosphorylated urinary NPM are readily detectable. This is more likely due to NPM leakage across the cell membrane than NPM degradation during the brief period of ischemia.13 Mass spectrometry reveals that urinary NPM exhibits the same pattern of phosphorylation and dephosphorylation as NPM harvested from renal cells and fresh issue. The observation that both total and phosphorylated NPM are detectable within hours after renal ischemia, well before a rise in serum BUN/creatinine is evident, supports their further evaluation as early AKI markers that can be used to initiate effective AKI treatment. Despite marked organ dysfunction, total NPM is no longer detected in urine 48 hours after the insult, suggesting that it reflects acute renal epithelial cell injury. Although total urinary NPM could result from detached, sublethally injured renal cells, we suspect that phosphorylated NPM is specific for lethal cell injury. This is strongly supported by our observations that wild-type NPM overexpression in renal cells is nontoxic,13 whereas stress phosphomimic NPM exacerbates ischemic cell death. The magnitude of differential NPM phosphorylation during ischemic stress is uncertain and will likely require quantitative mass spectrometry to be adequately addressed. In biology, it is unusual for protein to exist in a single form, because phosphatase and kinase activity act as opposing forces to regulate protein function.55 On the basis of compelling data gleaned from our phosphoproteomic study and peptide treatment, we propose that NPM is a potential marker of acute renal cell injury that can be physiologically linked to effective AKI treatment.
NPM has been proposed by others to be a Bax chaperone,34,47 and it is strongly implicated in regulating Bax-mediated ischemic renal cell injury.13 However, this study shows for the first time that a blocking peptide designed to interfere with NPM-Bax interaction actually decreases NPM-Bax complex formation in vitro and in vivo by about 50% (Figure 9), improves renal cell survival (Figure 7), and substantially improves kidney function (i.e., reduced both BUN and creatinine) by 50%. Although a single effective inhibitor is unlikely to afford complete protection, a cocktail of site-specific peptides directed at multiple steps in the NPM-Bax cell death pathway could improve renoprotection. The recent observation that NPM phosphorylation has interdependent regulatory effects on oligomer assembly and protein partner binding20 (e.g., with Bax) supports the use of multiple reagents. We acknowledge that NPM- and Bax-independent forms of cell death likely contribute to organ failure after ischemia.56 However, substantial evidence suggests that Bax contributes to the forms of regulated cell death most often detected after renal ischemia, including apoptosis, necrosis,57–59 and more recently, necroptosis, a biochemically distinct form of cell death.57,60
Peptides are ideal for treating renal diseases that primarily target the PTEC. Peptides have few untoward side effects partly due to their extremely short serum t1/2.61 Furthermore, peptides that contain the NLS sequence selectively accumulate in the kidney. In fact, >94% of an intravenous radiolabeled peptide attached to the NLS sequence accumulates in the murine kidney.35 It is highly likely the megalin receptor, located on the brush border of the PTEC, mediates renal peptide uptake and markedly prolongs its t1/2.62 In this study, NLS-containing peptides uniquely designed to interfere with separate NPM functions significantly protected cells against ischemic stress (Figure 7) as effectively as CRISPR-mediated NPM knockdown (Figure 6). The seemingly modest 20%–30% improvement in cell survival could promote organ recovery by more rapidly replacing cells lost during ischemia, a key determinant of organ recovery after AKI.28,63 In contrast to universally lethal AKI in control, 75%–88% of animals survived after a single dose of the peptide administered 0–3 hours after releasing the renal pedicle clamps, showing for the first time its efficacy in treating ischemic AKI, a disease the presently lacks an effective treatment. Although peptide administration did not completely protect against AKI, treatment was sufficient to promote animal survival, equivalent to avoiding RRT in human AKI. The 3-hour therapeutic window for peptide treatment is similar to that reported for thrombolytic treatment in acute myocardial infarction64 and ischemic stroke,65 insults in which Bax-mediated apoptosis and necrosis have been reported.58,66–68 Although relatively narrow, this treatment window exceeds that reported for other forms of programmed cell death in renal ischemia-reperfusion injury, including necroptosis, that can only be given before the insult.69
The remarkable efficacy of interfering with a single-cell death pathway is likely due to the facts that (1) Bax-mediated apoptosis and necrosis represent a continuum of cell death rather than distinct pathways13,70,71,72; (2) cytosolic NPM translocation, deoligomerization, and NPM-Bax complex formation occur before cells are irreversibly committed to die13,34 (unlike blocking downstream caspases); and (3) renal cell death is partly responsible for the subsequent oxidant injury and inflammation that accompany organ failure.72 As a result, early NPM detection linked with early treatment is especially attractive.
Although elegant studies have dissected the biochemical and structural requirements for outer membrane permeabilization by Bax,66,73 none have identified the signals or proteins responsible for mitochondrial Bax accumulation during stress. In response to stress, approximately 75%–80% of cell Bax colocalizes with mitochondria without a known mitochondrial localizing sequence74 or an identified outer membrane receptor,73 suggesting that Bax requires a cytosolic chaperone to target mitochondria. Our results show that virtually identical phosphorylation changes at specific amino acid residues convert NPM from an essential protein to a toxic Bax chaperone in both mice and humans and are a potential early marker of ischemic renal cell injury and an attractive target for therapeutic intervention. The presence of both NPM and Bax in all mammalian cells raises the intriguing possibility that this regulatory cascade also participates in tissue injury in other ischemia-susceptible organs.
Disclosures
None.
Acknowledgments
This work was supported by National Institutes of Health grant DK-053387 (to S.C.B.) and a Boston University Center for Translational Sciences Institute grant (to S.C.B.).
Footnotes
Published online ahead of print. Publication date available at www.jasn.org.
- Copyright © 2019 by the American Society of Nephrology
References
In this issue
