Kim-1 Targeted Extracellular Vesicles: A New Therapeutic Platform for RNAi to Treat AKI

Significance Statement

AKI is a frequent clinical problem without definitive therapies. We developed an efficient RNAi therapy against AKI by engineering red blood cell-derived extracellular vesicles (REVs) with targeting peptides and therapeutic siRNAs. REVs targeted with Kim-1–binding peptide LTH efficiently delivered P65 and Snai1 siRNAs to the injured tubules, leading to reduced expression of P-p65 and Snai1. Dual suppression of P65 and Snai1 inhibited renal inflammation and fibrosis in mice subjected to ischemia/reperfusion injury and unilateral ureteral obstruction, and blunted the chronic progression of ischemic AKI. This study provides an efficient platform, REV_LTH, for the targeted delivery of therapeutics into injured tubular cells, and suggests the viability of targeting P65 and Snai1 as a therapeutic avenue for AKI.

Visual Abstract

Abstract

Background

AKI is a significant public health problem with high morbidity and mortality. Unfortunately, no definitive treatment is available for AKI. RNA interference (RNAi) provides a new and potent method for gene therapy to tackle this issue.

Methods

We engineered red blood cell–derived extracellular vesicles (REVs) with targeting peptides and therapeutic siRNAs to treat experimental AKI in a mouse model after renal ischemia/reperfusion (I/R) injury and unilateral ureteral obstruction (UUO). Phage display identified peptides that bind to the kidney injury molecule-1 (Kim-1). RNA-sequencing (RNA-seq) characterized the transcriptome of ischemic kidney to explore potential therapeutic targets.

Results

REVs targeted with Kim-1–binding LTH peptide (REV_LTH) efficiently homed to and accumulated at the injured tubules in kidney after I/R injury. We identified transcription factors P65 and Snai1 that drive inflammation and fibrosis as potential therapeutic targets. Taking advantage of the established REV_LTH, siRNAs targeting P65 and Snai1 were efficiently delivered to ischemic kidney and consequently blocked the expression of P-p65 and Snai1 in tubules. Moreover, dual suppression of P65 and Snai1 significantly improved I/R- and UUO-induced kidney injury by alleviating tubulointerstitial inflammation and fibrosis, and potently abrogated the transition to CKD.

Conclusions

A red blood cell–derived extracellular vesicle platform targeted Kim-1 in acutely injured mouse kidney and delivered siRNAs for transcription factors P65 and Snai1, alleviating inflammation and fibrosis in the tubules.

AKI, a common syndrome with high in-hospital morbidity and mortality, is associated with increased long-term risks of CKD, ESKD, and other organ dysfunction.^1–4 Currently, there is no definitive therapy for established AKI in clinic.

RNA interference (RNAi) therapeutics, such as small-interfering RNA (siRNA) and microRNA, are shedding light on the treatment of various diseases by targeting the undruggable disease-causing genes.^5–7 For example, an siRNA targeting P53 was designed to protect against ischemic and nephrotoxic AKI.^8,9 However, achieving efficient delivery of RNAi therapeutics to target organs or specific cell types remains a major translational challenge.^10–12 Although a wide range of delivery nanocarriers such as liposomes, dendrimers, and calcium phosphate nanoparticles have been developed, these synthetic materials have concerns of immunogenicity, cytotoxicity, and rapid clearance from the circulation.^10–12 Hence, a more safe and effective strategy for RNAi delivery remains elusive.

Extracellular vesicles (EVs) are nanoscale membrane particles secreted by most mammalian cells that have emerged as important mediators of intercellular communication by transferring protein and nucleic acid contents between cells.^13,14 As the natural delivery tool, EVs possess many favorable properties, such as stability, biocompatibility, low immunogenicity, and biologic barrier permeability, enabling them to be versatile nanocarriers for a large variety of cargoes.^15–19 Recent studies have successfully loaded RNAi therapeutics, including siRNAs, short hairpin RNAs, microRNAs, and antisense oligonucleotides into EVs, leading to robust gene silencing without apparent side effects.^20–23 Moreover, EVs can be further functionalized with peptides or other surface ligands that confer tissue-specific targeting.^15,16 Thus, an EV-based delivery platform shows considerable promise to manipulate RNAi drugs for AKI therapy. However, to date, there is no effective targeting method to the injured kidney, and no proper therapeutic targets for AKI.

Here, we developed an EV-based delivery platform for RNAi therapeutics to protect against AKI. As shown in Figure 1A, the delivery platform is on the basis of: (1) a peptide that is capable of recognizing and reaching the injured kidney; (2) red blood cell (RBC)–derived EVs (REVs); and (3) potential therapeutic targets for RNAi therapy. Specifically, phage display was used to identify peptides that bind to the kidney injury molecule-1 (Kim-1), which is recognized as a marker for tubular injury during AKI.^24–27 Then, RNA-sequencing (RNA-seq) was performed to characterize the transcriptome of ischemic kidney to explore the potential therapeutic targets. We demonstrated that REVs targeted with the Kim-1–binding LTHVVWL peptide allow for an efficient delivery of P65 and Snai1 siRNAs to the injured tubules, thereby mitigating kidney injury and fibrosis development in murine models of AKI after renal ischemia/reperfusion (I/R) injury or unilateral ureteral obstruction (UUO).

Figure 1.

Identification of Kim-1–targeting peptides by phage display. (A) Schematic illustration of the preparation of REV-based kidney-targeted RNAi therapy against AKI. (B) Table of the five candidate Kim-1–binding peptides identified by in vitro phage display. (C) His-tagged peptides were coated on nickel beads, then incubated with kidney lysates from I/R mice. The captured Kim-1 protein was analyzed with Western blot to examine the binding affinity of candidate peptides to Kim-1. n=3 independent experiments. (D, E) For in vivo biodistribution, mice subjected to unilateral renal I/R were administered intravenously with FITC-labeled peptides. n=3 mice per group. (D) Ex vivo imaging of kidneys (left: sham, right: I/R) and quantitative analysis of the I/R/sham ratio at 6 hours after injection. (E) Correlation analysis between the average radiance of FITC-labeled peptides and Kim-1 protein in I/R kidney. (F) Different concentrations of FITC-labeled peptides were incubated with Kim-1–coated plates, and the bound peptides were imaged with IVIS spectrum. n=3 independent experiments. (G) Representative confocal images showing the colocalization of FITC-labeled LTH or Scrbl with Kim-1 in sections from I/R kidney. Scale bar, 10 μm. Quantification on the basis of 24 sections from three mice. (H) Representative confocal images showing the binding of FITC-labeled LTH or Scrbl to Kim-1-overexpressed TECs. Scale bar, 20 μm. Quantification on the basis of 30 fields from three independent experiments. The specificity of LTH in recognizing Kim-1 is determined by the ratio of LTH⁺Kim-1⁺/LTH⁺ tubules or cells. Data are presented as mean±SD, *P<0.05, **P<0.01 versus LTH, one-way ANOVA.

Methods

In vitro Biopanning

The Ph.D.-7 phage display library (New England BioLabs) was used for in vitro biopanning experiments. Briefly, ELISA plates were precoated with 40 μg of Kim-1 purified protein (Cloud-Clone Corp., China) at 4°C overnight, and unbound protein was removed by washing with PBS containing 0.05% Tween (PBS-T), followed by blocking with 5% BSA at 30°C for 2 hours. Phages at a titer of 1 × 10¹¹ were added to Kim-1–coated plates and incubated at 37°C for 2 hours with gentle shaking. The unbound phages were washed out with 0.05% PBS-T whereas the Kim-1–bound phages were eluted with glycine-HCl buffer (pH 2.2), followed by neutralizing with Tris-HCl buffer (pH 9.1). Eluted phages were titered and amplified using Escherichia coli TG1 for subsequent cycles. After five rounds of biopanning, the polyclonal phage mixture from each round was tested for recognition of Kim-1 by polyclonal phage ELISA. Then 96 phage clones randomly picked from round two and round four were tested for binding to Kim-1, using monoclonal phage ELISA. The top five clones were selected for verification and the inserted DNA was identified by high-throughput sequencing at the AtaGenix Co. Ltd.

Binding Affinity Analysis of Candidate Peptides

In vivo Verification

All of the peptides with a purity >95% were synthesized and provided by Chinapeptide Co. Ltd. Mice with unilateral renal I/R injury were injected intravenously with FITC-labeled peptides at a dose of 25 mg/kg upon reperfusion. Mice were euthanized after 6 hours of injection, and major organs including the kidney, heart, lung, liver, and spleen were harvested for imaging with IVIS Spectrum imaging system (PerkinElmer).

Kim-1 Pull-down Assay

Nickel Dynabeads (Beaver) were precoated with 100 μg of His-tagged peptides (Chinapeptide) for 2 hours at 4°C, then washed three times with washing buffer (20 mM Phosphate Buffer, 500 mM NaCl, and 75 mM imidazole, pH 7.4) to remove unbound peptides.¹⁸ Then 300 μg of protein lysates from I/R kidney were added to the recovered beads with peptides attached, and the lysate and bead mixtures were allowed to mix using a bench-top rotator for 1 hour at 4°C. After incubation, beads were washed with washing buffer three times, and bound proteins were eluted with elution buffer (20 mM Phosphate Buffer, 500 mM NaCl, and 500 mM imidazole, pH 7.4), and detected using Western blot with antibody against Kim-1 (AF1817, R&D Systems).

In vitro Binding Assay

ELISA plates were coated with 4 μg of recombinant Kim-1 (Cloud-Clone Corp., China) at 4°C overnight. After washing the wells three times with 0.05% PBS-T, the plates were blocked with 5% of BSA in PBS at room temperature for 1 hour. Different concentrations of FITC-labeled MFPSSFL, IQPFWVI, LTHVVWL and scrambled peptides were then added into the Kim-1–coated plates and incubated at 37°C for 2 hours. After washing the wells five times with 0.05% PBS-T, the bound peptides in each well were quantified by IVIS Spectrum imaging system.

REV Purification from RBCs

Blood samples were first collected in a tube containing sodium citrate from C57BL/6 mice (18–20 g), then diluted with PBS, and added slowly into a centrifugation tube containing the Ficoll-Paque layer (GE Healthcare). After centrifugation at 500 g for 30 minutes at 18°C, the serum and buffy coat were carefully removed, and isolated RBCs were diluted in PBS and treated with 10 μM calcium ionophore (Abcam) at 4°C for 2 days. To purify REVs, RBCs and cell debris were removed by centrifugation at 2000 g for 20 minutes and 10,000 g for 30 minutes at 4°C. This resulting supernatant was then filtered using 0.45 μm filters, and a pellet was recovered at 100,000 g in a TY70Ti rotor after 2 hours of ultracentrifugation (Beckman). Purified REVs were resuspended in sterile PBS for the next experiments. Nanoparticle tracking analysis (NTA) (ZetaView PMX 110, Particle Metrix) and transmission electron microscopy (TEM) were performed to examine the size distribution and morphology of REVs, respectively. EV-associated markers (Alix, sc-53540, Santa Cruz Biotechnology; CD63, sc5275, Santa Cruz Biotechnology; and CD81, 10037, Cell Signaling Technology) and RBC marker (hemoglobin A, ab92492, Abcam) were detected by Western blotting. The protein concentration of the REVs was quantified using the bicinchoninic acid protein assay kit (ThermoFisher Scientific).

TEM Analysis of EVs

The EV samples were fixed 1:1 with 2% glutaraldehyde for 30 minutes. A fixed sample of 10 μl was pipetted on a 200-mesh nickel grid for 15 minutes, and subsequently stained with 2% phosphotungstic acid for 5 minutes at room temperature, then air dried. Images were acquired using a Hitachi HT7700 transmission electron microscope.

Conjugation of LTH Peptide to REVs

LTH peptides were conjugated onto REVs via a copper-free azide-alkyne cycloaddition reaction, as described previously.^28,29 First, the reactive azadibenzocyclooctyne (DBCO) groups were introduced to REV surface using a heterobifunctional crosslinker. Next, 10 μM DBCO-sulfo-NHS (Sigma) was added to 1 mg/ml REVs in PBS, and the DBCO-sulfo-NHS and REVs mixture was allowed to mix using a bench-top rotator for 4 hours at room temperature. Unconjugated DBCO-sulfo-NHS was removed by ultrafiltration with a 100 kDa tube (Millipore), twice. Then, the DBCO-conjugated REVs were ready to link to azide-functionalized peptide via triazole linkages. Next, 2 μM LTH or scrambled peptide with azide was added to DBCO-conjugated REVs in PBS, and the reaction mixture was mixed on a rotator at 4°C for 12 hours. To purify REVs from the reaction mixture, samples were ultrafiltrated twice with a 100 kDa tube, followed by ultracentrifugation at 100,000 g for 2 hours. To verify whether LTH was successfully conjugated onto REVs, FITC-labeled LTH was loaded onto PKH26 (Sigma Aldrich) labeled REVs, then applied to a glass slide and imaged by a confocal microscope (FV3000, Olympus). To quantify the amount of conjugated LTH, the FITC fluorescence intensity of 1 mg/ml REV_LTH was detected using a fluorescence detection plate reader, with excitation at 495 nm and emission at 520 nm. A standard curve of free FITC-LTH was used to calculate the concentration of peptide on REV_LTH.

Cell Culture and Intervention

HEK293 cells were cultured in DMEM/F12 (Gibco) supplemented with 10% FBS (Gibco) and 1% penicillin-streptomycin (Invitrogen). Immortalized mouse tubular epithelial cells (TECs) (a gift from J. B. Kopp, National Institutes of Health) were cultured in DMEM/F12, supplemented with 10% FBS and 1% penicillin-streptomycin. TECs and HEK293 cells were transfected with a lentivirus-expressing mouse HAVCR1 gene with RFP (GeneChem) to overexpress and label Kim-1 in vitro. To test the binding of LTH to Kim-1, 5 μM of FITC-labeled LTH was incubated with TECs expressing Kim-1–RFP (10⁵ cells per confocal dish) for 6 hours, then cells were fixed for immunofluorescence staining. To test the targeting of REV_LTH, 10 μg of REV_LTH was applied to HEK293 cells expressing Kim-1–RFP (10⁵ cells per confocal dish), and immediately observed under an FV3000 confocal microscope for live-cell imaging.

In vivo Distribution of REV_LTH

FITC-labeled LTH or scrambled peptide was loaded onto REVs, and 200 μg of REV_LTH or REV_Scrbl was injected intravenously into I/R mice upon reperfusion. Mice were euthanized and perfused with 30 ml of cold PBS after 12 hours of injection. Major organs including the kidney, heart, lung, liver, and spleen were harvested for the imaging with IVIS Spectrum imaging system. Moreover, kidneys were obtained and embedded in optimum cutting temperature medium and frozen. Kidney sections were stained with Kim-1 (MA5–28211, Invitrogen) to analyze the homing of REV_LTH and REV_Scrbl to injured tubules.

RNA-seq

RNA was isolated from the cortex of sham or I/R kidney, and was checked for the integrity and quantity using the Agilent RNA PicoChip Kit and the Agilent 2100 Bioanalyzer System. The Illumina TruSeq kit was used for cDNA library preparation. cDNA libraries and next-generation sequencing were performed at the Biomarker Co. Ltd. Seven samples were sequenced with an Illumina HiSeq2500, obtaining 30–40 million reads per sample. Genes were considered significantly differentially expressed if the multiple test corrected P value was <0.05 and absolute fold change >2. Analysis of the differentially expressed genes was performed using BMKCloud (www.biocloud.net).

Loading of siRNA to REV_LTH

All of the siRNAs used in the study were synthesized by GenePharma Co. Ltd., which comprised a 19 bp duplex region followed by 2 bp phosphorothioated overhangs. Pyrimidines were modified with 2′-o-methyl (2′-OMe), and a triethylene glycol-cholesterol was conjugated to the 5′ sense strand. The sequences of siP65 and siSnai are as follows: (1) siP65: sense: GGAGUACCCUGAAGCUAUATT; antisense: UAUAGCUUCAGGGUACUCCTT. (2) siSnai1: sense: GGAAGAUCUUCAACUGCAATT; antisense: UUGCAGUUGAAGAUCUUCCTT. Loading of siRNA by coincubation of REV_LTH with cholesterol-conjugated siRNA (cc-siRNA) was performed as previously described.^30–32 Specifically, 100 μg of REV_LTH was mixed with 10 μg of cc-siRNA in 100 μl RNase-free PBS at 37°C for 90 minutes with gentle shaking. siRNA-loaded REV_LTH was purified from the mixture by ultracentrifugation at 100,000 g for 2 hours. Notably, freshly prepared siRNA-loaded REV_LTH were used for every single assay, in both in vivo and in vitro experiments. To assay the siRNA loading efficiency, the cc-siRNA to be loaded were labeled with Cy3 at 5′ antisense strand. After loading siRNA as described above, the siRNA-loaded REV_LTH were precipitated down by ultracentrifugation, and the unloaded Cy3–cc-siRNA were collected from the supernatant. The Cy3 fluorescence of the supernatant (unloaded siRNA) was measured with a fluorescence detection plate reader with excitation at 547 nm and emission at 570 nm, using a calibration curve for the quantification. The siRNA loading efficiency was calculated through the equation below:

$$\text{siRNA loading efficiency}=\frac{\text{Input siRNA}-\text{free siRNA}}{\text{Input siRNA}}$$

To visualize the siRNA-loaded REV_LTH, Cy3-labeled cc-siRNA and FITC-labeled LTH were loaded onto 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine perchlorate (DID) (Invitrogen) labeled REVs, and images were taken using FV3000 confocal microscope.

RNase and FBS Digestion Experiment

To determine the localization of siRNA in REV_LTH, RNase A digestion with or without Triton X-100 was carried out. Purified REV_LTH-siP65 was incubated with 2 mg/ml of RNase A (Beyotime) at 37°C for 30 minutes, followed by addition of 10× concentrated RNase inhibitor (Beyotime). When both RNase A and Triton X-100 treatments were required, purified REV_LTH-siP65 were first subjected to treatment with 1% Triton X-100 (Beyotime) for 30 minutes at 37°C, after which RNase A was added. Then, the samples were loaded into 1% for electrophoresis in tris acetate buffer to test the degradation of siP65. The gel was run at 120 V for 20 minutes and subsequently imaged via a gel imaging system (Tanon 3500). To examine whether 2′-OMe modification can improve the stability of siRNA on REV_LTH, purified REV_LTH-siP65 with or without 2′-OMe modification was incubated with 50% FBS at 37°C for 8 hours, and the degradation of siP65 was evaluated using gel retention assay as described above.

Animal Models and Therapeutic Experiments

All animal experiments were performed in accordance with standard guidelines for the care and use of laboratory animals and were approved by the Institutional Animal Care and Use Committee of Southeast University (20190112009). Male C57BL/6 mice (8–10 weeks old, weighing 20–22 g) were purchased from the National Model Animal Center of Nanjing University, and housed in ventilated cages on a 12/12 hour light/dark cycle at 21–23°C and 40%–60% humidity. Mice were allowed free access to food and water. Bilateral I/R was induced by clamping of both renal arteries for 35 minutes at 37°C with a bilateral flank approach. The success of I/R was monitored by checking the kidney color after clamping and after removing the clamps. In sham operations, both kidneys were exposed but without induction of ischemia. Warm saline was intraperitoneally injected after the surgery for volume supplementation. To establish a model of CKD transitioned from ischemic AKI, a 40-minute unilateral left renal I/R approach was adopted. For the UUO model, the left ureter of mice was exposed by flank incision and ligated with 6–0 silk at two points, close to the renal pelvis. RNAi treatment was performed at the time of reperfusion for ischemic AKI and 1 day after UUO for obstructive AKI. Specifically, 10 μg of siRNA associated with 100–150 μg of REV_LTH-siRNA were administered by tail-vein injection, once a day for 5 consecutive days. Bilateral I/R injury, unilateral I/R injury, and UUO were maintained for 5 days, 4 weeks, and 14 days, respectively.

Histologic Analysis

Kidneys were fixed in 4% formaldehyde, embedded in paraffin, and sectioned to 5 μm thickness. To evaluate kidney injury score, three random tissue sections per mouse were assessed on the basis of periodic acid–Schiff (PAS) staining and scored using ImageJ software (National Institutes of Health).³³ Briefly, a graticule grid consisting of 100 grids was superimposed on each image (×20 magnification). The histology of each grid was scored as normal (0 = normal histology) or abnormal (1 = TEC swelling, exfoliation, or necrosis, brush border loss, cast formation, tubular dilation, tubular atrophy, or immune cell infiltration), and then all 100 grids were summed to obtain an injury score for each image. Results for each mouse were averaged. To assess interstitial fibrosis, at least ten random images of Masson’s trichrome staining were obtained at ×20 magnification and the percentage of renal fibrosis as defined by blue staining was calculated. The mean values of the fibrotic area were reported.

Immunohistochemistry and Immunofluorescence Staining

Immunohistochemistry staining was performed using the UltraSensitive SP IHC Kit (Maixin) according to the manufacturer’s protocol. Briefly, formalin-fixed, paraffin-embedded kidney sections were deparaffinized and rehydrated, and antigens were retrieved by boiling the sections for 10 minutes in EDTA solution (pH 8). Endogenous peroxidase was blocked with 4% hydrogen peroxide. The sections were incubated with primary antibodies against P-p65 (3033, Cell Signaling Technology) and Snai1 (ab180714, Abcam) overnight at 4°C, followed by incubation with the biotinylated secondary antibody and streptavidin peroxidase solution. Diaminobenzidine (Maixin) was used as chromogen. For immunofluorescence staining, fixed cells or kidney sections were permeabilized with 0.1% Triton X for 10 minutes. After blocking with 5% BSA for 1 hour, primary antibodies against KIM-1 (MA5–28211, Invitrogen), P-p65 (3033, Cell Signaling Technology), Snai1 (ab180714, Abcam), CD68 (ab955, Abcam), and CD3 (ab16669, Abcam) were incubated at 4°C overnight, followed by incubation with the corresponding secondary antibodies for 1 hour at room temperature. Images were obtained by FV3000 confocal microscope. Isotype-matched antibodies were used as negative controls (Supplemental Figure 1).

Quantitative Real-time PCR Assay

Tissues were snap-frozen in liquid nitrogen upon harvesting. RNA was extracted using the TRIzol (Takara) then reverse transcribed using the PrimeScript RT Master Mix (Takara) following the manufacturer’s instructions. Real-time quantitative PCR (RT-qPCR) was performed using TB Green Premix Ex Taq (Takara). Expression levels were normalized to β-actin, and the results were reported as 2^-ΔCt values. Primers used are listed in Supplemental Table 1.

Western Blot

Renal tubules were isolated from kidneys of mice by differential sieving for protein detection, as described previously.³⁴ Briefly, renal cortex was dissected and minced into pieces, followed by grinding with an 80-mesh stainless steel sieve and then filtering through a 100-mesh steel sieve. Tubule fractions in 100-mesh steel sieve were collected and centrifuged at 1500 rpm for 5 minutes. The supernatant was discarded, and the sediment washed once with sterile DMEM was collected and homogenized in radioimmunoprecipitation assay buffer containing protease inhibitors (Roche). Protein content was determined using a bicinchoninic acid protein assay kit (ThermoFisher Scientific). Using Bis-Tris gel (Invitrogen), 30–50 μg of protein was separated by SDS electrophoresis and transferred onto polyvinylidene difluoride membrane (Millipore) with a wet-transfer system. Membrane was blocked in 5% BSA in tris buffered saline–Tween 20 for 1 hour at room temperature and incubated with primary antibody overnight at 4°C. After washing the membrane with tris buffered saline–Tween 20 it was incubated with HRP-conjugated secondary antibody (Abcam) for 90 minutes at room temperature. The signal was detected using the ECL Detection System (GE Healthcare). Primary antibodies were as follows: KIM-1 (AF1817, R&D Systems), P-p65 (3033, Cell Signaling Technology), Snai1 (ab180714, Abcam), and β-actin (sc47778, Santa Cruz Biotechnology).

Statistical Analysis

Data were presented as the mean±SD. Differences between two groups were tested using the two-tailed t test. Differences among multiple groups were tested by one-way ANOVA. P<0.05 was considered significant.

Results

Identification of Kim-1–targeting Peptides by Phage Display

Kim-1 is a type I transmembrane glycoprotein and is selectively expressed on injured TECs,^24–27 which are the nidus of injury during AKI.^35–37 First, we examined the expression pattern of Kim-1 in mice subjected to bilateral renal I/R, and found that Kim-1 was significantly upregulated in the proximal tubules at the early stage of ischemia even before the pathologic changes (Supplemental Figure 2), suggesting Kim-1 is a proper candidate for realizing kidney targeting.

Next, in vitro biopanning was carried out with purified Kim-1 protein. After five rounds of screening, the top five positive phage clones from round four displaying the LTIYSHD (LTI), MSTPLLG (MST), MFPSSFL (MFP), IQPFWVI (IQP), and LTHVVWL (LTH) sequences were selected for further evaluation (Figure 1B, Supplemental Table 2). To investigate the Kim-1–binding affinity of candidate peptides, kidney lysates from I/R mice that contained abundant endogenous Kim-1 protein were incubated with His-tagged peptide-coated nickel beads, and proteins directly bound to the beads were then analyzed with Western blot. Stronger Kim-1 bands were detected in LTI, MST, MFP, IQP, and LTH relative to scrambled peptide, suggesting all five candidate peptides can bind to Kim-1, among which LTH showed the strongest binding affinity (Figure 1C). We also tested the ability of FITC-labeled candidate peptides to recognize the injured kidney in mice with unilateral renal I/R. Quantification of the FITC signal showed that LTI, MFP, IQP, and LTH accumulated in the I/R kidney (R) at levels 7.3-fold, 7.9-fold, 8.1-fold, and 14.4-fold higher than in sham kidney (L), respectively (Figure 1D). However, no significant I/R-to-sham contrast ratio was detected in mice administered with MST, most of which accumulated in the lung and liver (Figure 1D and Supplemental Figure 3A), an indication MST is unable to recognize the injured kidney. Correlational analyses revealed a negative correlation between LTI level and Kim-1 expression in I/R kidney (Figure 1E and Supplemental Figure 3B), suggesting LTI is not specific for Kim-1 in vivo. Moreover, we evaluated the binding of MFP, IQP, and LTH to recombinant Kim-1 in vitro, and found both IQP and LTH strongly bound to Kim-1 at high concentrations; however, only LTH exhibited a high binding affinity at low concentrations (Figure 1F). These results indicate LTH is a preferred candidate for Kim-1 targeting. Notably, the specificity of LTH in recognizing Kim-1⁺ tubules in vivo or Kim-1⁺ TECs in vitro was 69.4±18.6% and 83.5±12.3%, respectively (Figure 1, G and H).

LTH Efficiently Directs REVs to the Injured Kidney

To construct an LTH-functionalized REV-based delivery system (REV_LTH), as shown in Figure 2A, REVs were isolated from the calcium ionophore-treated mouse RBCs, and LTH was conjugated onto REVs via copper-free azide-alkyne cycloaddition reaction (click chemistry).^28,29 Briefly, DBCO groups were first introduced to amine-containing molecules on REV surface using DBCO-sulfo-NHS. Then, azide-modified LTH reacted with DBCO on REVs to form stable triazole linkages through copper-free click chemistry. NTA revealed that, compared with the native REV, there was a slight increase in the diameter of REV_LTH, with the highest peak around 118 nm (Figure 2B). The morphology of REV_LTH was a typical cup-shaped bilipid structure under a TEM (Figure 2B), which is consistent with previously described EVs.^38,39 Western blot showed that REV and REV_LTH were enriched in EV markers (Alix, CD63, CD81) and hemoglobin A, the major RBC protein (Figure 2C). These data illustrated the identity and purity of REV_LTH. To evaluate the loading efficiency of LTH, we labeled REVs and LTH with PKH26 and FITC, respectively, and found approximately 76% of REVs were colocalized with FITC signal (Figure 2D), suggesting a successful conjugation of LTH onto REVs. By comparison with a fluorescent standard curve of free LTH-FITC, it was calculated that 1 mg/ml REV_LTH contained an average of 835 nM peptides. The results collectively indicate a successful loading of LTH onto the REV surface by click chemistry.

Figure 2.

LTH efficiently directs REVs to the injured kidney. (A) Schematic diagram of conjugating LTH to REV surface by a two-step reaction. (B) Size distribution of REV and REV_LTH. Representative TEM image of REV_LTH. (C) Western blot analysis of EV markers (Alix, CD63, CD81) and RBC marker (Hemoglobin A) in REV and REV_LTH. (D–H) For imaging analysis, the scrambled and LTH peptide were labeled with FITC. (D) Representative confocal image of REV_LTH. Yellow shows the colocalization of FITC-labeled LTH with PKH26-labeled REV. Scale bar, 1 μm. Quantification on the basis of 30 fields from three independent experiments. (E) Ex vivo imaging of kidneys from unilateral I/R mice injected intravenously with REV_Scrbl and REV_LTH. n=4 mice (REV_Scrbl); n=6 mice (REV_LTH). (F) Representative confocal images showing the colocalization of REV_LTH with Kim-1–positive tubules in sections from I/R kidney. Scale bar, 30 μm. Quantification on the basis of 30 sections from three mice per group. (G) Live-cell imaging of the cellular uptake of REV_LTH in HEK293 cells expressing Kim-1-RFP at indicated times. Scale bar, 5 μm. See Supplemental Video 1. Zoomed image indicates the binding of REV_LTH to Kim-1. (H) Representative confocal images of REV_Scrbl and REV_LTH in Kim-1⁺ HEK293 cells after 30 minutes of incubation. Scale bar, 20 μm. Quantification on the basis of 150 cells from three independent experiments. Data are presented as mean±SD *P<0.05, ***P<0.001 versus REV_Scrbl, two-tailed t test.

Subsequently, we examined the tissue distribution of REV_LTH. After the induction of renal unilateral I/R injury, mice were intravenously injected with REV_LTH and REV_{Scramble (Scrbl)}, whose peptides were labeled with FITC. Ex vivo imaging revealed an approximately five-fold increase in renal (I/R) accumulation of REV_LTH compared with REV_Scrbl (Figure 2E), and a minor, nonstatistically significant reduction in pulmonary, hepatic, and splenic accumulation of REV_LTH (Supplemental Figure 4). Specifically, robust amounts of REV_LTH, but not nontargeted REV_Scrbl, spread into the Kim-1⁺ tubules (Figure 2F). Notably, whereas REV_LTH showed an increased accumulation in the injured kidney and Kim-1⁺ tubules, substantial numbers of REV_LTH were sequestered in the liver and spleen (Supplemental Figure 4), suggesting REVs targeted with LTH are still unable to escape the nonspecific phagocytosis by the reticuloendothelial system.^40,41 Moreover, we transfected HEK 293 cells with RFP-labeled Kim-1 and accessed their uptake of REV_LTH using confocal live-cell imaging. As shown in Figure 2G and Supplemental Video 1, REV_LTH trafficked from the extracellular region to the cell membrane, and then captured into the cytosol by interacting with Kim-1. Meanwhile, LTH significantly augmented the REV uptake in Kim-1⁺ cells (Figure 2H). All of these findings demonstrate that LTH efficiently guides the homing of REV_LTH to the injured tubules, which could be harnessed as a promising vector for AKI therapy.

P65 and Snai1 Are Potential Therapeutic Targets Involved in I/R Pathogenesis

Although considerable progress has been made in elucidating the cellular and molecular mechanisms driving AKI, there are no specific treatments for AKI or prevent its progression to CKD per se. To identify the potential therapeutic targets, we performed RNA-seq to characterize the transcriptome signature of mouse I/R kidney. Then 1 day after reperfusion, 1850 genes were differentially expressed (968 upregulated and 882 downregulated) between sham and I/R kidney (Supplemental Figure 5 and Supplemental Table 3). We observed an enhanced expression of Havcr1 (Kim-1), Lcn2 (neutrophil gelatinase-associated lipocalin), and Krt20, suggesting a severe tubular injury caused by I/R.^24,25,42,43 However, defining a therapeutic target from so many differentially expressed genes is difficult to achieve.

Transcription factors regulate downstream genes and serve as the convergence points of signaling pathway, thus targeting transcription represents a highly promising intervention strategy. A total of 45 upregulated transcription factors were identified in I/R kidney (Figure 3A and Supplemental Table 4). Among them, Sox9 and Stat3 can activate and regulate kidney tissue repair programs,^44–46 indicating an endogenous kidney repair at the early stage of injury. Here, we focused on the transcription factors that may promote the progression of AKI. Rela (P65) belongs to the NF‐κB family, which plays a central role in the induction and progression of inflammation.⁴⁷ Inhibition of NF‐κB pathway has been reported to decrease inflammation and ameliorate ischemic, obstructive, and toxic AKI.^48–50 Snail is recognized as a major transcription factor that induces epithelial-mesenchymal transition (EMT). Injury-induced reactivation of Snai1 in TECs triggers a partial EMT that markedly contributes to renal inflammation and fibrosis.^51,52 Therefore, we speculated that P65 and Snai1 can be effective therapeutic targets for AKI. The elevated P65 and Snai1 mRNA levels were confirmed by RT-qPCR (Figure 3B). Immunostaining showed that P-p65 and Snai1 were strongly induced in I/R tubules (Figure 3C), while colocalizing with Kim-1⁺ TECs (Figure 3D). Next, we investigated whether suppressing P65 and Snai1 was sufficient to alleviate kidney injury and block the AKI to CKD transition.

Figure 3.

P65 and Snai1 are potential therapeutic targets involved in I/R pathogenesis. (A) Heat map of the upregulated transcription factors in I/R kidney compared with sham kidney. n=3 mice (sham); n=4 mice (I/R). (B) RT-qPCR analysis of P65 and Snai1 mRNA levels in kidney tissues. n=8 mice (sham); n=18 mice (I/R). (C) Immunohistochemical analysis of P-p65 and Snai1 expression in kidney sections. Scale bar, 50 μm. Zoom panels show the nuclear localization of P-p65 and Snai1 in tubules of I/R kidney. (D) Representative confocal images of P-p65/Kim-1 and Snai1/Kim-1-immunostained I/R kidney. Scale bar, 20 μm. Data are presented as mean±SD **P<0.01, ***P<0.001 versus sham, two-tailed t test.

Targeting of P65 and Snai1 in Tubules Using siRNA-loaded REV_LTH

To efficiently deliver siRNAs, cholesterol-conjugated 2′-OMe–modified P65 and Snai1 siRNAs (siP65, siSnai1) were tethered to REV_LTH, respectively, by coincubation. A cholesterol moiety enables spontaneous membrane association, and 2′-OMe modification can improve the stability of siRNA.³⁰ NTA and TEM analysis revealed the size distribution and morphology of REV_LTH were largely unchanged on siRNA loading (Figure 4A). Using Cy3-labeled siRNAs, we quantified their loading efficiency on the basis of the fluorescence intensity, and found around 73.3% of siP65 and 72.7% of siSnai1 were loaded to REV_LTH, respectively, whereas lacking cholesterol moiety reduced the loading of both siRNAs to 10%, and controls without REV_LTH showed pelleting as low as 9% (Figure 4B). Figure 4C further proved the loading of both siP65 and LTH to REVs. Next, RNase A digestion was performed to determine the localization of siRNAs on REV_LTH. In total, <40% of siP65 was retained with REV_LTH after RNaseA alone treatment, and none was detected when the REV membrane was damaged by Triton (Figure 4D), suggesting most siP65 is present on the surface of REV_LTH rather than in the lumen, which is consistent with previous reports.³⁰ Given the instability of siRNAs in vivo, we assessed whether 2′-OMe modification can protect REV_LTH-siP65 from serum degradation. As expected, approximately 80% of 2′-OMe–modified siP65 versus 20% of unmodified siP65 remained intact after incubation in 50% FBS at 37°C (Supplemental Figure 6A), and a prolonged circulation time of REV_LTH-siP65 caused by 2′-OMe modification (Supplemental Figure 6B). These results indicate the approach we devised for loading siP65 and siSnai1 to REV_LTH is feasible and effective, with high stability. Furthermore, we found REV_LTH delivered 2.3-fold more Cy3-siP65 into I/R kidney compared with nontargeted REV_Scrbl (Figure 4E), suggesting displaying siRNAs on the surface of REV_LTH is not likely to affect its targeting capacity.

Figure 4.

Targeting of P65 and Snai1 in tubules using siRNA-loaded REV_LTH. (A) Size distribution and representative TEM images of REV_LTH-siP65 and REV_LTH-siSnai1. (B) The loading efficiency of Cy3-labeled siRNA with or without cholesterol modification or cholesterol-modified siRNA without REV_LTH. n=3 independent experiments. ***P<0.001 versus REV_LTH+chol-siRNA. (C) Representative confocal image of REV_LTH-siP65. White shows the colocalization of FITC-labeled LTH and Cy3-labeled siP65 with DID-labeled REVs. Scale bar, 1 μm. (D) Gel retention assay of siP65 in REV_LTH-siP65 treated with RNase or RNase combined with Triton. n=3 independent experiments. ***P<0.001 versus untreated control. (E) Ex vivo imaging of I/R kidney from unilateral I/R mice administered intravenously with REV_Scrbl-siP65 and REV_LTH-siP65. n=3 mice (REV_Scrbl-siP65); n=6 mice (REV_LTH-siP65). **P<0.01 versus REV_Scrbl-siP65. (F, G) Mice were subjected to 35 minutes of bilateral renal I/R injury and received treatment every 24 hours upon reperfusion, five times in total. All of the mice were sacrificed at day 5 post-I/R injury. (F) Western blot analysis of P-p65 and Snai1 protein levels in tubules. n=4 mice per group. *P<0.05, **P<0.01, ***P<0.001 versus REV_LTH-siRNA. (G) Immunohistochemical analysis of P-p65 and Snai1 expression in I/R kidney treated with REV_LTH-siScrbl or REV_LTH-siRNA. Scale bar, 50 μm. Quantification on the basis of four mice, with at least five sections counted in each. ***P<0.001 versus REV_LTH siScrbl. Data are presented as mean±SD one-way ANOVA (B, D, F), two-tailed t test (E, G).

We then verified the functionality of REV_LTH-mediated RNAi in mice who underwent bilateral renal I/R injury and received treatment on reperfusion once a day for 5 consecutive days. Notably, REV_LTH-siP65 and REV_LTH-siSnai1 induced a significant reduction of P-p65 and Snai1 protein expression in tubules, respectively, as measured by Western blot and immunostaining, whereas REV_LTH with scrambled siRNA or nontargeted REV_Scrbl showed no noticeable RNAi effects (Figure 4, F and G, Supplemental Figure 7, A and B). By contrast, nontargeted REV_Scrbl-siP65 and REV_Scrbl-siSnai1 markedly inhibited the mRNA levels of P65 and Snai1 in the lung and liver, which was partly improved by REV_LTH-mediated delivery (Supplemental Figure 7, C and D). These results demonstrate that REV_LTH efficiently delivers siP65 and siSnai1 to the injured kidney, leading to a robust suppression of the candidate therapeutic targets.

REV_LTH-siP65/siSnai1 Combination Therapy Alleviates I/R-induced Ischemic AKI

We first evaluated the therapeutic efficacy of REV_LTH-siP65/siSnai1 combination therapy in murine ischemic AKI. Mice were subjected to 35 minutes of bilateral I/R and received either REV_LTH-siP65/siSnai1 or REV_LTH-siScrbl at the time of reperfusion, as outlined in Figure 5A. We found there were no differences in serum creatinine and BUN at 24 hours, indicating similar degrees of initial injury and that a single dose of REV_LTH-siP65/siSnai1 is insufficient to improve ischemic AKI. After a total of five doses of REV_LTH-siP65/siSnai1, serum creatinine and BUN were significantly decreased on day 5, compared with those of mice received REV_LTH-siScrbl (Figure 5, B and C), accompanied by a reduction of urine levels of neutrophil gelatinase-associated lipocalin (Supplemental Figure 8A). Consistent with improved kidney function, I/R-induced tissue damage, including necrotic tubules, cast formation, dilated tubules, and immune cell infiltration, was significantly alleviated by REV_LTH-siP65/siSnai1 (Figure 5, D and E). We also tested parameters associated with renal inflammation and fibrosis regulated by P65 and Snai1. We found the interstitial infiltration of macrophages and T cells (Figure 5F) and the mRNA expression levels of proinflammatory cytokines (TNFα, IL6, CCL2) (Figure 5G) and fibrotic markers (Vimentin, ACTA2, COL1A1) (Figure 5H), and the apoptosis and G2/M-arrested TECs (Supplemental Figure 8, B and C)in I/R kidney, were markedly repressed by the treatment of REV_LTH-siP65/siSnai1. These results demonstrate that REV_LTH-siP65/siSnai1 combination therapy was capable of inhibiting I/R-induced renal inflammation and fibrosis development.

Figure 5.

REV_LTH-siP65/siSnai1 combination therapy alleviates I/R-induced ischemic AKI. (A) Schematic diagram of the experimental design. In brief, mice were subjected to bilateral renal I/R and received REV_LTH-siP65/siSnai1 or REV_LTH-siScrbl every 24 hours upon reperfusion, five times in total. (B) Serum creatinine and (C) blood urea nitrogen at day 0, 1, and 5 post-I/R. n=6 mice (sham, I/R, siScrbl); n=8 mice (siP65/siSnai1). (D) Quantification of kidney injury on the basis of PAS staining. n=6 mice (sham, I/R, siScrbl); n=8 mice (siP65/siSnai1). (E) Representative images of PAS staining. Scale bar, 100 μm. (F) Immunostaining of macrophages (CD68) and T cells (CD3) in the tubulointerstitial region. Scale bar, 20 μm. Quantification on the basis of six mice with at least five sections counted in each. (G) RT-qPCR analysis of proinflammatory cytokine mRNA levels in kidney tissues. n=6 mice per group. (H) RT-qPCR analysis of fibrotic factor mRNA levels in kidney tissues. n=6 mice per group. (I, J) Mice were subjected to 40 minutes of unilateral renal I/R, and REV_LTH-siP65/siSnai1 or REV_LTH-siScrbl was administered upon reperfusion and continued every 24 hours for 5 consecutive days. Mice were euthanized at day 28 after reperfusion. n=6 mice per group. (I) Representative images of PAS staining of I/R kidneys treated with siP65/siSnai1 or siScrbl. Scale bar, 100 μm. (J) Representative images of Masson trichrome staining and quantification of fibrotic area. Scale bar, 100 μm. Data are presented as mean±SD *P<0.05, **P<0.01, ***P<0.001 versus I/R or siScrbl, one-way ANOVA (B–D), two-tailed t test (F–J).

To analyze the potential of REV_LTH-siP65/siSnai1 combination therapy in preventing AKI to CKD transition, a 40-minute unilateral I/R approach was adopted, and RNAi treatment was administered daily for 5 consecutive days, commencing on day 0 immediately after reperfusion. Then 4 weeks after I/R, when compared with REV_LTH-siScrbl treatment, the tubulointerstitial damage, including tubular atrophy, flattening and sloughing of TECs, cast formation, immune cell infiltration, and extracellular matrix accumulation, was dramatically ameliorated by the early treatment of REV_LTH-siP65/siSnai1 (Figure 5, I and J). Notably, the combination therapy also reduced the protein expression of P-p65 and Snai1, and the mRNA levels of proinflammatory cytokines and fibrotic markers in the kidneys at day 28 post-I/R (Supplemental Figure 9). Moreover, we found the renoprotective effects of the siP65/siSnai1 combination were greater than those of individual siP65 or siSnai1 (Supplemental Figure 10), suggesting the synergistic effects derived from the combination therapy. These data indicate that REV_LTH-siP65/siSnai1 combination therapy can interrupt AKI to CKD progression.

REV_LTH-siP65/siSnai1 Combination Therapy Attenuates UUO-induced Obstructive AKI

Further, we extended our studies to murine obstructive AKI to verify the efficacy of REV_LTH-siP65/siSnai1 combination therapy, in which the Kim-1 was also robustly induced in the proximal tubules (Supplemental Figure 11).⁵³ Then 24 hours after UUO, REV_LTH-siP65/siSnai1 or REV_LTH-siScrbl was intravenously administered daily for 5 consecutive days (Figure 6A). Histologic analysis of kidneys from REV_LTH-siP65/siSnai1-treated mice presented a recovered morphology with a low degree of tubular injury and less collagen deposition on day 14 after UUO (Figure 6, B–E). Coincident with improved kidney injury, macrophage and T cell infiltration (Figure 6F), mRNA expression levels of proinflammatory cytokines and fibrotic factors (Figure 6, G and H), apoptosis of TECs, and G2/M cell cycle arrest (Supplemental Figure 12)were notably lower in REV_LTH-siP65/siSnai1-treated mice than those from mice treated with REV_LTH-siScrbl. Taken together, these results indicate that combined inhibition of P65/Snai1 blocks the progression of UUO-induced kidney injury.

Figure 6.

REV_LTH-siP65/siSnai1 combination therapy attenuates UUO-induced obstructive AKI. (A) Schematic diagram of the experimental design. In brief, mice were treated daily starting at day 1 after UUO surgery with REV_LTH-siP65/siSnai1 or REV_LTH-siScrbl, and were sacrificed on day 14 after surgery. (B, D) Representative images of PAS staining and quantification of kidney injury. Scale bar, 50 μm. n=6 mice (sham, UUO, siScrbl); n=8 mice (siP65/siSnai1). (C, E) Representative images of Masson trichrome staining and quantification of fibrotic area. Scale bar, 100 μm. n=6 mice (Sham, UUO, siScrbl); n=8 mice (siP65/siSnai1). (F) Immunostaining of macrophages (CD68) and T cells (CD3) in the tubulointerstitial region. Scale bar, 20 μm. Quantification on the basis of six mice with at least five sections counted in each. (G) RT-qPCR analysis of proinflammatory cytokine mRNA levels in kidney tissues. n=6 mice per group; (H) RT-qPCR analysis of fibrotic factor mRNA levels in kidney tissues. n=6 mice per group. Data are presented as mean±SD *P<0.05, ***P<0.001 versus UUO or siScrbl, one-way ANOVA (B, C), two-tailed t test (F–H).

We also examined the toxicity of the RNAi therapy. The concentrations of aspartate aminotransferase and alanine aminotransferase in the serum of REV_LTH-siP65/siSnai1-treated UUO mice showed no obvious differences from those of sham mice (Supplemental Figure 13), indicating a lack of toxicity to the liver, the organ that nonspecifically takes up REVs. Additionally, tissue histology further supported no significant toxic damage in any of the major organs (Supplemental Figure 13).

Discussion

RNAi-mediated gene therapy enables precise and personalized treatment of diverse diseases. For AKI therapy, delivery of RNAi therapeutics into specific kidney cells and lack of effective therapeutic targets are major challenges. In this study, we developed a novel REV-based therapeutic platform whereby Kim-1–binding LTH peptides and combinations of siP65/siSnai1 were deployed to protect against AKI. We demonstrated that REVs targeted with LTH efficiently deliver siP65/siSnai1 to injured tubules in ischemic or obstructive kidney, promoting renal recovery by relieving tubulointerstitial inflammation and fibrosis.

EVs have recently emerged as a promising delivery system for RNA drugs. For example, EVs derived from dendritic cells and fibroblasts have been electroporated with siRNAs for the treatment of Alzheimer’s disease²¹ and pancreatic cancer,⁵⁴ respectively. However, the safety and toxicity of EVs from different cell sources is less characterized. Here, we prepared EVs from RBCs because (1) RBCs are intrinsically biosafe, biocompatible, and nonimmunogenic; (2) RBCs are the most abundant cells that can be easily extracted from peripheral blood; and (3) the lack of both nuclear and mitochondrial DNA in RBCs further improves the security of their EVs by avoiding unwanted biologic effects on recipient cells. We found a sufficient number of REVs were harvested from calcium ionophore-treated mouse RBCs, which was consistent with previous reports that Ca²⁺ can provoke the secretion of EVs.^22,55,56 These attractive features make REVs as a nanocarrier with great potential for clinical translation. Notably, native REVs derived from human RBCs of group O blood have been utilized for the delivery of RNA drugs, including antisense oligonucleotides, Cas9 mRNA, and guide RNAs.²² Our study further offered a valid evidence that REVs are well tolerated and can be functionalized with therapeutics and targeting moieties.

TECs are the central nidus of injury during AKI of different etiologies including ischemia, toxins, and obstruction.^26,57 Protecting TECs from injury is of paramount importance to promote renal recovery from AKI. Therefore, delivery of therapeutics into the injured TECs is the primary concern in AKI therapy. Our recent study reported that macrophage-derived EVs are able to deliver IL-10 into the injured TECs through the interaction of integrin α_Lβ₂ on the EV surface with ICAM-1 on TECs.⁵⁸ However, ICAM-1 is not restricted on TECs, but also overexpressed on endothelial cells and immune cells during injury. Thus, a more specific targeting device remains to be developed. In this study, we exploited in vitro phage display in our screens for peptides that recognize Kim-1, because Kim-1 specifically presents on the surface of injured TECs.^24,25 Importantly, we identified the peptide LTH, which showed high binding affinity to Kim-1, enhanced the renal accumulation of REVs at the sites of Kim-1⁺ tubules in ischemic AKI. Moreover, considering TECs highly express Kim-1 not only in AKI but also in CKD,⁵⁹ we propose that REV_LTH could be utilized to target injured TECs in multiple kidney diseases with wide applicability.

Another major concern relating to AKI therapy is the lack of effective therapeutic targets. We performed RNA-seq to clarify the critical genes in driving AKI and focused our analysis on the transcription factors. In general, tubular cells have a great capacity for self-renewal so the kidney can rapidly undergo repair.⁶⁰ In accordance with this notion, transcription factors such as Sox9 and Stat3 that regulate endogenous kidney repair were confirmed in I/R kidneys. However, injured TECs may undergo cell cycle arrest^61–63 and partial EMT,^51,52 which prevents complete repair and leads to inflammation, fibrosis, and the AKI to CKD transition. EMT-related transcription factors, Snai1 and Twist1, are shown to trigger partial EMT and G2/M arrest of TECs, thereby instigating a proinflammatory and profibrotic TEC state.^51,52 Of note, we found a robust increase in mRNA and protein levels of Snai1 in the injured tubules, and the upregulation of P65 expression. Recently, the single nucleus RNA-seq uncovered a cluster of unrepaired proximal TECs in I/R kidney, which has regulon activity for both Relb and NFκB,⁶⁴ providing further evidence that inhibition of P65 and Snai1 may represent a new therapeutic strategy to improve renal repair and reduce fibrosis. Impressively, we demonstrated that employing REV_LTH as “Trojan horses” for the delivery of siP65/siSnai1 efficiently downregulated protein levels of P-p65 and Snai1 in tubules, thereby alleviating I/R- or UUO-induced tubular injury, interstitial inflammation, and fibrosis. Hence, our study revealed that RNAi targeting critical transcription factors is a viable therapeutic strategy for protecting kidney structure and function during injury.

Notably, although Kim-1 is a well-validated tubular injury marker, it is also expressed in immune cells such as T cells and B cells,^65,66 and thus REV_LTH-mediated RNAi may cause off-target gene silencing in these cells. In addition, although REV_LTH can localize to the injured tubules with the aid of Kim-1–targeting peptide, a number of them are unavoidably captured by the liver and spleen of the reticuloendothelial system as they travel through the systemic circulation. Even so, no significant organ toxicity was observed with REV_LTH-mediated RNAi therapy, and targeting Kim-1 remains one of the potential strategies for realizing kidney-targeted drug delivery.

In summary, we have developed a Kim-1–targeted delivery platform-REV_LTH and highlighted its potential in the delivery of siP65/siSnai1 as a promising RNAi therapy for AKI. Our data demonstrated that REV_LTH decorated with Kim-1–binding LTH peptide efficiently home to and deliver siP65/siSnai1 therapeutics to the injured tubules. Consequently, dual suppression of P65 and Snai1 significantly improved I/R- and UUO-induced kidney injury, and abrogated AKI to CKD transition. Our study suggests a transformative potential of REV_LTH-siP65/siSnai1 in controlling renal inflammation and fibrosis with minimal toxicities, and provides a unique platform that may accelerate the development of gene therapy in AKI and other kidney diseases.

Disclosures

B.C. Liu, T.T. Tang, and B. Wang report being inventors on patent 202011592251.5 (in a Preliminary Examination Procedure) related to this work, filed by State Intellectual Property Office of People’s Republic of China. All remaining authors have nothing to disclose.

Funding

This work was supported by National Natural Science Foundation of China grants 82030024, 81720108007, 81670696, 81470922, and 31671194), National Key Research and Development Program grant 2018YFC1314000, Clinic Research Center of Jiangsu Province grant BL2014080, Major State Basic Research Development Program of China grant 2012CB517706, and Postgraduate Research and Practice Innovation Program of Jiangsu Province grant KYCX18_0171.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2020111561/-/DCSupplemental.

Supplemental Figure 1. Images of isotype control of the immunostaining.

Supplemental Figure 2. Expression of Kim-1 in mice with renal I/R.

Supplemental Figure 3. Tissue distribution of candidate Kim-1-binding peptides.

Supplemental Figure 4. Tissue distribution of REV_LTH.

Supplemental Figure 5. Transcriptional profiling of the kidneys from Sham and I/R mice.

Supplemental Figure 6. Stability analysis of the REV_LTH-siP65.

Supplemental Figure 7. Off-target effects of REV_LTH-siP65 and REV_LTH-siSnai1.

Supplemental Figure 8. Therapeutic efficacy of the combined RNAi therapy in ischemic AKI.

Supplemental Figure 9. REV_LTH-siP65/siSnai1 combination therapy attenuates AKI-CKD transition.

Supplemental Figure 10. Comparison of the efficiency of individual or combined RNAi therapy in preventing AKI-to-CKD transition.

Supplemental Figure 11. Expression of Kim-1 in the kidneys of UUO mice.

Supplemental Figure 12. Therapeutic efficacy of the combined RNAi therapy in obstructive AKI.

Supplemental Figure 13. Assessment of toxicity in REV_LTH-siP65/siSnai1-treated UUO mice.

Supplemental Table 1. Primers used in this study.

Supplemental Table 2. In vitro biopanning for Kim-1-binding peptides.

Supplemental Table 3. List of differentially expressed genes.

Supplemental Table 4. List of upregulated transcriptional factors.

Supplemental Video 1. Live-cell imaging.